Dipyrimidine amines: a novel class of chemokine receptor type 4 antagonists with high specificity

Abstract

The C-X-C chemokine receptor type 4 (CXCR4)/stromal cell derived factor-1 (SDF-1 or CXCL12) interaction and the resulting cell signaling cascade play a key role in metastasis and inflammation. On the basis of the previously published CXCR4 antagonist 5 (WZ811), a series of novel nonpeptidic anti-CXCR4 small molecules have been designed and synthesized to improve potency. Following a structure-activity profile around 5, more advanced compounds in the N,N'-(1, 4-phenylenebis(methylene)) dipyrimidin-2-amines series were discovered and shown to possess higher CXCR4 binding potential and specificity than 5. Compound 26 (508MCl) is the lead compound and exhibits subnanomolar potency in three in vitro assays including competitive binding, Matrigel invasion and Gα(i) cyclic adenosine monophosphate (cAMP) modulation signaling. Furthermore, compound 26 displays promising effects by interfering with CXCR4 function in three mouse models: paw inflammation, Matrigel plug angiogenesis, and uveal melanoma micrometastasis. These data demonstrate that dipyrimidine amines are unique CXCR4 antagonists with high potency and specificity.

Figure 1

Structures of nonpeptidic CXCR4 antagonist 1, potential CXCR4 antagonist template 2, potential CXCR4 antagonist lead compound 3, peptidic CXCR4 antagonist 4, and potential CXCR4 antagonist 5.

Figure 2

Comparison of inhibition of CXCR4/CXCL12-mediated cAMP modulation by anti-CXCR4 compounds. (A) CXCR4 positive cell line U87CD4CXCR4 was treated with chemokine CXCL12; (B) CCR5 positive cell line U87CD4CCR5 was treated with CCL5. With pre-treatment (15 min at room temperature) of compounds, 1, 5, 21c, or 26 at various concentrations, the effect of 150 ng/ml of CXCL12 or CCL5 on cAMP reduction in the presence of 5 µM forskolin was measured by using the TR-FRET based LANCE assay kit. While our anti-CXCR4 compounds are effective in counteracting CXCL12 function at as low as 10 nM, 1 requires almost 1000 nM to significantly block CXCL12 function, which is 1000 times less efficient than our compounds in blocking the Gα_i pathway. Compound 5 is our previously reported analog; compounds 21c and 26 are comparable or better than compound 5 in blocking CXCL12 function. The tested compounds have no effect on CCR5 cells.

Figure 3

Suppression effect of anti-CXCR4 compounds on carrageenan-induced mouse paw inflammation. Acute paw inflammation was induced by subcutaneous injection of 50 µL of λ-carrageenan in one hind paw. The mice in the treatment group were all administered CXCR4 antagonists at 10 mg/kg i.p., while 4 was administered at 300 µg/kg and 30 min following carrageenan challenge and daily thereafter. Control animals received corresponding i.p. injections of vehicle. Panel (A) shows the control mouse with left paw induced inflammation by carrageenan; panel (B) shows the 4 treated mouse with left paw induced inflammation by carrageenan with about 50% suppression. The bar graph shows that all tested compounds are effective at suppressing paw inflammation at different degrees. Compounds 10g, 10e, 26, and 21b showed effects comparable to 4.

Figure 4

Inhibitory effect of anti-CXCR4 compounds on Matrigel plug angiogenesis assay in vivo. The mice in the CXCR4 antagonist-treated group received daily subcutaneous injections of the selected compounds (two plugs per mouse) at 10 mg/kg, and 4 at 300 µg/kg. Ten days after matrigel injection, the animals were sacrificed, and the Matrigel plugs were excised, photographed and sliced for H&E staining and processed for hemoglobin assay. Analog 4-treated Matrigel plugs revealed no significant angiogenesis, while the control group exhibited significantly more blood vessels clearly shown by H&E staining (the black arrows in the picture). The column graph shows the tested compounds can inhibit angiogenesis from 30% to 68%, while compound 26 delivers almost the same efficacy as 4 at about 70% angiogenesis inhibition.

Figure 5

Inhibition effect of compound 26 compared to 4 on lung fibrosis induced by Bleomycin. Representative H&E-stained histopathologic sections of untreated (A), 4 was administered at 300 µg/kg, i.p. (B), and 26 administered at 10 mg/kg, i.p. (C) lung tissues on Day 20 after bleomycin treatment. Lung fibrosis is shown by small black arrows in the images. Column graph shows 4 and 26-treated groups experience a significant decrease in lung fibrosis.

Figure 6

Inhibition effect of 26 compared to 4 in a uveal melanoma micrometastasis animal model. Hepatic tissues were collected and fixed in 10% formalin, processed, and H&E stained, and the number of hepatic micrometastasis was counted under a microscope. Micrometastatic clones are shown by small black arrows in the images. The murine model with ocular melanoma metastatic to the liver was treated with 4 at 300 µg/kg i.p. or compound 26 at 10 mg/kg, i.p. once per day starting at 4th day after uveal melanoma inoculation. The animals presented significantly fewer micrometastases than those in the control (PBS-treated) group. Compound 26 decreased the numbers of hepatic micrometastases in a mouse model of human uveal melanoma with efficacy similar to 4, both are about 50% inhibition.

Scheme 1^a

^a reagents and conditions: 1. 2-amino-fluoropyridines, NaBH(OAc)₃, HOAc, ClCH₂CH₂Cl, 61–64%; 2. 2-amino-pyrimidine, NaBH(OAc)₃, HOAc, ClCH₂CH₂Cl, 82%; 3. DMP, CH₂Cl₂, 94%; 4. ArNH₂, NaBH(OAc)₃, HOAc, ClCH₂CH₂Cl, 65–69%.

Scheme 2^a

^a reagents and conditions: 1. 12, Cs₂CO₃, DMF, 75%; 2. 14, NaHCO₃, THF, 94%; 3.fluoropyrimidines (16a–b), DIPEA, DMF, 65–70%; 4. SOCl₂, MeOH, quant.

Scheme 3^a

^a reagents and conditions: 1. 12, DIPEA, DMF, 96%; 2. SOCl₂, MeOH, then 20a–c, DIPEA, DMF,63–72% (2 steps); 3. 22, DIPEA, DMF, 94%; 4. SOCl₂, MeOH, then 20b, DIPEA, DMF, 65% (2 steps); 5. mCPBA, CH₂Cl₂, then 25, dioxane, 25% (2 steps).