Skip to main content
. 2022 Mar 24;11(7):1102. doi: 10.3390/cells11071102

Table 3.

Toxicity prediction outcomes as determined by MTS and NRU assays on WJSCs in two- (2D) and three-dimensional (3D) culture environments.

RC Rat-Only Weight Regression
WJSC (mg/kg)		 RC Rat-Only Millimole Regression
WJSC (mg/kg)
Chemicals Hazard Category RC Rodent (mg/kg) MTS NRU MTS MTS NRU MTS
2D 3D 2D 3D
Mercury(II) chloride 1 1 308 U 233 U 302 U 343 U 246 U 335 U
Cycloheximide 1 2 38 U 15 U - 29 U 10 U -
Sodium
arsenite 		2 41 139 U 126 U 139 U 88 U 78 U 89 U
Sodium dichromate dihydrate 2 50 72 U 32 111 U 65 U 25 108 U
Cadmium(II) chloride 3 88 164 (105) * 356 U - 135 (77) * 326 U -
Sodium
fluoride 		3 180 419 U (274) * 379 U 361 U 173 (105) * 154 145
Propranolol hydrochloride 4 470 488 500 575 619 636 751
Atropine sulfate monohydrate 4 639 1593 1373 - 4033 U 2416 U -
Potassium chloride 5 2602 2976 2981 2575 2411 3386 2033
Trichloroacetic acid 5 4999 1637 O 1344 O - 1849 O 1466 O -
Sodium hypochlorite 6 10,328 1140 O 945 O 1110 O 777 O 622 O 753 O
Glycerol 6 12,691 6401 6343 - 6702 6301 -
Correct prediction
score (%) **		 N = 12 5/12 (41.7)
[6/12 (50)] *		 5/12 (41.7) - 5/12 (41.7)
[5/12 (41.7)] *		 5/12 (41.7) -
N = 7 2/7 (29)
3/7 (42.9)		 - 2/7 (29) 3/7 (42.9) - 3/7 (42.9)
% Prediction improvement N = 12 +9/+17 +9/+17 - +9/+0 +9/+0 -

U = underprediction (underestimation of toxicity); O = overprediction (overestimation of toxicity). Chemicals in bold = correct prediction of toxicity class by both regression types (weight/millimole). The 7 chemicals in italics are those selected for confined group analysis based on N = 7 chemicals (one per hazard class; two from class 2) in 2D or 3D culture conditions. * Bracketed data represent predicted LD50 values and correct prediction scores resulting from in vitro (2D or 3D) toxicity evaluation in serum-free conditions, using the MTS assay. ** Prediction scores for BALB/c 3T3 and NHK cells (ICCVAM validation study) were 33.3/25% and 33.3/41.7% for the RC-rat only weight and millimole regressions, respectively; % prediction improvement was compared to the available in silico data from the NRU assay for BALB/c 3T3 and NHK cells, respectively.