*3.4. Sulfated Quinazolinones Do Not Inhibit MMP-9*

We next turned our attention to MMP-targeting selectivity studies. Considering that MMP-9 is the most closely related metalloenzyme to MMP-8, and as it belongs to the "intermediate" S<sup>1</sup> ′ pocket category like MMP-8, we screened sulfated benzofurans and sulfated quinazolinones against MMP-9. NSGM **26** also inhibited MMP-9 reasonably well (Figure 6). This was not too unusual, because its inhibition of MMP-8 was predicted to arise from coordination with the catalytic Zn2<sup>+</sup> and histidine triad (see Figure 5B), which is known to be the origin of a lack of selectivity associated with many inhibitors of MMPs. Another sulfated benzofuran, NSGM **29**, which did not inhibit MMP-8, also showed >80% inhibition of MMP-9. Thus, the future use of sulfated benzofurans as MMP-8 hits is likely to be beset with such non-selectivity concerns.

**Figure 6.** Screening of NSGM **15**–**42** at 100 µM against MMP-9. Experiments were performed using a commercially available assay kit (Abcam, Cambridge, U.K.) in at least duplicates. Error bars represent ± 1 SD.

In contrast, the sulfated quinazolinones did not inhibit MMP-9 significantly (Figure 6), thus displaying >10-fold selectivity for MMP-8 over MMP-9. Considering that the predicted mode of binding of sulfated quinazolinones does not involve a strong engagement of the catalytic Zn2<sup>+</sup> or the histidine triad (see Figure 5C), the observation of a lack of MMP-9 activity is encouraging. The sulfate group of **38**, inserted in the S<sup>1</sup> ′ pocket, forms hydrogen and ionic bond interactions with A220 and R222, respectively, of MMP-8. At the corresponding positions in MMP-9, there is neither an arginine nor an alanine, which supports the observed lack of inhibition. In fact, the arginine at 222 is found in only one other MMP (MMP-28), and the alanine at 220 is not present in any other MMPs [45]. Thus, it is possible that NSGM **38** is an excellent hit with a high potential for selectivity of MMP-8. This hypothesis will need to be further experimentally validated through extended selectivity studies.

Overall, our "hit" NSGMs are the first group of GAG-related inhibitors of MMP-8. Given the urgent need and lack of efforts directed towards developing MMP-8 inhibitors, these NSGMs offer a unique starting point for further structure-guided hit-to-lead optimization studies. Two major optimization routes could be considered, namely: (1) Our modeling insights convey that the use of a biphenyl substituent, instead of a phenyl substituent, on the sulfated quinazolinone would lead to a greater reach of a potential inhibitor inside the S<sup>1</sup> ′ pocket. This should enhance both the potency and selectivity. In fact, a peptide-hydroxamate containing a biphenyl substituent was previously shown to inhibit MMP-8 with a low nanomolar potency by occupying the S<sup>1</sup> ′ pocket [49]. However, because of the strong Zn+<sup>2</sup> chelating nature, metabolic instability, and toxicity associated

with hydroxamates, these molecules failed in the clinical trials [3,16]. (2) Increasing the polarity of the linker by changing it from an alkyl-based chain to an ethylene glycol-based chain would likely result in polar interactions with the residues lining the S<sup>2</sup> and S<sup>3</sup> pockets. This should also enhance the potency and perhaps selectivity.
