Chronic Intra-Renal Insulin Replacement Reverses Diabetes-Induced Natriuresis and Diuresis

Abstract

We recently showed that sustained natriuresis in Type I diabetic dogs was due to the decrease in insulin rather than the hyperglycemia alone. The sodium retaining action of insulin appeared to require hyperglycemia and it completely reversed the diabetic natriuresis and diuresis. This study tested whether the sodium retaining effect was due to direct intrarenal actions of insulin. Alloxan-treated dogs (D; n=7) were maintained normoglycemic using 24hr/day i.v. insulin replacement. After control measurements, i.v. insulin was decreased to begin a 6-day diabetic period. Blood glucose increased from 84±6 to an average of 428 mg/dL on days 5 and 6, sodium excretion increased from 74±8 to 98±7 mEq/day over the 6 days, and urine volume increased from 1645±83 to 2198±170 mL/day. Other dogs (DIr, n=7) were subjected to the same diabetic regimen, but in addition, insulin was infused continuously into the renal artery at 0.3 mU/kg/min during the 6-day period. This did not affect plasma insulin. Blood glucose increased from 94±10 to an average of 380 mg/dL on days 5 and 6, but sodium excretion averaged 76±5 and 69±8 mEq/day during control and diabetes, respectively. The diuresis also was prevented. GFR increased only in Dir dogs, and there was no change in MAP in either group. This intra-renal insulin infusion had no effect on sodium or volume excretion in normal dogs. Intra-renal insulin replacement in diabetic dogs caused a sustained increase in tubular reabsorption that completely reversed diabetic natriuresis. Insulin+glucose may work to prevent salt wasting in uncontrolled Type II diabetes.

Introduction

The chronic renal and cardiovascular actions of insulin are an enigma. Hyperinsulinemia correlates strongly with hypertension in metabolic syndrome and diabetes, ^1-4 but there is no consensus on whether there is a cause and effect relationship. On one hand, the possibility that hyperinsulinemia could cause hypertension is supported by acute insulin infusion studies that show stimulation of sodium reabsorption by insulin in animals ^5-7 and humans. ^8-11 Chronic sugar feeding studies in rats report hypertension linked to hyperinsulinemia and renal insulin action. ^{12, 13} In addition, chronic insulin infusion causes renal sodium retention and increases blood pressure in rats. ^{14, 15} Thus, insulin-mediated renal sodium retention is hypothesized to be at least one component that explains hypertension in hyperinsulinemic conditions such as metabolic syndrome or Type II diabetes. ^{3, 16, 17}

On the other hand, there is no direct experimental evidence outside rat models ^{14, 15} for chronic sodium retaining or hypertensive actions of insulin. In fact, numerous chronic insulin infusion studies in dogs ^18-20 have refuted the insulin hypotheses by failing to show hypertension or direct sodium retention. Chronic insulin infusion into the renal artery of normal dogs also failed to cause sustained sodium retention or hypertension. ²¹ When considered along with the absence of hypertension in patients with insulinoma, ^{22, 23} it is understandable that other mechanisms such as oxidative stress and aldosterone have been proposed to explain the link between insulin resistance, renal sodium handling, and hypertension in metabolic syndrome and diabetes. ^{2, 24}

Therefore, hypotheses that insulin can cause sustained renal sodium retention and hypertension lack chronic experimental support outside of rat models. However, the failure of hyperinsulinemia to cause sustained sodium retention or hypertension in dogs and insulinoma patients does not necessarily warrant ignoring the supportive data and abandoning a potential sodium-retaining effect of insulin. This is because there have been no chronic experiments testing the renal actions of insulin under conditions that actually represent the metabolic syndrome or diabetic milieu. In other words, the euglycemic hyperinsulinemic approach used to isolate the effect of insulin in most insulin infusion experiments, including previous dog studies, ^18-20 does not represent the insulin-glucose relationship either in Type I diabetes (low insulin-high glucose) or metabolic syndrome/Type II diabetes (high insulin-high glucose).

We addressed that recently by testing whether the natriuresis caused by induction of Type I diabetes was due, at least in part, to the loss of insulin and its sodium-retaining action rather than being due solely to the osmotic diuretic effect of hyperglycemia. ²⁵ We reported that the sustained natriuresis in Type I diabetic dogs (∼ 400 mg/dl for 6 days) was obliterated if the decrease in circulating insulin was prevented. That was the first evidence in a non-rodent model for a sustained sodium retaining action of insulin. The results also suggested that insulin only stimulates renal tubular sodium reabsorption in the presence of hyperglycemia. However, all manipulations of insulin in that study were systemic, and there is evidence that central actions of insulin can increase sympathetic nervous system activity. ^{26, 27} Therefore, in the present study we infused insulin 24 hr/day for 6 days into the renal artery of diabetic dogs to test the hypothesis that insulin, only in the presence of hyperglycemia, acts directly on the kidneys to stimulate sodium reabsorption chronically.

Methods

Studies were conducted in conditioned male mongrel dogs weighing approximately 25 kg and all experimental protocols were approved by the Institutional Animal Care and Use Committee at Georgia Health Sciences University. Dogs were instrumented as follows during a single, sterile surgical procedure under isoflurane anesthesia: Via a left flank incision, a flow probe (3PSB; Transonic, Ithaca, NY) was affixed around the left renal artery and a tygon catheter was inserted in the left renal artery using the method of Herd and Barger. ²⁸ The right kidney was removed via a right flank incision. A Data Sciences (DSI St. Paul, MN) TA11PA-D70 blood pressure unit was implanted in the right femoral artery, and standard fluid-filled Tygon catheters were implanted in the right femoral vein and in the left femoral artery and vein. The catheters and probe cable were tunneled subcutaneously to the scapular region and exteriorized, and the dogs were fitted with a polypropylene jacket equipped with a pocket to protect and hold the catheters and flow probe cable.

After one week of recovery, dogs were placed in individual metabolic cages and connected to the infusion pumps and flow meter. Briefly, dogs are fitted with a 7″×10″ curved, padded piece of plexiglass under their jacket. It serves as a platform to connect the catheters and flow probe cable to the infusion lines and flow meter cable. It also is the base for the flexible stainless steel conduit that connects (i.e. tethers) the dog to the customized electrical/hydraulic swivel mounted at the top-center of the cage. The infusion lines and flow meter cable run through this conduit. The swivel connects to the infusion pumps and flow meter. This system allows continuous intravenous and intrarenal infusion and electrical connectivity while the dogs have completely unrestricted, 360-degree freedom of movement in the cages 24 h/day. Approximately 2 weeks were allowed for the dogs to acclimate to the metabolic cages and be trained to lie quietly for blood sampling.

Salt and Water Balance and Glucose Control

Immediately after placement in the cages, 24 hr/day infusion of 0.9% saline (approximately 475 ml/day) was begun in all dogs. Combined with feeding a low-sodium diet (Hills H/D; 3, 13 oz cans per dog per day), this enabled maintenance of constant sodium intake throughout the study. In addition, all dogs received approximately 975 ml of sterile water vehicle per day iv., and 48 ml of heparinized (1%) sterile water vehicle per day through the renal artery catheter throughout the study. Drinking water was available ad libitum.

During this 2-week acclimation period, dogs were divided randomly into two groups: Type I diabetes (D, n=7) and Type I diabetes with chronic intra-renal insulin infusion (Dir, n=7). Alloxan (50 mg/kg) was given to all dogs to decrease endogenous insulin. This allowed us to control plasma insulin chronically throughout the experiment. To avoid potential renal complications from concentrating alloxan, the iv. saline infusion was increased to 1000 ml/day for 2 days preceding alloxan administration, and each dog was given mannitol (50 ml of a 250 mg/ml solution, iv.) 30 minutes immediately preceding alloxan administration. Every day after the alloxan adminstration, fasting blood glucose was measured in all dogs at 0800 hours. Blood glucose reached diabetic levels 1-3 days after alloxan, at which point dogs were placed on insulin replacement therapy. Regular insulin was infused iv. 24 hr/day via the saline bag (insulin replacement; Insulin Rx), and the dose was adjusted daily in each dog based on the daily glucose measurement. By the end of the 2-week acclimation period, all dogs reached stable normal daily fasting blood glucose and stable daily Insulin Rx dose.

Experimental Protocol

After the acclimation period, control period measurements were begun. Control period conditions were alloxan-treated dogs that received Insulin Rx 24 hr/day to maintain normal fasting blood glucose. After control measurements, a 6-day period of Type I diabetes was induced in all dogs by decreasing the Insulin Rx dose. In the DIr dogs (n=7), insulin (0.3 mU/kg/min) was added to the 48 ml/day intra-renal sterile water vehicle infusion, whereas no insulin was added to the intra-renal infusion for the D dogs (n=7). After 6 days of diabetes, control conditions were resumed in both groups and normal glucose levels were restored for recovery period measurements.

Blood Sampling and Analytical Procedures

Fasting blood samples (21 hrs postprandial) were drawn during the control period, on diabetes days 2 and 5, and during the recovery period in all dogs in both groups. On the same days, glomerular filtration rate (GFR) was determined from the total plasma clearance of I¹²⁵-labeled iothalamate (Glofil; QOL Medical, Kirkland, WA) over a 3-hour period from 0800-1100 while the dogs rested quietly in their cages. Blood pressure and renal blood flow (RBF) measurement began at 1400 hours every day and continued through to 0800 hours the next morning. The blood pressure and flow signals were sampled for 10 seconds each minute at 100 Hz using the A.R.T. software from DSI.

Urine sodium, potassium, and lithium concentrations were determined by atomic absorption, plasma electrolytes were measured by ion-sensitive electrodes (MEDICA Easy Electrolytes, Bedford, MA), plasma protein concentration was measured by refractometry, blood glucose was measured with an Accu-Check meter (Roche, Indianapolis, IN), urine glucose was measured using a glucose assay kit (Sigma-Aldrich, St. Louis, MO), osmolality was measured by freezing point depression (Advanced Instruments, Norwood, MA), plasma insulin was measured using an EIA kit from ALPCO Diagnostics (Salem, NH), and plasma renin activity (PRA) was measured by radioimmunoassay (Diasorin, Stillwater, MN). Daily electrolyte and water balances were calculated as: intake - output - insensitive loss, where insensitive loss equaled average intake during the control period - average output during the control period. Therefore, cumulative balance equals zero on the last day of the control period.

Data from D and DIr dogs were analyzed with 2-factor repeated measures analysis of variance (ANOVA) using GraphPad Prism software. At p < 0.05 for the within-subjects F-test in the ANOVA, Dunnett's test was used to determine which experimental-period day differed from control. The control for time in this experimental design is the recovery period, and for variables with multiple recovery period days, an average recovery period value was determined and used in the ANOVA. At p < 0.05 for the between-subjects F-test in the ANOVA, bonferroni test was used to determine on which days the two groups differed. Statistical significance versus control was * p<0.05, and for the DG versus the D group ^† p<0.05. All data are expressed as mean ± SEM.

Results

Glucose and Insulin

Figure 1 shows blood glucose in diabetic dogs (D; top panel) and diabetic dogs with intra-renal insulin infusion (DIr; bottom panel). During the control period, plasma insulin averaged 4.3±1.1 and 4.1±0.5 uU/ml (Table 1) and blood glucose averaged 84±6 and 94±10 mg/dl (Figure 1) in the D and DIr dogs, respectively. During diabetes, the increase in blood glucose was significant and not different between groups, increasing similarly on day 1, tending to be lower in the DIr dogs on days 2-4, and averaging approximately 400 mg/dl in both groups on days 5 and 6 (Figure 1). Systemic plasma insulin concentration decreased significantly in both groups to approximately 2.5 uU/ml with no difference between groups (Table 1). All values returned towards control levels when Insulin Rx was resumed for the recovery period.

Figure 1.

Blood glucose in diabetic dogs (D, n=7, panel A) and diabetic dogs with intra-renal insulin infusion (DIr, n=7, panel B). C = control day, D = diabetes day, IR = intra-renal insulin diabetes day, and R = recovery. Insulin Rx = 24 hr/day iv. insulin replacement infusion. Data are mean ± SEM. * = p < 0.05 vs control period.

Table 1.

Plasma composition and extracellular fluid volume in diabetic dogs (D) and diabetic dogs with chronic intra-renal insulin infusion (DIr).

Variable	Group		Experimental Period
		Control	Day 2	Day 5	Recovery
P Insulin	D	4.3 ± 1.1	2.1 ± 0.5 *	2.7 ± 0.6	5.0 ± 1.0
	DIr	4.1 ± 0.5	2.8 ± 0.2	2.6 ± 0.3 *	4.1 ± 0.7
PRA	D	0.33 ± 0.13	0.12 ± 0.06	0.20 ± 0.05	0.83 ± 0.33
	DIr	0.39 ± 0.10	0.91 ± 0.30 *,†	0.68 ± 0.20 *	0.38 ± 0.10*
P Na+	D	144.0 ± 1.1	137.3 ± 0.5 *	135.2 ± 0.6 *	144.0 ± 1.0
	DIr	144.6 ± 0.4	140.1 ± 2.0 *	139.1 ± 1.0 *,†	143.7 ± 0.4
P K+	D	4.6 ± 0.1	5.5 ± 0.4	4.9 ± 0.1	4.5 ± 0.1
	DIr	4.7 ± 0.4	5.1 ± 1.8	4.7 ± 0.8	4.5 ± 0.1
P Osm	D	311 ± 2	323 ± 4 *	325 ± 6 *	307 ± 4
	DIr	310 ± 1	313 ± 3 †	315 ± 2 †	311 ± 4
Hct	D	42 ± 2	38 ± 2 *	36 ± 2*	39 ± 2
	DIr	43 ± 1	41 ± 1	42 ± 1	40 ± 1
P Protein	D	7.4 ± 0.3	8.0 ± 0.2	7.9 ± 0.3	7.4 ± 0.3
	DIr	7.2 ± 0.2	7.2 ± 0.2	7.2 ± 0.2	6.9 ± 0.1
ECFV	D	7839 ± 233	7738 ± 194	7107 ± 286 *	7537 ± 349 *
	DIr	7333 ± 300	7140 ± 115	7223 ± 278	6466 ± 53*

P _Insulin= plasma insulin (uU/ml), PRA = plasma renin activity (ng AngI/ml/hr), P _Na+= plasma sodium (mEq/L), P _K+= plasma potassium (mEq/L), P _Osm= plasma osmolality (mOsm/kg), Hct = hematocrit (%), P _Protein = plasma protein (g/dl), ECFV = extracellular fluid volume (ml). Data are mean ± sem,

^*p<0.05 vs. control,

^†p<0.05 vs. D group.

D group n = 7and DIr group n = 7.

Urine Sodium and Volume Excretion

Sodium intake averaged 88±1 and 85 ± 2 mEq/day for the D and Dir dogs, respectively. Urine sodium excretion (UNaV) increased in both groups on day 1 of diabetes (Figure 2). This is similar to the day 1 natriuresis we measured previously in Type I diabetic dogs and in diabetic dogs with circulating insulin maintained at control levels. ²⁵ The natriuresis was sustained over the 6-day period in the D group, with some waning near the end likely due to the decreasing blood pressure and withdrawal of pressure natriuresis. In the DIr dogs, however, UNaV decreased rapidly to levels not different from control (Figure 2) despite sustained hyperglycemia (Figure 1, bottom panel). Cumulative sodium balance averaged (-) 187±82 and (-) 25±16 mEq in the D and DIr dogs, respectively, after 6 days of diabetes. The different recovery-period UNaV response pattern between the two groups also was interesting. Whereas UNaV decreased rapidly in the D dogs at the start of the recovery period, there was an opposite response in the DIr dogs. These patterns are consistent with a sodium retaining response during diabetes in the DIr dogs and sodium losing response in the D dogs. The increase in fractional lithium reabsorption (Figure 3) during the diabetic period also supports a sodium retaining response in the DIr dogs. Moreover, the significant attenuation of urinary glucose excretion in the Dir dogs, considered together with lithium reabsorption, is consistent with stimulation of proximal sodium-glucose co-transport.

Figure 2.

Urinary sodium excretion in diabetic dogs (D, n=7, panel A) and diabetic dogs with intra-renal insulin infusion (DIr, n=7, panel B). C = control day, D = diabetes day, IR = intra-renal insulin diabetes day, and R = recovery. Insulin Rx = 24 hr/day iv. insulin replacement infusion. Data are mean ± SEM. * = p < 0.05 vs control period.

Figure 3.

Fractional lithium reabsorption in diabetic dogs (D, n=7, panel A) and diabetic dogs with intra-renal insulin infusion (DIr, n=7, panel B). C = control day, D = diabetes day, IR = intra-renal insulin diabetes day, and R = recovery. Insulin Rx = 24 hr/day iv. insulin replacement infusion. Data are mean ± SEM. * = p < 0.05 vs control period. ^† = p < 0.05 vs D group.

Urine volume (Figure 4) tracked with UNaV, and the elimination of diabetic diuresis by intra-renal insulin is consistent with the elimination of diabetic natriuresis in the DIr dogs. Extracellular fluid volume also decreased significantly only in the D dogs (Table 1), and there was a significantly greater decrease in cumulative water balance in D vs. DIr dogs (Table 2). The greater increase in urinary glucose excretion (Table 2) in the D vs. DIr dogs is consistent with their greater urine sodium and water losses.

Figure 4.

Urine volume in diabetic dogs (D, n=7, panel A) and diabetic dogs with intra-renal insulin infusion (DIr, n=7, panel B). C = control day, D = diabetes day, IR = intra-renal insulin diabetes day, and R = recovery. Insulin Rx = 24 hr/day iv. insulin replacement infusion. Data are mean ± SEM. * = p < 0.05 vs control period.

Table 2.

Renal clearance and excretion data in diabetic dogs (D) and diabetic dogs with chronic intra-renal insulin infusion (DIr).

Variable	Group		Experimental Period
		Control	Day 2	Day 5	Recovery
GFR	D	54 ± 2	53 ± 2	56 ± 1	58 ± 2
	DIr	53 ± 2	60 ± 1 *,†	63 ± 3 *,†	51 ± 1
U osm V	D	631 ± 114	1452 ± 249 *	1440 ± 251 *	929 ± 279
	DIr	671 ± 40	1133 ± 162 *,†	988 ± 167 *,†	803 ± 103
UK+V	D	47 ± 6	64 ± 8 *	47 ± 5	45 ± 7
	DIr	45 ± 4	61 ± 4 *	37 ± 8 *,†	48 ± 4
U Glucose V	D	ND	233 ± 53 *	208 ± 57*	ND
	DIr	ND	15 ± 3 *	11 ± 3 *,†	ND
C Osm	D	1.64 ± 0.10	3.62 ± 0.19 *	3.60 ± 0.19 *	2.32 ± 0.47
	DIr	1.50 ± 0.09	2.51 ± 0.36 *	1.93 ± 0.43	1.78 ± 0.31
C H2O	D	-0.50 ± 0.08	-2.11 ± 0.17*	-2.00 ± 0.22 *	-0.91 ± 0.43 *
	DIr	-0.38 ± 0.07	-1.30 ± 0.26 *,†	-0.94 ± 0.36 *,†	-0.56 ± 0.24
C Li	D	11.9 ± 2.2	13.3 ± 4.9	14.8 ± 4.7 *	10.0 ± 1.6
	DIr	16.3 ± 1.4	6.9± 1.6 *,†	6.8 ± 1.3 *,†	5.4 ± 1.1 †
C Na	D	0.33 ± 0.03	0.60 ± 0.04 *	0.46 ± 0.03 *	0.30 ± 0.05
	DIr	0.27 ± 0.02	0.39 ± 0.04 †	0.30 ± 0.03 †	0.35 ± 0.05
Bal H2O	D	0	-1075 ± 468 *	-857 ± 746 *	1098 ± 1282 *
	DIr	0	-475 ± 186 *,†	-468 ± 197 *,†	-616 ± 555 *,†

GFR = glomerular filtration rate (ml/min), U_osmV = osmolar excretion (mOsm/day), U_osmV= potassium excretion (mEq/day), U_GlucoseV = urine glucose excretion (g/day), C_osm= osmolar clearance (ml/min), C_H2O= free water clearance (ml/min), C_Li= lithium clearance (ml/min), C_Na= sodium clearance (ml/min), Bal_H2O= cumulative water balance. Data are mean ± sem,

^*p<0.05 vs. control;

^†p<0.05 vs. D group.

D group n = 7 and DIr group n = 7. n = 7 per group for U_GlucoseV.

Hemodynamics

Mean arterial pressure (MAP) did not change significantly in either group, but there were trends towards decreased MAP in the D dogs versus increased MAP in the DIr dogs (Figure 5). Renal blood flow (RBF) was measured continuously, 19 hrs/day with a Transonic flow probe, and there was no change in RBF in either group - remaining essentially “flat” in both groups - throughout all periods. Glomerular filtration rate (GFR) also did not change in the D group, but it increased significantly in the DIr group during diabetes (Table 2). Plasma renin activity (PRA) also increased only in the DIr group (Table 1).

Figure 5.

Mean arterial pressure in diabetic dogs (D, n=7, panel A) and diabetic dogs with intra-renal insulin infusion (DIr, n=7, panel B). C = control day, D = diabetes day, IR = intra-renal insulin diabetes day, and R = recovery. Insulin Rx = 24 hr/day iv. insulin replacement infusion. Data are mean ± SEM. * = p < 0.05 vs control period.

Intra-Renal Insulin in Normal Dogs

Our intra-renal insulin dose of 0.3 mU/kg/min is the same dose shown previously not to cause sustained sodium retention or hypertension during chronic intra-renal infusion in normal dogs. ²¹ We confirmed that in a separate group of 4 dogs surgically instrumented, housed, and maintained the same as the diabetic dogs. The only difference was that they were not given alloxan and thus were normoglycemic. Figure 6 shows that 6 days of intra-renal insulin infusion had no significant, sustained effect on UNaV. There was no change in MAP, and Table 3 shows that there were no changes in plasma insulin or blood glucose, consistent with previous results at this dose. ²¹

Figure 6.

Urinary sodium excretion in normal dogs with intra-renal insulin infusion (NI, n=4). C = control day, I = intra-renal insulin day, and R = recovery. Data are mean ± SEM.

Table 3.

Plasma insulin, glucose, and renin activity in Normal dogs with chronic intra-renal infusion of glucose (NG), insulin (NI), and glucose + insulin (NGI)

Variable	Group		Experimental Period
		Control	Day 2	Day 5	Recovery
P Insulin	NG	3.2 ± 0.6	3.1 ±0.6	3.2 ±0.3	4.7 ± 1.3
	NI	2.4 ±0.5	3.4 ± 0.6	3.3 ± 0.7	3.7 ± 1.0
	NGI	4.8 ± 1.4	4.3 ± 0.3	4.0 ± 0.5	5.0 ± 0.6
B Glucose	NG	94 ± 1	94 ± 2	102±6	94 ± 4
	NI	104 ± 6	92 ± 5	103 ± 7	91 ± 9
	NGI	93 ± 5	99 ± 3	102 ± 2	97 ± 3
PRA	NG	0.4 ± 0.2	0.3 ± 0.2	0.4 ± 0.2	0.1 ± 0.1
	NI	0.2 ± 0.1	0.3 ± 0.2	1.0 ± 0.9	0.1 ± 0.1
	NGI	0.1 ± 0.1	1.2 ± 0.2 *	2.5 ± 1.7 *	0.4 ± 0.2

P _Insulin= plasma insulin (uU/ml), B_Glucose= blood glucose (mg/dl), PRA = plasma renin activity (ng AngI/ml/hr). Data are mean ±SEM.

^*= p < 0.05 vs. Control.

Intra-Renal Insulin + Glucose in Normal Dogs

Our previous ²⁵ and current data show that insulin stimulates sodium reabsorption only in the presence of hyperglycemia. To determine whether adding glucose to the intra-renal insulin infusion could trigger a sodium-retaining action, we infused insulin (0.3 mU/kg/min) plus glucose (17 mg/min or 24 g/day) intra-renally in 7 dogs surgically instrumented, housed, and maintained the same as the diabetic dogs. This dose of glucose was calculated to raise intra-renal blood glucose by approximately 14 mg/dl, similar to the modest increase early in metabolic syndrome. Infusing glucose alone at this dose in 6 dogs had no effect on UNaV or any measured variable, similar to the effect of insulin alone. However, the combined infusion significantly decreased UNaV over the 6-day period, followed by rebound natriuresis during the recovery period (Figure 7). There was no change in MAP, and Table 3 shows that there were no changes in plasma insulin or blood glucose, but PRA increased significantly (Table 3).

Figure 7.

Urinary sodium excretion in normal dogs with intra-renal glucose + insulin infusion (NGI, n=7). C = control day, GI = intra-renal glucose + insulin day, and R = recovery. Data are mean ± SEM. * = p < 0.05 vs control period.

Discussion

The primary finding from this study is that intrarenal insulin replacement in Type I diabetic dogs completely reversed the natriuresis and diuresis caused by onset of hyperglycemia. The antinatriuretic action was sustained for the 6-day diabetic period. This replicates our previous finding ²⁵ that 6 days of hyperglycemia in the 400 mg/dl range did not cause sustained natriuresis or diuresis if plasma insulin was not allowed to decrease from normal levels. However, all experimental manipulation of insulin and glucose in that study were intravenous, which raised the possibility of contributions from systemic mechanisms such as the sympathetic nervous system. ^{26, 27} This study isolated the sustained sodium retaining effect of insulin to the kidney.

Our previous study ²⁵ had a complicated experimental design. Our premise was that the euglycemic, hyperinsulinemic approach used to isolate the effect of insulin in most insulin infusion experiments, including previous dog studies, ^18-20 does not represent the insulin-glucose relationship either in Type I diabetes (low insulin-high glucose) or metabolic syndrome/Type II diabetes (high insulin-high glucose). Therefore, euglycemic hyperinsulinemia could not test the effect of changes in the endogenous system as it related to either Type I or Type II diabetes. We addressed this problem with a chronic study in dogs that tested the renal effects of the decrease in insulin that causes Type I diabetes. ²⁵ We used chronic, 24 hr/day iv. insulin and glucose infusion in alloxan-treated dogs to create a 6-day condition of normal plasma insulin and 400 mg/dl hyperglycemia. The hyperglycemia was not different from the Type I diabetic control dogs; the only difference was that plasma insulin did not decrease from baseline levels. The primary finding was that the sustained natriuresis and diuresis measured in the Type I diabetic dogs was reversed completely in the dogs in which insulin was not allowed to decrease.

Three important mechanistic questions arose: 1) How did insulin cause antinatriuresis if the dogs were not hyperinsulinemic? 2) Is the antinatriuretic effect of insulin due to direct renal actions? 3) What are the renal mechanisms for the sodium retaining effect? The first question is important because the iv. insulin infusion in our previous study ²⁵ did not induce hyperinsulinemia, but simply maintained baseline insulin action in alloxan-treated dogs. Thus, the sodium-retaining effect occurred only when the dogs became overtly diabetic: Hyperglycemia in the presence of baseline insulin triggered a sodium-retaining action that was powerful enough to reverse completely the diabetic natriuresis and diuresis. This suggested that insulin + glucose, rather than isolated hyperinsulinemia, has a sustained sodium retaining action.

The present study confirmed that, because intra-renal insulin infusion had a sodium retaining effect only in hyperglycemic dogs. Hall et al. ²¹ used a very similar experimental model to show that chronic intra-renal insulin infusion (0.3 mU/kg/min) did not cause sustained sodium retention or hypertension in normal dogs. We confirmed that using the same intra-renal insulin dose in normal dogs (Figure 6). However, when that intra-renal insulin dose was administered in dogs with Type I diabetes, the diabetes-induced natriuresis and diuresis were reversed. We further explored this relationship by testing whether addition of glucose to the intra-renal insulin infusion in normal dogs would cause sustained sodium retention. Figure 7 shows that the combined intra-renal infusion of insulin and glucose caused mild but significant antinatriuresis, even though neither one affected sodium excretion when infused individually. Further studies will be needed to determine whether a longer infusion period and/or higher doses of glucose and/or insulin would cause greater sodium retention and possibly hypertension in normal dogs. Our previous ²⁵ and present results suggest that the antinatriuretic effect functions as a preventer of renal salt wasting in uncontrolled diabetes, rather than causing increased sodium balance.

In addition to confirming that insulin + glucose decreases sodium excretion, these results addressed our 2nd mechanistic question by isolating the effect of insulin to the kidney. The plasma insulin and glucose data show that there was no spillover into the systemic circulation either in normal or diabetic dogs. This is similar to findings of Hall et al., ²¹ who also showed that spillover did occur in normal dogs at a 2-fold higher intra-renal insulin dose of 0.6 mU/kg/min. Based on those data and our own data in normal dogs, we speculate that the 0.3 mU/kg/min intra-renal dose we used in the diabetic dogs, in which plasma insulin was decreased 50%, restored rather than increased renal plasma insulin levels. We cannot rule out whether renal plasma insulin increased slightly above normal, but the absence of spillover nonetheless indicates that insulin acted directly on the kidneys to cause antinatriuresis..

There are several intriguing possibilities to address our 3rd question regarding how insulin and glucose might act on the kidneys to stimulate sodium reabsorption. First, it is important to establish that the mechanism indeed was through tubular sodium reabsorption, because intrarenal insulin infusion in the diabetic dogs prevented the natriuresis, diuresis, and increased sodium clearance in the face of increased GFR and decreased renal vascular resistance. The proximal tubule in particular is implicated, because the increases in GFR, fractional lithium reabsorption, and PRA are consistent with the hypothesis that glucose stimulates increased proximal tubule sodium transport and causes withdrawal of the tubuloglomerular feedback signal at the macula densa. ^{29, 30} The marked attenuation of urinary glucose excretion in the Dir dogs, consistent with our previous study, ²⁵ provides further support by suggesting increased activity of the proximal tubule sodium glucose co-transporter (SGLT). A puzzling aspect of our results, however, is that this proximal sodium reabsorption mechanism has been proposed as an explanation for the increase in GFR in normal Type I diabetes. ²⁹ Indeed, we reported increased GFR, RBF, and PRA in chronically-instrumented Type I diabetic rats, and suggested that combined effects of increased proximal tubule sodium reabsorption and renal volume loss could play a role in those responses. ³¹ We expected similar renal changes in the D dogs, perhaps being amplified by adding intra-renal insulin in the Dir dogs. However, there was no increase in lithium reabsorption, PRA, or GFR in the normal diabetic dogs in this study or our previous study. ²⁵ Although we cannot explain the apparent lack of increased fractional proximal sodium reabsorption in the D dogs, the data suggest strongly that it is increased in the Dir dogs. The mechanism is not known, but serum and glucocorticoid-inducible kinase 1 (SGK1) has been shown to regulate proximal tubular glucose transport ³² and to be stimulated by glucose ^{32, 33} and insulin. ^{34, 35} Therefore, an intriguing question for future study is whether insulin and glucose can drive proximal tubule sodium transport in Type II diabetes through a mechanism involving SGK1.

The stimulation of the renin-angiotensin system by intra-renal insulin infusion in the Dir dogs suggests angiotensin II (AngII) is another possible explanation for the antinatriuresis. Indeed, the increase in GFR combined with no change in RBF is tempting to ascribe to increased angiotensin II, such that the increase in filtration fraction would stimulate proximal tubular sodium and fluid reabsorption. However, increased tubular reabsorption in diabetes also has been proposed to increase GFR directly through changes in glomerular filtration pressures, independent of feedback signaling through the macula densa. ³⁶ Thus, the changes in GFR, RBF, PRA and proximal sodium reabsorption may be linked directly or may be independent, parallel responses to increased proximal tubular reabsorption. Future studies that block the renin-angiotensin system will be needed to address this and the role of AngII in mediating the antinatriuresis.

Although there is evidence that insulin can stimulate sodium chloride reabsorption in the proximal tubule and loop of Henle ^{6, 7}, most evidence implicates action in the distal nephron ^{8, 9, 11, 14}, possibly through modulation of ENaC or sodium-chloride co-transport ¹⁴. Insulin regulates ENaC, ^{14, 16, 37-39} and, interestingly, glucose can activate ENaC. ^{14, 40} However, an interactive effect with insulin is not known. It also is interesting that insulin and aldosterone activate ENaC independently via SGK1. ^{16, 37, 38, 41} We have no data yet implicating ENaC, but it is intriguing to consider because aldosterone antagonism markedly attenuated sodium retention and hypertension in obese, hyperinsulinemic dogs even though there was not a statistically significant increase in aldosterone. ⁴² In addition, resistant hypertension is prevalent in obese patients and responsive to aldosterone receptor blockade even though aldosterone may not be elevated. ^{43, 44} This has been ascribed to interaction with insulin via ENaC. ¹⁶ Thus the potential role of ENaC in mediating the sustained, antinatriuretic effect of insulin we discovered in dogs is particularly exciting as it may relate to resistant hypertension and metabolic syndrome.

Perspectives

It is tempting to hypothesize that that the antinatriuretic effect of insulin and glucose could cause hypertension if it occurred in the context of impaired kidney function, thus providing a potential link to the hypertension of metabolic syndrome and Type II. However, it is critical to note that the effect we measured in the Type I diabetic dogs in this study and our previous study ³⁵ was to reverse diabetic natriuresis and diuresis. Thus, there was not an increase in cumulative sodium balance in either case. Rather, insulin acted to prevent, or reverse, the decrease in sodium and water balance. Because of this, we hypothesize that protecting against glucose-induced renal sodium loss is a normal, physiologic function of insulin that has not been recognized previously. Therefore, maintenance of sodium balance during the progression of metabolic syndrome and uncontrolled Type II diabetes, in which plasma insulin and glucose both are elevated chronically, may be due to a cooperative effect of insulin and glucose in the renal tubule to counteract and prevent progressive sodium loss from the sustained hyperglycemia. This hypothesis, the tubular mechanism for the effect, and the potential that this mechanism could raise blood pressure in certain conditions, remain to be tested.