Inhibitors of SPSB-iNOS interactions as a potential novel class of anti-infectives

Abstract

The SPRY-domain of the SOCS box protein 2 (SPSB2) plays an important role in the proteasomal degradation of inducible nitric oxide synthase (iNOS). SPSB2 knockout mice show prolonged expression of iNOS and enhanced killing of persistent pathogens such as Mycobacterium tuberculosis and Leishmania major, suggesting that inhibitors of the SPSB2-iNOS interaction represent a potential novel class of anti-infectives. In this study, attempts to discover small molecule inhibitors of SPSB2-iNOS interaction were performed using in silico guided fragment-based drug design (FBDD) approach. The best fragment hit STK441224, however, was found to bind promiscuously to SPSB2 by saturation transfer difference spectroscopy (STD), Carr-Purcell-Meiboom-Gill (CPMG), ¹⁹ F and [¹ H,¹⁵N]-HSQC NMR experiments, with an estimated KD of 1.8 mM by surface plasmon resonance (SPR). Further predictions by SiteMap and FTMap revealed that the iNOS binding site of SPSB2 is less druggable, explaining the poor outcome from the current FBDD campaign. Thus, other approaches to discover potent and specific inhibitors of the SPSB2-iNOS interaction were explored. Utilising an in silico structure-based drug design approach, a disulphide-bridged cyclic peptide Ac-c[CVDINNNC]-NH2 was designed and synthesised. It was found to bind to the iNOS binding site on SPSB2 with a KD of 4.4 nM, as shown by SPR, [¹H,¹⁵N]-HSQC and ¹⁹F NMR experiments, with approximately 70-fold improvement in affinity, compared to the linear peptide DINNN (KD ≈ 318 nM). An in vitro assay on macrophage cell lysates further showed a complete inhibition of SPSB2-iNOS interactions by the cyclic peptide. In addition, the solution structure of the cyclic peptide was found to closely match that of the crystal structure of SPSB2-bound linear peptide DINNN with a backbone RMSD of 1.21 A. The designed peptide was also found to be stable against pepsin, trypsin and α-chymotrypsin, and in human plasma. The disulphide-bridged cyclic peptide, however, is reductively labile. To generate redox-stable inhibitors of SPSB2-iNOS interaction that would retain activity in the cell cytoplasm, two cyclic peptide analogues, one containing a thioether bridge (CP1) and the other a lactam bridge (CP2), as well as four cyclic peptidomimetics (M1-M4), incorporating organic moieties as cyclisation linkers, were generated. All analogues were able to bind to the iNOS binding site of SPSB2, with five of the six analogues binding with stronger affinities (3-15 fold) than the linear peptide DINNN (CP1, KD 31 nM; CP2, KD 21 nM; M1, KD 29 nM; M2, KD 99 nM; M3, KD 54 nM; M4, KD 465 nM), as determined by 19F NMR and SPR, respectively. All analogues were able to compete with full-length iNOS for binding to SPSB2 in macrophage cell lysates. As CP2 is the most potent redox-stable analogue, with more sites for derivatisation and is easier to synthesise compared to other analogues, CP2 is being used as the template to generate analogues for macrophage-targeted delivery studies. Binding studies by SPR and 19F NMR revealed an approximately 5-fold improvement in binding affinity of rhodamine B isothiocyanate (RBITC)-conjugated cyclic peptide analogue CP4 to the iNOS binding site of SPSB2 (KD ≈ 4 nM), while the oligohistidine-conjugated analogues CP5 and CP6 showed a modest 1-2 fold drop in their binding affinities to SPSB2 compared to CP2 (KD ≈ 33-51 nM). Imaging studies of mannose or GalNAc glycopolymer-conjugated analogues of CP4 (i.e. CP7 and CP8, respectively) by confocal laser scanning microscopy revealed that both analogues were taken up by bone marrow-derived macrophages but not HEK 293 cells. On the other hand, oligohistidine-conjugated analogue CP6 showed signs of endosomal escape after 9 h of incubation, although coincident signs of cell necrosis or apoptosis were observed in some of these macrophages. In summary, several potent and stable cyclic peptide and peptidomimetic inhibitors of SPSB2-iNOS interactions were identified in this study. One of these analogues was able to be derivatised for macrophage-targeted delivery studies without negatively affecting their binding to SPSB2. This study also showed that both mannose and GalNAc glycopolymers are viable choices for macrophage-targeted delivery although more work to improve the endosomal escape of this new class of cyclic peptide inhibitors of SPSB2-iNOS interaction to the cytoplasm of macrophages without causing cell injury is needed.