Figure 4S. ADMET profile comparison of lead compounds

Predicted absorption, distribution, and toxicity parameters (admetSAR 3.0) for the lead spiro compounds KB-282 and KB-281 relative to the hydroxamate reference octanediamide and other HDAC8 reference inhibitors.

(A) Key ADMET parameters (higher = more favourable)

(B) Absorption and distribution

Mean Caco-2 permeability of the leads (0.650) is about 40.6× that of the hydroxamate reference (0.016).

KB-282 shows the lowest predicted hERG inhibition and mutagenicity; its DILI signal is higher than the hydroxamate and warrants monitoring.

ADMET data summary

Compound	Caco-2	HIA	BBB	PPB^a	VDss	hERG (1 µM)	DILI	Hepatotox.	Carcinog.	Mutagen.	DDI score^b
KB-282	0.619	0.835	0.599	100.7	−0.883	0.017	0.737	0.404	0.357	0.136	1.82
KB-281	0.681	0.972	0.735	105.7	−0.335	0.026	0.826	0.505	0.463	0.502	3.16
Octanediamide	0.016	0.775	0.215	90.4	−0.611	0.027	0.499	0.711	0.524	0.576	2.77
NCC149	0.453	0.982	0.884	93.0	−0.030	0.013	0.595	0.385	0.560	0.569	4.34
PCI-34051	0.122	0.912	0.661	86.5	−0.281	0.026	0.478	0.416	0.526	0.366	3.41
Entinostat	0.286	0.963	0.721	81.5	−0.081	0.019	0.507	0.482	0.695	0.784	3.81

ADMET observations

Oral absorption: both leads show higher predicted Caco-2 permeability (KB-282 0.619; KB-281 0.681) than the hydroxamate reference (0.016); KB-281 also shows high intestinal absorption (HIA 0.972).
Cardiac and genotoxic risk: KB-282 has the lowest predicted hERG inhibition (0.017) and mutagenicity (0.136) in this comparison, and lower predicted hepatotoxicity (0.404) than octanediamide (0.711).
DILI trade-off: KB-282's predicted drug-induced liver injury probability (0.737) exceeds that of the hydroxamate reference (0.499) — a signal to monitor in subsequent experimental work.
Drug–drug interactions: KB-282 has the lowest cumulative CYP-inhibition (DDI) score (1.82), about 53% below the mean of NCC149 and PCI-34051 (3.88).
Drug-likeness: KB-282 and KB-281 satisfy Lipinski's rule of five with no violations (Supplementary S19); KB-281 (477.4 Da) approaches the molecular-weight threshold.
Overall: the lower hERG, mutagenicity, and DDI signals support KB-282 as the lead candidate, while the elevated DILI probability is the principal liability to track.

Sources. Absorption/distribution: Supplementary S7 and S22. Toxicity: Supplementary S10 and S23. CYP-inhibition / DDI: Supplementary S9 and S24. Drug-likeness: Supplementary S19. Interpretation. For hERG, DILI, hepatotoxicity, carcinogenicity, and mutagenicity, lower values indicate lower predicted risk. In panel A the three toxicity axes are plotted as safety scores (1 − risk) so that, for every axis, larger area means a more favourable profile. ^aPPB values are raw admetSAR 3.0 output; values above 100 reflect model extrapolation beyond physiological bounds and should be read as high-binding predictions, not literal percentages. ^bDDI score = sum of six CYP-inhibition probabilities (CYP1A2, 2B6, 2C9, 2C19, 3A4, 2D6); lower indicates lower predicted interaction potential.