Superoxide (O(2)(-)) enhances tubuloglomerular feedback by scavenging nitric oxide at the macula densa. However, the singling pathway of O(2)(-) production in the macula densa is not known. We hypothesized that the increase in tubular NaCl concentration that initiates tubuloglomerular feedback induces O(2)(-) production by the macula densa via NAD(P)H oxidase, which is activated by macula densa depolarization. We isolated and microperfused the thick ascending limb of the loop of Henle and attached macula densa in rabbits. A fluorescent dye, dihydroethidium, was used to detect O(2)(-) production at the macula densa. When luminal NaCl was switched from 10 to 80 mM, a situation of initiating maximum tubuloglomerular feedback response, O(2)(-) production significantly increased. To make sure that the shifts in the oxyethidium/dihydroethidium ratio were due to changes in O(2)(-), we used tempol (10(-4) M), a stable membrane-permeant superoxide dismutase mimetic. With tempol present, when we switched from 10 to 80 mM NaCl, the increase in oxyethidium/dihydroethidium ratio was blocked. To determine the source of O(2)(-), we used the NAD(P)H oxidase inhibitor apocynin. When luminal NaCl was switched from 10 to 80 mM in the presence of apocynin, O(2)(-) production was inhibited by 80%. To see whether the effect of increasing luminal NaCl involves Na-K-2Cl cotransporters, we inhibited them with furosemide. When luminal NaCl was switched from 10 to 80 mM in the presence of furosemide, O(2)(-) production was blocked. To test whether depolarization of the macula densa induces O(2)(-) production, we artificially induced depolarization by adding valinomycin (10(-6) M) and 25 mM KCl to the luminal perfusate. Depolarization alone significantly increases O(2)(-) production. We conclude that increasing luminal NaCl induces O(2)(-) production during tubuloglomerular feedback. O(2)(-) generated by the macula densa is primarily derived from NAD(P)H oxidase and is induced by depolarization.