As discussed on the
Boldenone-mediated hepatorenal impairments... thread, I believe that the
direct neurotoxicity of AAS is AR-mediated (and that estrogens tend to have an opposite effect). More significant, however, are the
indirect effects, e.g., as indicated by uncontrolled hypertension (high BP), as the most significant factor contributing to nephrotoxicity. Given the cardiac maladaptations induced by AAS via multiple pathways, there are several pathways (direct in nephrons and indirect in, e.g., vascular endothelial cells) by which AAS induce nephrotoxicity.
It's mostly folly to try to attempt to predict with confidence where on a continuum Dbol might fall with respect to nephrotoxicity, but it's fair to say it's less nephrotoxic than tren per mg (because of its aromatization and reduced AR potency and tren's even greater haematopoietic effects).
Since Dbol does tend to increase blood volume and therefore stroke volume/index (i.e., 15 mg yields a 20% increase in stroke volume) reflecting an increase to systolic BP, its profound haemodynamic effects are similarly supported by more mechanistic evidence that it's the most potent haematopoietic (increase to HCT/Hb) agent commercially available (for therapeutic use, legitimately) in humans (per mg). Though I think virtually everyone will agree tren is more potent at this, certainly at doses >= 350 mg/wk (a dose that was regarded as absurd [pros commonly used 76 - 152 mg/wk hexahydrobenzylcarbonate at most] until Bill Roberts proposed it as the recommended initial dose).