Fibrosis is at the core of the high morbidity and mortality rates associated with the complications of diabetes and obesity, including diabetic nephropathy (DN), without any US Food and Drug Administration-approved drugs with this specific target. We recently provided the first evidence that the matricellular protein CCN3 (official symbol NOV) functions in a reciprocal manner, acting on the profibrotic family member CCN2 to inhibit fibrosis in a mesangial cell model of DN. Herein, we used the BT/BR ob/ob mouse as a best model of human obesity and DN progression to determine whether recombinant human CCN3 could be used therapeutically, and the mechanisms involved. Eight weeks of thrice-weekly i.p. injections (0.604 and 6.04 μg/kg of recombinant human CCN3) beginning in early-stage DN completely blocked and/or reversed the up-regulation of mRNA expression of kidney cortex fibrosis genes (CCN2, Col1a2, TGF-β1, and PAI-1) seen in placebo-treated diabetic mice. The treatment completely blocked glomerular fibrosis, as determined by altered mesangial expansion and deposition of laminin. Furthermore, it protected against, or reversed, podocyte loss and kidney function reduction (rise in plasma creatinine concentration); albuminuria was also greatly reduced. This study demonstrates the potential efficacy of recombinant human CCN3 treatment in DN and points to mechanisms operating at multiple levels or pathways, upstream (eg, protecting against cell injury) and downstream (eg, regulating CCN2 activity and extracellular matrix metabolism).