| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
RENAL HEMODYNAMICS AND CARDIORENAL INTEGRATION
1Division of Nephrology, The Kidney Research Centre, Ottawa Health Research Institute, and 2Department of Pathology, University of Ottawa, Ottawa, Ontario, Canada
Submitted 26 June 2007 ; accepted in final form 10 April 2008
 |
ABSTRACT |
|---|
 TOP ABSTRACT METHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
hyperfiltration; tubuloglomerular feedback; neuronal nitric oxide synthase
For more than 25 years, the effect of loss of renal mass has also attracted attention because of demonstrations of the acceleration of preexisting renal disease and changes in the control of GFR in the remnant kidney (4, 15). The combined effects of reduction of renal mass and diabetic nephropathy on renal function and progression of diabetic nephropathy have aroused clinical (11) and experimental interest. Again, these investigations usually focused on the insulin-treated rat STZ model of DM1 (23, 27).
With respect to whole kidney GFR, 30 years ago, Gartner (12) documented hyperfiltration in the obese db/db (B6.Cg-m+/+Leprdb/J) model of DM2. Shortly thereafter, Bower et al. (6), in a key study using the DM2 db/db mouse, documented exacerbation of diabetic nephropathy induced by uninephrectomy (Nx) and suggested that alterations in glomerular hemodynamics subsequent to Nx influence the rate of development of diabetic glomerular lesions. Few studies have attempted to address in vivo single-nephron effects of DM2, because appropriate rat models are not readily available and in vivo micropuncture assessments in hyperglycemic DM2 mouse models with glycosuria can be technically overwhelming (36). We recently established that single-nephron GFR (SNGFR) is significantly elevated in the db/db mouse compared with heterozygote and wild-type (WT) controls (21) and reported associated changes in the tubuloglomerular feedback (TGF) system, as well as effects related to extracellular fluid volume (ECV) changes and acute neuronal nitric oxide (NO) synthase (nNOS) inhibition.
Several intrarenal factors may contribute to the early increase in whole kidney GFR and SNGFR in experimental models of diabetes mellitus (17–19, 29). These factors include, separately or in combination, hypertrophy of glomerular and tubular structures, altered tubular transport, changes in TGF responses, and alterations in intrarenal NO synthase (NOS) activity. The tubular hypothesis of glomerular hyperfiltration, supported by several studies in the STZ DM1 rat preparation, points to enhanced proximal tubule fluid reabsorption in early diabetes as a key determinant of hyperfiltration (31, 32).
In the present investigations, we evaluated functional and structural changes following Nx in the gene-targeted hyperglycemic obese db/db mouse model of DM2. We hypothesized that 1) Nx of 3 wk duration leads to an additional increase in whole kidney GFR and SNGFR in the already hyperfiltering db/db mice, 2) this additional increase in SNGFR is likely associated with a resetting of the TGF system, 3) chronic NOS1 inhibition attenuates the hyperfiltration, and 4) Nx in these DM2 mice eventually leads to exacerbation of the lesions of diabetic glomerulopathy. More specifically, to understand possible mechanisms underlying filtration changes in our db/db DM2 mouse model, in the context of the tubular hypothesis of glomerular hyperfiltration (32), we focused on three single-nephron parameters before and after Nx, using WT and sham-operated mice as controls, in the four groups of mice: 1) distal tubule-measured SNGFR (without modification of flow to the macula densa site), 2) proximal SNGFR with zero distal flow, and 3) proximal tubule fluid reabsorption rates to assess net fluid transport.
|
METHODS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
3 wk before the stressful micropuncture experiments. Thus in vivo studies were carried out in 13- to 15-wk-old db/db or WT mice 3 wk after Nx or the sham procedure. In addition to micropuncture and histopathology studies in 15-wk-old mice, histopathological changes were assessed in 30- and 40-wk-old mice. Experimental Groups
Groups 1–4.
Effects of Nx were assessed by micropuncture studies on four groups of 13- to 15-wk-old mice 3 wk after Nx or sham Nx: db/db Nx (group 1) and WT Nx (group 3) mice with corresponding db/db sham and WT sham mice (groups 2 and 4, respectively).
Groups 5 and 6.
The possible effects of chronic nNOS inhibition on SNGFR and related parameters were examined in db/db Nx mice that ingested S-methylthiocitrulline (SMTC) at 12 and 24 µg/day (groups 5 and 6, respectively) in drinking water beginning 2 days after Nx and continuing for 2–3 wk. These ingested SMTC doses are similar on a per-weight basis to those used in rats by Komers et al. (16), who were able to modify the lesions of diabetic nephropathy after 12 wk of ingestion.
All studies were approved by the Animal Care Committee of the University of Ottawa.
Micropuncture Preparation
Mice were anesthetized with thiobutabarbital sodium (100 mg/kg ip; Inactin, Sigma-Aldrich, St. Louis, MO) and ketamine hydrochloride (100 mg/kg im; Bimeda-MTC, Cambridge, ON, Canada). After anesthesia, the mouse was placed on a heated mouse operating table; a tracheostomy was performed, and PE-90 tubing was inserted. The left carotid artery was cannulated with pulled PE-10 tubing for continuous blood pressure measurement and blood sampling for glucose and inulin measurements. Cannulation of the left jugular vein with pulled PE-10 tubing provided two lines for infusion of fluid and thiobutabarbital sodium anesthesia. The bladder was cannulated with PE-50 tubing. The left kidney was exposed by a flank incision, separated from the surrounding fat, and carefully dissected away from the adrenal gland, while the ureter remained attached to the kidney. Because of fat accumulation in the neck, and especially in the abdomen, in db/db mice, the surgical procedure was more difficult. Because of the excessive intra-abdominal fat, a longer flank incision was made in the linea alba to provide space for the kidney cup, and special care was taken to prevent excess heat and fluid losses from the larger wound by using Surgicel (Ethicon). Surgicel was placed just below the kidney cup, covering the entire incision and, thereby, preventing excessive widening of the abdominal wound. This operative approach allowed sufficient fat to be shifted, so that the kidney could be positioned appropriately after it was immobilized in a Lucite cup (10–12 mm). The kidney was set in 0.9% NaCl agar, with a surface opening created to hold a shallow layer of heavy mineral oil. The mice were infused with a priming dose of 5–10 µCi of [H3]inulin followed by a sustaining infusion of an equal amount of 2.25% albumin in 0.9% saline at 3.3 ml·100 g body wt–1·h–1 for the duration of the experiment.
Clearance Studies
For determination of GFR and other whole animal parameters, mice were prepared as described above, but without isolation of the left kidney. The protocol was designed to replicate the time, infusion, and blood sampling of the micropuncture experiments. Blood samples were taken for hematocrit, glucose, and inulin. Urine was collected under oil for determination of inulin concentration and urine flow from both kidneys. The mice were infused with a priming dose of 5–10 µCi of [H3]inulin and a sustaining infusion of an equal amount per hour of [3H]inulin in 0.9% saline. Again, this protocol was designed to replicate the time, infusion, postoperative days, age, and blood sampling of the micropuncture experiments.
SNGFR and TGF Responses
SNGFR was determined at the late-proximal and distal tubular sites by the usual technique of [H3]inulin infusion, sparse use of Lissamine green dye, and timed collections of quantitative tubular flow by careful oil block insertions. To assess TGF activity, we measured, in a paired fashion, proximal-distal SNGFR differences as described by Lorenz et al. (22). This method is sensitive to SNGFR changes in the "normal"-to-low range, is more likely than the stop-flow pressure (PSF) technique to detect TGF responses to very low flows, and may be more sensitive to changes in ECV (22, 31).
Nx and Sham Nephrectomy
Male WT (25–30 g body wt) and 8- to 10-wk-old db/db (40–50 g body wt) mice with hyperglycemia, polydipsia, and polyuria were selected for Nx by surgical technicians of the University of Ottawa's Animal Care Veterinarian Services in a sterile, climate-controlled facility. The mice were placed in a restrainer without anesthesia, and one hindleg was extended. For blood glucose determination, the lower surface of the hindleg was shaved so that the saphenous vein could be visualized and needled. At 1 h before surgery, buprenorphine (0.05 mg/kg sc; RB Pharmaceuticals) was administered, and baseline body weight was determined. Anesthesia was induced with 0.5–3% isoflurane. The isoflurane was mixed with oxygen at a rate of 1 l/min, and before the first incision, 1 ml of normal saline was injected subcutaneously and a BNP ophthalmic ointment (Vetcom) was applied. For kidney access, the dorsal area was shaved and prepped, and a small incision was made dorsally so that the right kidney could be pulled out and isolated. The fat at the lower pole of the kidney was grasped and blunt dissected (with care taken to spare the adrenal gland), and three 4-0 silk ligatures were passed under the ureter, renal artery, and renal vein; one ligature was tied distally, and two were tied proximally. A cut was made between the ligatures, the proximal ties were shortened and the distal tie was removed, and the kidney was removed. The incision was carefully checked for bleeding, the peritoneum was closed with 6-0 prolene suture using a simple interrupted pattern, and the skin was closed with autoclips. The right kidney was promptly weighed after the fat was removed. The mice were kept in an incubator until they were conscious and moving. Postoperative care included subcutaneous administration of 1 ml of sterile saline for dehydration and 0.05 mg/kg of buprenorphine over 3 days according to Animal Care Veterinarian Services guidelines.
Sham mice underwent the same procedure, except the kidneys were only touched by the instruments: no incision was made. All Nx and sham mice were allowed 3 wk to recover before subsequent micropuncture experiments.
Histopathology
For histopathological assessment, whole left kidneys from usually two or three mice per group were bivalved, fixed in buffered formalin, embedded in paraffin, and sectioned at 3 µm. Each kidney was stained with hematoxylin and eosin, Jones silver methenamine, periodic acid-Schiff, Masson's trichrome, and Sirius red. One full half-section was stained with hematoxylin and eosin and assessed for glomerular diameter. All glomeruli in that section were measured using an ocular grid micrometer at x400 magnification. Approximately 60–100 glomeruli were examined per section. Mesangial matrix expansion was assessed using the Sirius red-stained sections. High-resolution digital photographs of 50 glomeruli sectioned through, or close to, the hilum were prepared from the Sirius red-stained sections from each animal. These photographs were subsequently quantified by assessment of the ratio of surface area of mesangium to total glomerular surface area with color-assisted image analysis using ImagePro software (version 5.1).
Statistical Analyses
For two-group comparisons, two-tailed unpaired t-tests were used. For four-group comparisons, ANOVA was used to test the hypothesis that a given parameter was equal in the four groups (e.g., db/db Nx, db/db sham, WT Nx, and WT sham). If the hypothesis of equality between groups was rejected, then an unpaired two-tailed t-test or Tukey's post test was used to determine group differences. In these tests, data involving repeated measures were averaged for each mouse. Paired t-tests were used to compare parameters such as SNGFR within a single group. For a single-group analysis of one parameter, where each mouse served as its own control, the paired t-test data included each repeated measure as a separate observation. The significance level to reject the null hypothesis was set at 0.05 for each test.
Statistical analysis for histopathological assessment was done using SPSS 15.0. Glomerular diameters of db/db mice, with and without Nx, were assessed using planned comparisons between Nx and sham Nx mice at each age, i.e., 15, 30, and 40 wk (2-tailed t-tests). Planned comparisons of glomerular diameters of WT Nx and db/db Nx mice at 30 and 40 wk of age were also done. A one-way ANOVA was done to compare glomerular diameters at 15, 30, and 40 wk in db/db Nx mice. Post hoc analyses included polynomial contrasts and pairwise comparisons [Tukey's honestly significantly difference (HSD) test]. Similarly, glomerular mesangial ratios of db/db sham and db/db Nx mice were each compared at 15 and 30 wk (2-tailed t-test). Glomerular mesangial ratios at 30 wk in WT Nx and WT sham mice were also compared (2-tailed t-test). Glomerular mesangial ratios were also compared in WT Nx and db/db Nx mice at 30 and 40 wk (1-way ANOVA).
|
RESULTS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
Whole animal data from db/db Nx, db/db sham, WT Nx, and WT sham mice (groups 1–4) infused with a solution of 2.25% albumin in 0.9% saline at 3.3 ml·100 g body wt–1·h–1 for the duration of the experiment are summarized in Table 1. In our previous db/db mouse micropuncture investigation (21), we noted that whole kidney GFR tended to decline during the experiment. In the present study, with use of a constant albumin infusion, there was no significant GFR difference when the first period or the mean of two or three periods was considered, indicating that albumin has a stabilizing effect. In detailed preliminary experiments, we also assessed possible effects of this infusion on ECV, SNGFR, and TGF responsiveness of this sustaining infusion. We studied an additional group of db/db Nx mice infused with the same albumin solution, but at a 40% lower rate. Single-nephron responses were virtually identical to those obtained from group 1 (see below).
|
Histopathology
Histopathological results are shown in (Fig. 1).
|
2 = 0.20] and 40 wk [95.6 ± 2 vs. 108.3 ± 8.5 µm, F(1,754) = 76, P < 0.01, partial 2 = 0.09]. There was also a significant and marked difference in mean mesangial ratio between db/db Nx and WT Nx at 30 wk [0.46 ± 0.17 vs. 0.12 ± 0.02, F(1,800) = 64.3, P < 0.01, partial 2 = 0.10] and 40 wk [0.31 ± 0.15 vs. 0.17 ± 0.01, F(1,284) = 14.4, P < 0.01, partial 2 = 0.05].
Effects of time and Nx on glomerular diameter and mesangial ratio in db/db mice.
At the baseline 15-wk measurement, there was no difference between db/db Nx and db/db sham mice with respect to glomerular diameter (65.8 ± 14.0 vs. 66.5 ± 14.6 µm, P > 0.1). Similarly, there was no significant difference in glomerular diameter in db/db Nx and db/db sham mice at 30 or 40 wk, consistent with the absence of a separate effect of the Nx status on db/db mouse glomerular diameter. However, within the db/db Nx group, glomerular diameter did significantly increase with time. There was a significant difference in glomerular diameter at 15 wk compared with 30 or 40 wk [F(2,1854) = 187.7, P < 0.001, partial 2 = 0.17]. Post hoc polynomial contrast showed a linear trend of increasing diameter with time [F(1, 1854) = 347.6, P < 0.001], and post hoc analysis (Tukey's HSD) showed a significant difference between mean diameter in 15- and 30-wk-old, as well as 30- and 40-wk-old, db/db Nx mice (P < 0.001).
At 15 wk, there was no difference between the db/db Nx and db/db sham mice with respect to glomerular mesangial ratio (representing the mesangial matrix surface area vs. total glomerular surface area) (0.08 ± 0.06 vs. 0.04 ± 0.08 µm, P > 0.1); however, with time there was an increase of the mesangial matrix expansion in db/db Nx compared with db/db sham mice. Thus, in the 30-wk-old db/db Nx vs. db/db sham mice, the mesangial ratio was 0.46 ± 0.17 vs. 0.27 ± 0.02, P < 0.01, although glomerular diameters, as already stated, did not differ. In contrast, at 30 wk, the control WT Nx and WT sham mice showed no difference in the mesangial ratio (0.12 ± 0.05 vs. 0.18 ± 0.08, P > 0.05).
Arteriolar hyalinization. Although arteriolar thickening and hyalinization also appeared to be more pronounced in the db/db Nx than db/db sham mice (Fig. 1), this has not been further quantified. As previously described by Bower et al. (6), the periodic acid-Schiff-positive arteriolar hyaline-type thickening does have the appearance of flowing into the mesangial expansion. After comparison with the Sirius red stain, however, it seems that the mesangial expansion and arteriolar hyaline changes are separate, but geographically adjacent, processes with the same progression characteristics.
Micropuncture Data
Paired data (proximal and distal measurements from the same nephron) and data from a small number of unpaired collections are shown in Tables 2 and 3. Only paired data are shown in Table 3, and Fig. 2 shows the individual paired proximal-distal SNGFR values: 8–14 pairs per group corresponding to Table 3. Data in Table 2, derived usually from one or two nephrons per mouse, were meaned for each mouse. For each within-group comparison of these means (i.e., db/db Nx proximal SNGFR vs. db/db Nx distal SNGFR), paired t-testing was done. ANOVA and Tukey's post test were used for comparisons between the four groups. Because the data from the two SMTC-treated db/db Nx groups (groups 5 and 6) were virtually identical (not significantly different in any parameter), they were combined as a single group in Table 2 and only compared (using unpaired t-testing) with non-SMTC-treated db/db Nx mice (group 1).
|
|
|
With reference to the tubular hypothesis of diabetic hyperfiltration (see DISCUSSION), there were no significant changes in proximal fluid reabsorption between groups. Similarly, distal flow rates were unchanged, except for the significant difference between WT sham and db/db sham, where a significant increase was observed (5.2 ± 0.05 vs. 3.2 ± 0.03 nl/min, P < 0.05).
|
DISCUSSION |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
SNGFR and Proximal Tubule Fluid Reabsorption
In db/db Nx (group 1) mice, mean proximal and distal SNGFR were 3 nl higher than in the db/db sham group, but only the proximal SNGFR difference reached statistical significance using four-group post-ANOVA analysis. However, distal SNGFR was strikingly less in WT sham than WT Nx or db/db sham mice, suggesting an upward resetting of steady-state SNGFR (28).
With respect to control mechanisms possibly operative in our preparations, it is appropriate to consider whether the hyperfiltration we demonstrate is associated with changes in proximal tubule reabsorption or distal flow rates, as suggested by Thomson et al. (32). The distal SNGFR was significantly lower in WT sham (group 4) than in db/db sham (group 2) mice, as shown in Table 2. Accordingly, in comparing group 2 with group 4, in a sense evaluating the effect of DM2 on the response to the sham procedure, we document an increase in steady-state distal SNGFR without change in proximal tubular fluid reabsorption. In addition, the distal flow rate was significantly lower in WT sham (group 4) than in db/db sham (group 2) control mice, although proximal SNGFR was similar in both groups. Thus, under the free-flow conditions of our studies, we do not show an association of the rate of proximal fluid reabsorption or distal flow rates with SNGFR. Of course, to more comprehensively address the validity of the tubulocentric proposal (32) for the db/db mouse, specific studies are needed to evaluate the effects of the concurrent hyperglycemia, a reported suppressor of the Na+-glucose cotransporter (13) and downregulator of the TGF response in a rat model of DM1 (5). If feasible, measurement of PSF responses to higher-than-normal flow rates could possibly induce greater changes in proximal reabsorption and measured distal Cl– delivery. It would be also be desirable to eventually describe 1) the detailed morphology of db/db mouse proximal tubular epithelium, 2) its transapical membrane transport characteristics, i.e., Na+-glucose cotransport with and without hyperglycemia, and 3) whether ornithine decarboxylase plays a role as described in the STZ DM1 rat (31). Until other data emerge, it appears reasonable to suggest that hyperfiltration in the db/db diabetic mouse is likely sustained by a resetting of the TGF system at least in the range of zero-to-normal distal flow rates.
It is of course possible that hemodynamic or myogenic effects, rather than macula densa signaling per se, account for some of the changes we observed. A primary change in afferent arteriolar tone may occur in diabetic preparations and may account for a rightward shift in the TGF curve as described by Vallon et al. (34). Importantly, using direct videoscopic measurements, Hayashi et al. (14) demonstrated blunted myogenic responses to pressure by afferent arterioles from STZ DM1 rats. Such changes may be related to alterations in voltage-dependent Ca2+ channels and K+ channels (2).
Intrarenal NO Modulation of SNGFR
We and others have reviewed the literature relating to 1) changes in intrarenal NO metabolism in DM1 preparations, 2) how NO may attenuate intrarenal vasoconstrictor influences in diabetes, and 3) whether macula dense cells containing nNOS may be a paracrine source of NO vasodilator influences (17, 18, 29). Thus the TGF constrictor effect might be attenuated by higher levels of ambient NO in diabetes, thereby contributing to hyperfiltration. Indeed, intrarenal NO metabolism has been of interest in rodent models examining early and late effects of reduction of renal mass (30) and in the DM2 obese Zucker rat (10). In the remnant kidney, our real-time, direct tubular fluid NO concentration measurements (20) show a fourfold increase in tubular fluid NO concentration. Also, long-term effects of SMTC ingestion in the STZ DM1 rat with and without Nx have been described (16). We previously reported that, in db/db and db/m mice, an acute systemic infusion of SMTC that did not alter systemic blood pressure was associated with suppression of SNGFR (21). In the present study, we examined the possible effects of SMTC ingestion for 2–3 wk after Nx on SNGFR adaptation to Nx. Two doses of SMTC, similar to the doses used in the rat on a per-weight basis (16), in db/db Nx mice showed no effect on SNGFR or the TGF response (see RESULTS). It is possible that the effects of SMTC, although not apparent after 3 wk of ingestion, might have been evident if examined earlier. In rats, Ollerstam et al. (25) examined the effect of chronic selective inhibition of nNOS with 7-nitroindazole on the TGF system and found greater sensitivity at 1 wk but no change after 4 wk. There are also other uncertainties regarding the disposition of NOS inhibitors in vivo (9). Finally, the intrarenal metabolism and paracrine concentrations of NO may well be different in different rodent strains of DM1 and DM2. We previously showed that tubular fluid concentrations of NO in the db/db mouse model are vastly different from those in the STZ-induced diabetic mouse or the conventional STZ rat preparation (19).
The Preparation
The present study has limitations that are, in part, inseparable from its unique goal to evaluate for the first time the effects of nephrectomy on single-nephron function in this fragile db/db DM2 mouse model. In these db/db mice, hyperglycemia and hyperinsulinemia, factors already suggested to influence TGF responsiveness, are present. It is also possible that the GFR responses we measured after reduction of renal mass may have been suppressed in the more fragile hyperglycemic diabetic mice compared with WT mice. These and other issues have been reviewed by Bak et al. (3), who monitored the sequence of renal hypertrophy and hyperfiltration in the STZ diabetic rat. Other variables relate to our attempt to meticulously control for the effects of Nx. To better compare data with nondiabetic WT mice, we introduced sham controls, which entailed major surgical interventions (see METHODS). This procedure, per se, may have modified GFR or TGF responsiveness, and it is possible that other means of inducing nephron loss may differently modify GFR and TGF responses. We did not directly measure distal tubular Cl– concentrations, nor did we examine progressive PSF and TGF responses in the range of zero to higher-than-normal flow. Rather, we restricted our assessment to the normal flow range using distal SNGFR to assess so-called steady-state GFR, and we used proximal-distal SNGFR differences to assess the ambient restraining effect of the macula densa-TGF mechanism. Schnermann (28) and Vallon et al. (35) described in mice difficulties of determining, by microperfusion, effects of changes in very low loop flow rates on PSF and concluded that proximal-distal SNGFR differences reasonably assess TGF responses in the ambient flow range. Finally, it is appropriate to note that, in the present study, as in all in vivo micropuncture studies, we were restricted to evaluation of accessible superficial nephrons. In fact, juxtamedullary nephrons, with an intrinsically higher SNGFR and, reportedly, even further elevations in SNGFR in early diabetes (24), cannot be assessed in this experimental in vivo setting.
Perspectives and Significance
The worldwide epidemic of DM2 is the major cause of kidney failure requiring dialysis life support. The present study, in gene-targeted db/db obese mice with DM2, is the first to simultaneously assess progressive kidney damage along with single-nephron changes in early diabetic hyperfiltration with reduction of kidney mass by Nx. Our results show further hyperfiltration and kidney hypertrophy after Nx and evidence for an upward resetting of the control system regulating single-nephron filtration. We could not demonstrate changes in single- nephron filtration after 3 wk of ingestion of an inhibitor of NOS. At 4–6 mo after Nx, there was a marked increase in mesangial matrix expansion. Future in vivo investigations in these db/db mice should focus on single-nephron functional changes as GFR deteriorates, the effects of hyperglycemia, and factors regulating afferent arteriolar tone.
|
GRANTS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
|
REFERENCES |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Visit Other APS Journals Online |
