We have recently described reduced superficial nephron glomerular filtration rate (SNGFR) in chloride-depletion alkalosis (CDA) without volume depletion. To elucidate the mechanism of this phenomenon, we studied three degrees of increasing severity of CDA (groups CDA-1, 2, and 3) produced by one or two peritoneal dialyses against 0.15 M NaHCO3 and electrolyte infusions of different Cl and HCO3 content in Sprague-Dawley rats; control rats (CON) were dialyzed against and infused with Ringers-HCO3. Extracellular fluid (ECF) volume was assessed by blood pressure, hematocrit, plasma protein concentration, and 125I-albumin space; none of these variables differed among the four groups. Micropuncture of the latest proximal and earliest distal convolutions was carried out. As CDA intensified from CON to CDA-3 (plasma tCO2 25 +/- 1 to 43 +/- 1 meq/L; P less than 0.01), distally determined SNGFR declined progressively (40.9 +/- 1.7 to 28.3 +/- 1.8 nl/min; P less than 0.01), while in early distal tubule fluid, flow rate (8.6 +/- 0.7 to 3.4 +/- 0.6 nl/min) and Cl concentration (36 +/- 2 to 19 +/- 3 meq/L) decreased and osmolality (110 +/- 5 to 208 +/- 12 mosmol/kg) increased (P less than 0.01), and, in the loop segment, Cl reabsorption decreased progressively (2,009 +/- 112 to 765 +/- 128 peq/min; P less than 0.01). In early distal tubule fluid, Cl concentration correlated positively and osmolality negatively with distally determined SNGFR (P less than 0.05). Proximally determined SNGFRs did not differ among the four groups. Proximal tubule stop-flow pressure responses to increasing rates of orthograde perfusion of the loop segment from 0 to 40 nl/min did not differ between groups CON and CDA-2. We interpret these data to show that reductions in SNGFR in CDA in the rat can occur by tubuloglomerular feedback (TGF) in the absence of differences in ECF volume or of alterations in TGF sensitivity during metabolic alkalosis. Of the proposed signals for TGF sensed by the macula densa, distal tubule fluid osmolality or some related variable is the signal most compatible with our data.
J H Galla, D N Bonduris, P W Sanders, R G Luke