elsewhere (35, 36).Probes and in Situ Hybridization. RNA probes were transcribed from pGEM-1 transcription plasmid (Promega) that contained a 110-bp sequence of mouse SAA1 cDNA (p125) (37). This nucleotide sequence encompasses a domain coding for amino residues 30-66 that is highly conserved among various species and is 81% homologous with human SAA1 and SAA2 mRNAs and is 71% homologous with human apoSAA4 mRNA (refs. 4 3186The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. §1734 solely to indicate this fact.

Serum amyloid A (SAA) is a major acute-phase reactant and apoprotein of high density lipoprotein (HDL). SAA is encoded by a family of three active genes. We examined hepatic expression and searched for extrahepatic expression of the three SAA mRNAs after injection with casein or LPS. Studies using an SAA cDNA, which detects all three SAA mRNAs, revealed that after casein injection liver SAA mRNA was elevated approximately 1,000-fold. Adrenal gland expressed SAA mRNA at a low level (0.5% of hepatic level), and was the only extrahepatic tissue with elevated SAA mRNA after casein injection. The small intestine, primarily the ileum, and the large intestine of unstimulated control animals contained 5- and 15-fold higher SAA mRNA levels than control liver. LPS also elevated liver SAA mRNA approximately 1,000-fold. However, in contrast to casein injection, every extrahepatic tissue examined expressed SAA mRNA. Lung and kidney contained 2-5% and large intestine contained nearly 10% of SAA mRNA levels found in liver RNA. SAA mRNA levels were lower in the remaining tissues and ranged from 0.1% in the brain and pancreas to 1.0% in the small intestine, with the ileum containing 50-fold more than the duodenum. Analysis of liver with SAA1, SAA2, and SAA3 mRNA-specific oligonucleotide probes revealed that SAA1 and SAA2 mRNA were elevated approximately 50-fold higher than SAA3 mRNA after casein administration. LPS, however, induced all three SAA mRNAs equally. In extrahepatic tissues, SAA1, SAA2, and SAA3 mRNAs were expressed differentially and can be grouped into three general classes: tissues expressing all three genes, tissues expressing SAA1 and SAA3, and tissues expressing predominantly or only SAA3.

The serum amyloid A (SAA) proteins make up a multigene family of apolipoproteins associated with high density lipoproteins. They are of ancient origin; the filnding of a highly homologous protein in mammals and ducks indicates that SAAs have been in existence for at least 300 million years. The interspecies similarity among the SAAs makes the mouse, in which they have been most thoroughly studied, a reasonable model to use for defining the function(s) of this family of proteins in humans. Originally it was observed that the SAA proteins were made in the liver and represented a set ofproteins belonging to acute-phase reactants. SAA3 is a unique member of the SAA multigene family in mice in that its mRNA is also expressed in extrahepatic tissues by a variety of cell types, mainly macrophages and adipocytes. To date, nothing has been reported regarding the fate or function of the SAA3 translation product. To identify the SAA3 protein, we developed SAA3-specific antibodies by immunizing rabbits against a portion of SAA3 protein synthesized in a bacterial fusion protein expression system. Electroimmunoblot analysis of serum and lipoprotein fractions of it showed SAA3 to be associated with high density lipoproteins of mice treated with lipopolysaccharide. Furthermore, a continuous mouse macrophage cell line (J-774.1), when exposed to lipopolysaccharide, expressed SAA3 mRNA in a dose-dependent manner and secreted SAA3 protein. The expression and secretion of SAA3 by macrophages stimulated with lipopolysaccharide suggest a role for this SAA in local responses to injury and inflammation.

Inflammatory pathways are central mechanisms in diabetic kidney disease (DKD). Serum amyloid A (SAA) is increased by chronic inflammation, but SAA has not been previously evaluated as a potential DKD mediator. The aims of this study were to determine whether SAA is increased in human DKD and corresponding mouse models and to assess effects of SAA on podocyte inflammatory responses. SAA was increased in the plasma of people with DKD characterized by overt proteinuria and inversely correlated with estimated glomerular filtration rate (creatinine-based CKD-EPI). SAA was also elevated in plasma of diabetic mouse models including type 1 diabetes (streptozotocin/C57BL/6) and type 2 diabetes (BTBR-ob/ob). SAA mRNA (Nephromine) was increased in human DKD compared with non-diabetic and/or glomerular disease controls (glomerular fold change 1.5, P=0.017; tubulointerstitium fold change 1.4, P=0.021). The kidneys of both diabetic mouse models also demonstrated increased SAA mRNA (quantitative real-time PCR) expression compared with non-diabetic controls (type 1 diabetes fold change 2.9; type 2 diabetes fold change 42.5, P=0.009; interaction by model P=0.57). Humans with DKD and the diabetic mouse models exhibited extensive SAA protein deposition in the glomeruli and tubulointerstitium in similar patterns by immunohistochemistry. SAA localized within podocytes of diabetic mice. Podocytes exposed to advanced glycation end products, metabolic mediators of inflammation in diabetes, increased expression of SAA mRNA (fold change 15.3, P=0.004) and protein (fold change 38.4, P=0.014). Podocytes exposed to exogenous SAA increased NF-κB activity, and pathway array analysis revealed upregulation of mRNA for NF-κB-dependent targets comprising numerous inflammatory mediators, including SAA itself (fold change 17.0, P=0.006). Inhibition of NF-κB reduced these pro-inflammatory responses. In conclusion, SAA is increased in the blood and produced in the kidneys of people with DKD and corresponding diabetic mouse models. Podocytes are likely to be key responder cells to SAA-induced inflammation in the diabetic kidney. SAA is a compelling candidate for DKD therapeutic and biomarker discovery.

Thrombospondin 1 has been shown to be linked to PDGF-mediated mesangial cell proliferation and migration in vitro, but little is known regarding its expression or regulation in glomerular disease. Experimental mesangial proliferative nephritis was induced in rats by injection of anti-Thy1 antibody. Mesangial cell proliferation was associated with de novo expression of thrombospondin 1 mRNA (detected by Northern blot and in situ hybridization) and protein (by Western blot and immunostaining). Although some thrombospondin 1 was expressed by platelets and macrophages, double labeling showed that most thrombospondin 1 mRNA and protein were expressed by proliferating alpha-actin-positive mesangial cells. Thrombospondin 1 expression in anti-Thy1 nephritis was complement-dependent and could be reduced by treatment with anti-PDGF or anti-bFGF antibodies. Thrombospondin 1 could also be induced in normal rats by infusion of PDGF and in rats which were primed with low dose anti-Thy1 antibody by infusion of PDGF of bFGF. Thus, this study demonstrates that proliferating mesangial cells express thrombospondin 1 de novo in disease and that thrombospondin 1 expression in vivo is regulated by PDGF and bFGF.