Bacillus anthracis Edema Toxin Increases Fractional Free Water and Sodium Reabsorption in an Isolated Perfused Rat Kidney Model

ABSTRACT

Bacillus anthracis edema toxin (ET) consists of protective antigen (PA), necessary for host cell toxin uptake, and edema factor (EF), the toxic moiety which increases host cell cyclic AMP (cAMP). Since vasopressin stimulates renal water and sodium reabsorption via increased tubular cell cAMP levels, we hypothesized the ET would also do so. To test this hypothesis, we employed an isolated perfused rat kidney model. Kidneys were isolated and perfused with modified Krebs-Henseleit buffer. Perfusate and urine samples were obtained at baseline and every 10 min over 150 min following the addition of challenges with or without treatments to the perfusate. In kidneys perfused under constant flow or constant pressure, compared to PA challenge (n = 14 or 15 kidneys, respectively), ET (13 or 15 kidneys, respectively) progressively increased urine cAMP levels, water and sodium reabsorption, and urine osmolality and decreased urine output (P ≤ 0.04, except for sodium reabsorption under constant pressure [P = 0.17]). In ET-challenged kidneys, compared to placebo treatment, adefovir, an EF inhibitor, decreased urine cAMP levels, water and sodium reabsorption, and urine osmolality and increased urine output, while raxibacumab, a PA-directed monoclonal antibody (MAb), decreased urine cAMP levels, free water reabsorption, and urine osmolality and increased urine output (P ≤ 0.03 except for urine output with raxibacumab [P = 0.17]). Upon immunohistochemistry, aquaporin 2 was concentrated along the apical membrane of tubular cells with ET but not PA, and urine aquaporin 2 levels were higher with ET (5.52 ± 1.06 ng/ml versus 1.51 ± 0.44 ng/ml [means ± standard errors of the means {SEM}; P = 0.0001). Edema toxin has renal effects that could contribute to extravascular fluid collection characterizing anthrax infection clinically.

INTRODUCTION

Bacillus anthracis infection is characterized clinically by the marked retention and extravasation of fluid, which can compromise cardiopulmonary function and impair host defense and tissue repair (1–6). Although the loss of vascular integrity and lymphatic obstruction contribute to this fluid accumulation, the significant hyponatremia in patients in the 2001 U.S. inhalational and 2009 UK soft tissue outbreaks suggests that alterations in renal water and sodium handling may also contribute (3, 5, 7). Notably, of the two toxins produced by B. anthracis, edema toxin (ET) but not lethal toxin (LT) infusion reduced serum sodium levels in canines (8). While ET is known to produce localized edema when injected subcutaneously, understanding whether it alters renal water and sodium handling has important therapeutic implications for the fluid management of patients with severe B. anthracis infection.

Edema toxin consists of protective antigen (PA), the component necessary for host cell uptake of edema factor (EF), the toxic moiety (9, 10). Edema factor has calmodulin-dependent adenyl-cyclase activity and increases intracellular cyclic AMP (cAMP) levels (9, 11, 12). As renal tubular cell cAMP is a critical second messenger stimulating tubular water and sodium reabsorption, ET release during infection could stimulate both functions and contribute to fluid retention (13–15). Despite this potential, a recent PubMed search with the terms “anthrax edema toxin” and either “kidney” or “renal” produced only 14 references, none of which addressed the effects of ET on any aspect of renal physiology, including water and sodium handling. Therefore, we employed an isolated perfused rat kidney model to examine whether ET stimulates tubular water or sodium uptake and mechanisms potentially underlying such effects.

The renal effects of bacterial toxins have been investigated in the isolated perfused rat kidney model previously (16–21). In this model, the perfusate is delivered to and collected from the isolated kidney via the renal artery and vein, respectively, while urine is collected from the ureter. The draining perfusate is recirculated from the vein to the artery by a pump. The concentrations of electrolytes are measured in the perfusate and urine, and changes in renal water and sodium uptake as well as other renal functions are calculated.

Five experiments examined water and sodium reabsorption in this study. As a positive control, in the first experiment, kidneys were perfused with arginine vasopressin (AVP), which is known to stimulate cAMP-mediated water and, to a lesser extent, sodium reabsorption (13–15). In other experiments, the effects of perfusion with ET (EF plus PA) were compared to those with either PA alone (under constant flow or constant pressure [see below]); ET with raxibacumab, a PA-directed monoclonal antibody (MAb) (22); or ET with adefovir, a selective nucleoside inhibitor of EF cAMP production (12). In the experiments comparing ET versus PA or ET with adefovir versus ET without adefovir, two doses of ET were examined. Finally, because cAMP-stimulated tubular water and sodium uptake is mediated by the translocation of aquaporin 2 (AQP2) and epithelial sodium channels (ENaCs) to tubular cell apical membranes, respectively, the effects of ET on the location of AQP2 and ENaC in tubular cells was investigated by immunohistochemistry (13–15). Also, since this translocation produces the shedding of AQP2 and ENaC into the tubular lumen, these molecules were measured in urine.

Most experiments were performed with kidneys perfused with a constant flow to minimize fluctuations in perfusate delivery. However, ET can produce arterial vasodilation and potentially alter renal perfusion pressure and the glomerular filtration rate (GFR), thereby secondarily influencing water and sodium uptake. Therefore, to confirm that the effects of ET on water and sodium were direct tubular effects not related to alterations in perfusion pressure and the GFR, in one experiment, ET was studied in kidneys in which the flow was continuously adjusted to maintain constant perfusion pressure.

RESULTS

Figure 1 shows the overall design of the five perfused kidney experiments examining water and sodium reabsorption. The concentrations of agents and the numbers of animals used in these five experiments are summarized in Table 1. Table 2 shows the baseline flow, perfusion pressure, GFR, fractional free water reabsorption, urine output, and fractional sodium reabsorption for each of the five experiments.

FIG 1.

Timeline of experiments. T0, time zero.

TABLE 1.

Summary of dosages of challenges and treatments and numbers of kidneys employed in the isolated perfused kidney experiments

Expt	Method of perfusion	Challenge		Treatment		No. of kidneys tested
		Type	Dosea	Drug	Dose
Effect of challenge with vasopressin (AVP)	Constant flow	AVP	0.025 μM			3
		Diluent	0			5
	Constant flow	AVP	0.05 μM			6
		Diluent	0			7
	Constant flow	AVP	0.1 μM			5
		Diluent	0 μg/ml			5
Effect of ET challenge with constant flow	Constant flow	PA	2.0 μg/ml			7
		ET	1.0 μg/ml			7
	Constant flow	PA	4.0 μg/ml			7
		ET	2.0 μg/ml			6
Effect of ET challenge with constant pressure	Constant pressure	PA	1.0 μg/ml			9
		ET	0.5 μg/ml			9
	Constant pressure	PA	2.0 μg/ml			6
		ET	1.0 μg/ml			6
Effect of raxibacumab treatment with ET challenge	Constant flow	ET	1.0 μg/ml	Raxibacumab	10×b	8
		ET	1.0 μg/ml	Control	0	6
	Constant flow	ET	1.0 μg/ml	Raxibacumab	50×	3
		ET	1.0 μg/ml	Control	0	3
Effect of adefovir treatment with ET challenge	Constant flow	ET	1.0 μg/ml	Adefovir	20 μM	7
		ET	1.0 μg/ml	Control	0	5
	Constant flow	ET	2.0 μg/ml	Adefovir	20 μM	4
		ET	2.0 μg/ml	Control	0	4

^aET consists of edema factor and PA combined in a weight ratio of 1:2, and the dose of ET shown denotes the amount of EF employed.

^bMolar amount of either raxibacumab or a nonspecific (control) MAb compared to PA.

TABLE 2.

Baseline measurements^a

Parameter	Mean value ± SD
	Effect of challenge with vasopressin (AVP)		Effect of ET challenge with constant pressure		Effect of Raxi with ET challenge		Effect of adefovir with ET challenge		Effect of ET challenge with constant flow
	Diluent	AVP	PA	ET	ET + NS-Ab	ET + Raxi	ET + diluent	ET + adefovir	PA	ET
Flow (ml/min)	32.8 ± 1.0	31.4 ± 0.9	29.3 ± 1.3	27.5 ± 0.8	33.1 ± 1.7	36.1 ± 1.6	30.8 ± 0.7	30.8 ± 0.9	29.5 ± 0.9	30.4 ± 0.6
Perfusion pressure (mm Hg)	101.8 ± 1.4	100.7 ± 0.9	99.9 ± 0.9	100.8 ± 0.8	98.8 ± 0.5	97.7 ± 1.0	100.9 ± 1.0	101.8 ± 1.0	101.1 ± 1.6	100.9 ± 1.6
GFR (ml/min)	0.83 ± 0.03	0.88 ± 0.05	0.71 ± 0.05	0.77 ± 0.06	0.77 ± 0.04	0.83 ± 0.04	0.75 ± 0.03	0.79 ± 0.04	0.75 ± 0.03	0.77 ± 0.03
Fractional free water reabsorption	0.81 ± 0.01	0.81 ± 0.01	0.88 ± 0.01	0.88 ± 0.01	0.83 ± 0.01	0.84 ± 0.01	0.84 ± 0.01	0.84 ± 0.01	0.84 ± 0.01	0.83 ± 0.02
Urine output (μl/min)	157.2 ± 10.9	168.1 ± 11.1	86.6 ± 8.4	95.4 ± 10.5	132.8 ± 13.9	135.7 ± 11.0	123.5 ± 7.8	125.7 ± 9.6	126 ± 15	130 ± 15
Fractional sodium reabsorption	0.82 ± 0.01	0.81 ± 0.01	0.88 ± 0.01	0.88 ± 0.01	0.83 ± 0.01	0.84 ± 0.01	0.84 ± 0.01	0.85 ± 0.01	0.84 ± 0.02	0.84 ± 0.02

^aRaxi, raxibacumab; NS-Ab, nonspecific antibody.

Effect of vasopressin.

As a positive control for the isolated kidney model and for comparison to ET, the experiment to determine the effect of vasopressin examined three doses of arginine vasopressin (0.025, 0.05, or 0.10 μM) versus a diluent control. The effects of these three doses did not differ significantly (P ≥ 0.07) for any parameter, and these data were combined for analyses. Compared to the control, vasopressin increased free water reabsorption and urine osmolality and decreased urine flow in patterns that were significant overall or increased with time (P ≤ 0.04 overall or for the time interaction) (Fig. 2). Vasopressin increased cAMP levels early but not later (P = 0.0004 for the time interaction) and did not alter sodium reabsorption (P = 0.56) or the GFR (P ≥ 0.70) (data not shown) significantly.

FIG 2.

(A and D) Serial mean (±SEM) urine cAMP concentrations (A) and urine osmolality levels (D) in kidneys challenged with vasopressin or the diluent control alone and perfused under constant flow. (B, C, and E) Serial mean changes from the baseline (±SEM) for fractional free water reabsorption (B), the urine flow rate (C), and fractional sodium reabsorption (E). The numbers of kidneys employed in these experiments and baseline data for parameters shown as serial changes are noted in Tables 1 and 2, respectively.

Effects of ET under conditions of constant flow or pressure.

Two experiments investigated the effects of lower and higher ET doses on kidney cAMP release and kidney function under conditions of constant flow or constant pressure. In experiments with both constant flow and constant pressure, lower and higher ET doses had similar effects (P ≥ 0.18, except for the urine cAMP level, which was higher with high-dose ET in the constant-flow experiment [P = 0.001]), and data for the two doses were combined for analyses in each experiment. Under conditions of constant flow (Fig. 3) and constant pressure (Fig. 4), compared to PA, ET increased urine cAMP levels, fractional free water reabsorption, sodium reabsorption, and urine osmolality and decreased urine flow in patterns that were significant overall or increased with time (P ≤ 0.04 overall or for the time interaction, except for sodium reabsorption under constant pressure, which was increased but not significantly [P = 0.17]). Compared to PA, although ET decreased the GFR slightly during the middle of the perfusion period (P ≤ 0.05 for the time interaction) under constant flow, it did not alter the GFR under constant pressure (P ≥ 0.62) (Fig. 5). These findings in kidneys perfused with constant pressure indicate that increases in water and sodium reabsorption with ET were not due to changes in the GFR.

FIG 3.

(A and D) Serial mean (±SEM) urine cAMP concentrations (A) and urine osmolality levels (D) in kidneys challenged with ET or PA alone and perfused under constant flow. (B, C, and E) Serial mean changes from the baseline (±SEM) for fractional free water reabsorption (B), the urine flow rate (C), and fractional sodium reabsorption (E). The numbers of kidneys employed in these experiments and baseline data for parameters shown as serial changes are noted in Tables 1 and 2, respectively.

FIG 4.

(A and D) Serial mean (±SEM) urine cAMP concentrations (A) and urine osmolality levels (D) in kidneys challenged with ET or PA alone and perfused under constant pressure. (B, C, and E) Serial mean changes from the baseline (±SEM) for fractional free water reabsorption (B), the urine flow rate (C) and fractional sodium reabsorption (E). The numbers of kidneys employed in these experiments and baseline data for parameters shown as serial changes are noted in Tables 1 and 2, respectively.

FIG 5.

Serial GFRs for experiments with kidneys challenged with ET or PA alone and perfused under constant flow (A) or constant pressure (B). The numbers of kidneys employed in these experiments and baseline data for parameters shown as serial changes are noted in Tables 1 and 2, respectively. * and ** denote levels of significance for overall effects between study groups and for time interactions between groups, respectively.

Effect of raxibacumab on ET-challenged kidneys.

To examine whether the renal effects of ET were dependent on PA-mediated uptake of EF, the experiment to determine the effect of raxibacumab on ET-challenged kidneys compared either a low or a high dose of raxibacumab to nonspecific MAb (control) treatment in ET-challenged kidneys under constant flow. The effects of the raxibacumab doses did not differ (P ≥ 0.15), and these data were combined in analyses. In ET-challenged kidneys, compared to the control, raxibacumab decreased urine cAMP levels, free water reabsorption, and urine osmolality in patterns that were significant overall or increased with time (P ≤ 0.04 overall or for the time interaction) (Fig. 6). Raxibacumab increased urine flow but not significantly (P = 0.17) and did not alter sodium reabsorption (P = 0.95). Raxibacumab did not alter the GFR significantly (P = 0.60) (data not shown).

FIG 6.

(A and D) Serial mean (±SEM) urine cAMP concentrations (A) and urine osmolality levels (D) in kidneys challenged with ET, treated with raxibacumab or a nonspecific control MAb, and perfused under constant flow. (B, C, and E) Serial mean changes from the baseline (±SEM) for fractional free water reabsorption (B), the urine flow rate (C), and fractional sodium reabsorption (E). The numbers of kidneys employed in these experiments and baseline data for parameters shown as serial changes are noted in Tables 1 and 2, respectively.

Effect of adefovir on ET-challenged kidneys.

To examine whether EF-stimulated cAMP production was necessary for the effects of ET, the experiment to determine the effect of adefovir on ET-challenged kidneys compared the effects of adefovir to diluent (control) treatment on kidneys challenged with low or high ET doses under constant flow. Adefovir's effects on parameters did not differ with the two doses of ET except that urine flow and the GFR were higher with the high dose (P ≤ 0.05), and data from the two ET doses were combined in analyses. In ET-challenged kidneys, compared to the control, adefovir decreased urine cAMP levels, free water and sodium reabsorption, and osmolality levels (P ≤ 0.03 for the overall effects) and increased urine flow in patterns that were significant overall or increased with time (P ≤ 0.03 overall or for the time interaction) (Fig. 7). Adefovir increased the GFR at later time points (P = 0.04 for the time interactions) (data not shown).

FIG 7.

(A and D) Serial mean (±SEM) urine cAMP concentrations (A) and urine osmolality levels (D) in kidneys challenged with ET, treated with adefovir or a diluent, and perfused under constant flow. (B, C, and E) Serial mean changes from the baseline (±SEM) for fractional free water reabsorption (B), the urine flow rate (C), and fractional sodium reabsorption (E). The numbers of kidneys employed in these experiments and baseline data for parameters shown as serial changes are noted in Tables 1 and 2, respectively.

Aquaporin 2 and epithelial sodium channel component analysis of tissue and urine.

AQP2 staining was localized along tubular cell apical membranes in the fields of kidneys perfused with vasopressin or ET (n = 1 and 2, respectively) but not with HEPES (vasopressin diluent) or PA (n = 1 and 2, respectively) (Fig. 8). When the normalized staining intensity was examined along a line across tubular cells from the extraluminal region (interstitial space) to the tubular lumen for five tubules selected from each kidney, peak staining in ET- and vasopressin-perfused kidneys occurred over a much narrower distance than in PA- and HEPES-perfused kidneys and was localized at the apical (luminal) membrane (Fig. 8). Consistent with this localization with ET, free AQP2 levels were significantly higher in urine from ET-perfused kidneys than in urine from PA-perfused kidneys at 90 and 120 min of perfusion (5.52 ± 1.06 ng/ml [n = 14] versus 1.51 ± 0.44 ng/ml [n = 12] [means ± standard errors of the means {SEM}, averaged over the two time points]; P = 0.0001).

FIG 8.

Sample photomicrographs (A, C, E, G, I, and K) and AQP2 fluorescence intensity normalized to the maximum intensity for each image plotted along a 10-μm line crossing tubular cells from extraluminal to intraluminal regions in five selected tubules for each of the six kidneys (B, D, F, H, J, and L). In the photomicrographs, AQP2 staining is shown in pink, while nuclear staining and autofluorescence are shown in gray. Shown are photomicrographs and fluorescence intensity data for one kidney each challenged with AVP (A and B) or HEPES (AVP diluent) (C and D) and for two kidneys each challenged with edema toxin (ET for kidney 1 [ET 1] [E and F] and ET 2 [G and H]) or protective antigen alone (PA for kidney 1 [PA 1] [I and J] and PA 2 [K and L]). While the fluorescence intensity was dispersed widely along the plotted lines across tubules from kidneys challenged with HEPES and PA, this intensity tended to be localized over a short distance for tubules from AVP- and ET-challenged kidneys. As presented in Results, consistent with this narrow localization with ET but not PA, free AQP2 levels were significantly higher in urine from ET-perfused kidneys than in urine from PA-perfused kidneys at 90 and 120 min of perfusion (5.52 ± 1.06 ng/ml [n = 14 kidneys] versus 1.51 ± 0.44 ng/ml [n = 12] [means ± SEM, averaged over the two time points]; P = 0.0001).

While staining for ENaC subunits α and β was weak in kidneys, staining for subunit γ was stronger but diffuse in the fields surveyed in both ET- and PA-perfused kidneys (data not shown). However, in urine collected at 90 and 120 min, ENaC subunit α levels were higher in ET- than in PA-perfused kidneys (1.37 ± 0.21 ng/ml [n = 16] versus 0.70 ± 0.23 ng/ml [n = 16] [means ± SEM, averaged over the two time points]; P = 0.04). There were insufficient samples to measure urinary ENaC subunit β and γ levels (data not shown).

DISCUSSION

The present findings are the first that we are aware of that describe the direct effect of B. anthracis ET and its adenyl-cyclase activity on renal physiological function and the tubular uptake of water and sodium. These findings are highly relevant to the retention and extravasation of fluid that characterize severe B. anthracis infection. Challenge with ET in this isolated perfused rat kidney model increased urine cAMP levels and fractional free water reabsorption. Consistent with this reabsorption, ET increased urine osmolality and decreased urine flow. The effects of ET on water uptake, osmolality, and urine flow were similar to the changes that occurred when kidneys were exposed to vasopressin, an agent known to stimulate renal tubular free water uptake and concentrate urine. However, increases in urine cAMP levels were briefer with vasopressin than with ET, possibly because the former is rapidly metabolized (23). The findings that raxibacumab and adefovir decreased urine cAMP levels, water reabsorption, and urine osmolality and increased urine flow with ET challenge support the possibility that EF uptake and its stimulation of cAMP in renal tubular cells mediated the effects of ET. Challenge with ET also increased sodium reabsorption, although these increases were not as great in kidneys perfused with constant pressure as in kidneys perfused with a constant flow, and only adefovir inhibited this effect.

It seems very likely, based on known aspects of renal water handling and the findings here, that the impact of ET on water uptake was mediated by an effect on AQP2. Free water uptake in the renal tubule is regulated by water diffusion along an osmotic gradient through AQP2 water channels in the apical membrane of tubular cells (13, 14). Vasopressin increases free water absorption in part by attaching to arginine vasopressin 2 (AVP2) receptors on the basal membrane of collecting duct cells, which in turn activates the Gs/adenyl-cyclase system and increases intracellular cAMP levels. These increases in cAMP levels activate protein kinase A (PKA), leading to the translocation and insertion of active AQP2 channels into the apical membrane (13, 14). Other agents that increase the levels of intracellular cAMP (e.g., forskolin and calcitonin) and cAMP analogues (e.g., 8-Br–cAMP) stimulate PKA, AQP2 phosphorylation and membrane expression, and free water uptake (24–29). Protein kinase A blockade inhibits vasopressin-mediated AQP2 membrane expression and free water uptake (30–33). With respect to ET, the two anthrax toxin receptors (ANTR1 or TEM8 and ANTR2 or CMG2) that internalize PA and EF are expressed in renal tissue (34–36). Like vasopressin, EF stimulates cAMP gradients in cells starting in the perinuclear region, where A kinase-anchoring proteins participate in PKA-mediated AQP2 phosphorylation (37). ET-stimulated cAMP production is known to trigger PKA-dependent signaling, and PKA inhibition blocks the cellular effects of ET (38, 39). In the present study, the selective inhibition of cAMP production by EF with adefovir blocked increased free water uptake with ET. Finally, similarly to vasopressin, in the present study, ET concentrated the staining of AQP2 along the apical membrane of tubular cells, a histological effect previously noted for mouse renal tubular cells (40). Also, consistent with this localization of AQP2 in tubular cell apical membranes, ET increased urine AQP2 levels in the present study. Vasopressin similarly increases urine AQP2 levels in conjunction with increased free water uptake in rats (41).

The increased tubular cell cAMP levels and their potential effect on ENaC expression also provide a basis for the increased sodium reabsorption observed with ET. This possibility is supported by the inhibitory effect of adefovir on sodium uptake with ET. Also, ET increased urine ENaC subunit α levels, although these increases were not as high as those with AQP2. Vasopressin-stimulated cAMP production increases ENaC expression and sodium reabsorption, although the absence of such an effect with vasopressin here may be related to the relatively rapid metabolism of exogenously administered vasopressin (15).

The present findings suggest that ET could lead to water and possibly sodium retention during clinical B. anthracis infection, especially in the context of other mediators produced by this bacterium that can contribute to alterations in extravascular fluid volumes (10, 42). The bacterium's cell wall produces a robust inflammatory response capable of decreasing endothelial and epithelial barrier function (43, 44). Also, lethal toxin and other anthrax proteases can disrupt endothelial and epithelial barrier function (42). During systemic hypotension or shock with anthrax, it is likely that mineralocorticoid production would increase and also favor sodium and fluid retention (45). Thus, increased free water or sodium reabsorption with ET, combined with these other processes, may be an important mechanism leading to the retention and extravasation of fluid that complicates B. anthracis infection. Treatment with agents designed to inhibit ET could limit this detrimental fluid accumulation clinically and improve outcomes.

The concentrations of ET (EF combined with PA) employed in this isolated perfused rat kidney model were designed to simulate the estimated intravascular concentrations of these factors that had previously produced lethality rates of 50 to 75% in our ET-infused rat model (44). These concentrations were also close to the range of intravascular EF concentrations reported previously for rabbits challenged with 20 50% lethal doses (LD₅₀) of B. anthracis (strain A0843), measured 48 h after challenge by Western blotting and engineered immunoassays (1.9 ± 0.4 to 6.1 ± 2.1 μg/ml) (46). Of note, however, recent studies have shown that the plasma concentration of enzymatically active EF present during the development of lethal infection in mice following cutaneous B. anthracis challenge increased from 0.66 ± 0.08 log₁₀ pg/ml in the earliest stage to 4.28 ± 0.13 log₁₀ pg/ml in the final stage (47). The enzymatically active concentrations of EF that produced the changes in tubular function observed here are not known but may be very relevant when considering the concentrations of therapeutic inhibitors required to potentially reverse renal effects.

This study has several limitations. First, the isolated kidney model is viable for relatively short periods of time, and whether the effects of ET on renal tubular function persist is unclear. However, using both canine (8) and rat (our unpublished data) in vivo models, we showed that 24-h ET infusions progressively decreased serum sodium concentrations. Second, the concentration of ET employed in these experiments varied in some experiments based on the supply of toxin available for study. However, all experiments included kidneys challenged with EF at a concentration of 1 μg/ml. Finally, ET produced small changes in the GFR in kidneys perfused under constant flow. However, it did not alter the GFR in kidneys under constant pressure. Thus, the effects of ET on water and sodium reabsorption, urine osmolality, and urine flow did not appear to be related to the toxin's vascular effects (48).

In conclusion, anthrax infection is characterized clinically by extravascular fluid collection and tissue edema. The present findings are the first to suggest that ET release during infection contributes to these events via its systemic effects on renal handling of water and possibly sodium reabsorption. Further defining these effects may lead to therapeutic interventions that reduce the morbidity associated with this hallmark clinical manifestation of anthrax infection.

MATERIALS AND METHODS

Animal care.

This study was reviewed and approved by the Animal Care and Use Committee of the Clinical Center of the NIH.

Isolated perfused kidney experiments and groups.

Table 1 summarizes challenge and treatment doses and numbers of kidneys tested in five experiments. Figure 1 shows the timeline for the experiments. The first experiment compared the effects of challenge with vasopressin at 0.025 μM (n = 3) versus the diluent (n = 5), vasopressin at 0.05 μM (n = 6) versus the diluent (n = 7), or vasopressin at 0.1 μM versus the diluent (n = 5 each) on renal perfusion pressure, the GFR, urine cAMP levels, water and sodium reabsorption, and the urine flow rate and osmolality in kidneys isolated from Sprague-Dawley rats. Kidneys were perfused with a constant flow over 150 min following equilibration and confirmation of kidney stability (see below).

The second experiment presented examined whether ET altered isolated perfused kidney function under constant-flow conditions. This experiment compared the effects of challenge with a lower ET dose (EF at 1 μg/ml plus PA at 2 μg/ml) (n = 7) versus PA alone (2 μg/ml) (n = 7) or a higher ET dose (EF at 2 μg/ml plus PA at 4 μg/ml) (n = 6) versus PA alone (4 μg/ml) (n = 7). A third experiment compared a lower ET dose (EF at 0.5 μg/ml plus PA at 1 μg/ml) (n = 9) versus PA alone (1 μg/ml) (n = 9) or a higher ET dose (EF at 1 μg/ml plus PA at 2 μg/ml) (n = 6) versus PA alone (2 μg/ml) (n = 5) in kidneys perfused under constant pressure.

A fourth experiment examined the effect of raxibacumab, a PA-directed monoclonal antibody, on the renal effects of ET. In kidneys challenged with ET (EF at 1 μg/ml EF plus PA at 2 μg/ml), a lower or a higher raxibacumab dose (10 or 50 times the molar dose of PA [n = 5 and 3, respectively]) was compared to a nonspecific monoclonal antibody (control) (n = 5 and 3, respectively) under constant flow. A fifth experiment examined the effects of adefovir, a selective nucleoside inhibitor of EF adenyl-cyclase activity, on the renal effects of ET. Adefovir (20 μM) was compared to the diluent (control) in kidneys challenged with a lower ET dose (EF at 1 μg/ml plus PA at 2 μg/ml) (n = 7 with adefovir and n = 5 with the diluent) or a higher ET dose (EF at 2 μg/ml plus PA at 4 μg/ml) (n = 4 each with adefovir or the diluent) under constant flow.

Kidney isolation and perfusion.

In anesthetized (3 to 5% inhaled isoflurane) Sprague-Dawley rats (300 to 350 g), the right ureter was cannulated with PE10 tubing. A 19-gauge blunt needle was positioned in the renal artery, and in vivo kidney perfusion was initiated with a Baxter syringe pump. The kidney was then removed and suspended in a 37°C water chamber. The renal artery catheter was connected to a recirculating perfusion system containing 80 ml of the perfusate maintained at 37°C, continuously gassed with 95% O₂ and 5% CO₂, and recirculated with a Masterflex pump (Cole Parmer, Vernon Hills, IL). The renal ureter was accessed for urine collection. The renal vein emptied into the perfusion system. A pressure transducer measured renal artery perfusion pressure.

Kidneys were perfused with a modified Krebs-Henseleit (KH) buffer containing 118.5 mM NaCl, 25 mM NaHCO₃, 4.7 mM KCl, 1.5 mM MgSO₄, 1.5 KH2PO₄, 1.92 mM CaCl₂, 5.74 mM glucose, 4% bovine serum albumin (BSA), 1.67% dextran, and 20 amino acids (13 mM final concentration). The solution was equilibrated with 95% O₂ and 5% CO₂. The pH was adjusted to 7.40 by using 1 N HCl.

In the four constant-flow experiments, kidneys were perfused and equilibrated for 30 min while the perfusion flow rate was adjusted to maintain a perfusion pressure of 100 mm Hg (Fig. 1). Kidneys were excluded if the urine output was <0.5 ml over a 10-min period at the end of equilibration. After 20 min of equilibration, fluorescein isothiocyanate (FITC)-inulin was added to the system (0.2-mg/ml final concentration). Following equilibration, the perfusion flow rate was not changed further (i.e., constant flow was maintained). After 10 min of additional perfusion, at time zero, renal perfusion flow and pressure were measured, and a perfusate sample (300 μl) and all urine that accumulated over the interval were collected. Challenges and treatments were then added to the perfusion system as described above. Every 10 min thereafter, repeat measurements and samples were collected. Perfusate and urine samples were replaced in the perfusion system with an equal volume of fresh perfusate. In the one constant-pressure experiment, the perfusion flow rate was adjusted every 2 to 3 min throughout the experiment to maintain a perfusion pressure of 100 mm Hg.

Measurements and calculations for the isolated perfused kidney experiments.

The inulin concentration was measured by fluorescence with a plate reader (BioTek, Winooski, VT). Sodium, potassium, chloride, and glucose concentrations in perfusate and urine samples were measured by using the Stat Profile Critical Care Express system (Nova Biomedical, Waltham, MA). The cAMP level was measured by using a chemiluminescent direct immunoassay (Arbor Assays, Ann Arbor, MI). Urine osmolality was measured by freezing point depression (model 3320 osmometer [Advanced Instruments Inc.]; Cornell University Animal Health Diagnostics Center). Electrolytes and urine volumes were analyzed at 10-min intervals, and cAMP levels and urine osmolality were analyzed at 30-min intervals.

The formula for calculation of the GFR was GFR = (U_inulin × UFR)/P_inulin, where UFR is the urine flow rate (milliliters per minute), U_inulin is the urine inulin concentration (milligrams per milliliter), and P_inulin is the perfusate inulin concentration (milligrams per milliliter); the formula for calculation of fractional free water reabsorption (FRwater) was FRwater = 1 − (P_inulin/U_inulin); and the formula for calculation of the fractional reabsorption of sodium (R_Na) was R_Na = 1 − [(UFR × U_Na)/(GFR × P_Na)], where U_Na is the urine concentration of sodium (millimolar) and P_Na is the perfusate concentration of sodium (millimolar).

Aquaporin 2 and epithelial sodium channel component analysis of tissue and urine.

Immunohistochemistry staining of renal tissue was performed with an R&D Systems protocol. Paraffin-embedded kidney sections were deparaffinized in xylene, followed by an ethanol gradient, and washed with buffer (0.1 M Tris buffer, pH 7.4). After blocking for 1 h at room temperature in 0.1 M Tris buffer supplemented with 1% bovine serum albumin (Sigma-Aldrich, St. Louis, MO), 1% fetal bovine serum (GemCell; Gemini Bio-Products, West Sacramento, CA), 0.3% Triton X-100 (Supelco, Bellefonte, PA), and 2% goat serum (Jackson ImmunoResearch, West Grove, PA), tissue was incubated with a primary antibody to AQP2 (rabbit polyclonal; 1:2,500 dilution) or ENaC subunit α, β, or γ (rabbit polyclonal; 1:50 dilution) at 4°C overnight. Mark Knepper (NHLBI, Bethesda, MD) provided all primary antibodies. An isotype control antibody served as a negative control (Abcam, Cambridge, United Kingdom). After washing, tissues were incubated with goat anti-rabbit secondary antibody conjugated to Alexa Fluor 647 (AF647) (Abcam, Cambridge, UK) for 7 h at room temperature. Slides were washed with 0.1 M Tris buffer and mounted with ProLong Gold Antifade mounting reagent with 4′,6-diamidino-2-phenylindole (DAPI) (Life Technologies, Rockville, MD). Confocal images of mounted slides were acquired on a Zeiss 710 confocal microscope (Carl Zeiss Microimaging, Munich, Germany) using a 40× 1.3-numerical-aperture (NA) objective and a 633-nm single-photon laser. For all kidneys, several areas of at least 0.01 to 0.1 mm² were surveyed. Images were processed by using Imaris (Bitplane, Zürich, Switzerland) and Fiji image processing software, and localization data were obtained as follows: lines were drawn across tubular cells from the extra- to the intraluminal tubular spaces, and the fluorescence intensity in the far-red channel was plotted versus distance. Data were normalized to the maximal intensity for each image (given an arbitrary value of 100), and the normalized distribution of fluorescence was plotted against distance. This was done for 5 tubules from each of two kidneys for the ET and PA groups and from one kidney for the AVP and HEPES groups.

Urine samples collected after 90 or 120 min of perfusion with ET or PA were analyzed with rat AQP2 and amiloride-sensitive ENaC subunit alpha enzyme-linked immunosorbent assay (ELISA) kits (MyBioSource, San Diego, CA).

Toxin and treatment preparation.

Toxin components (PA and EF) were recombinant proteins prepared from Escherichia coli as previously described (49). ET consisted of EF with PA at a ratio of 1:2 on the basis of weight. The lipopolysaccharide (LPS) contents of PA and EF were 0.001 and 0.006 ng/μg, respectively. Raxibacumab (Human Genome Sciences) is a PA-directed human monoclonal antibody, and the control was an inactive, nonspecific monoclonal antibody (50). Adefovir dipivoxil (Sigma-Aldrich, St. Louis, MO) was dissolved in ethanol and then diluted to a working solution using KH buffer (12). Vasopressin (Arg8-vasopressin acetate) (Sigma-Aldrich) was diluted to a working solution in KH buffer.

Statistics.

Data were analyzed by using PROC Mixed in SAS version 9.3 software (SAS Institute Inc., Cary, NC). Three-way repeated-measures analysis of variance (ANOVA) accounting either for challenge type (ET versus PA or AVP versus the diluent), challenge dose, and time point or for treatment type (raxibacumab or adefovir versus the control), treatment dose, and time point for ET challenge were initially conducted to assess the effects of challenge or treatment, respectively. Since the effects of ET or treatment compared across doses for almost all parameters were not statistically different, these data were combined to increase the power of the analysis and limit the number of animals required for study. Two-way ANOVA was then used to assess the effects of challenge or treatment across time or overall. Baseline data for parameters measured at 10-min intervals are shown in Table 2, and subsequent data for these parameters are shown as changes from the baseline in figures. Throughout, data are presented as means ± SEM. A P value of ≤0.05 was considered significant.