α1-Antitrypsin (α1-AT) deficiency causes severe liver injury in a subgroup of patients. Liver injury is thought to be caused by retention of a polymerized mutant α1-ATZ molecule in the endoplasmic reticulum (ER) of hepatocytes and is associated with an intense autophagic response. However, there is limited information about what physiologic stressors might influence liver injury. In this study, we examined the effect of fasting in the PiZ mouse model of α1-AT deficiency, because fasting is a well-characterized physiological stressor and a known stimulus for autophagy. Results show that there is a marked increase in fat accumulation and in α1-AT-containing globules in the liver of the PiZ mouse induced by fasting. Although fasting induced a marked autophagic response in wild-type mice, the autophagic response was already activated in PiZ mice and did not further increase with fasting. PiZ mice also had a significantly decreased tolerance for prolonged fasting compared with wild-type mice (PiZ mice 0% survival of 72-h fast; wild-type 100% survivial). These results demonstrate an altered response to stress in the α1-AT-deficient liver, including inability to further increase an activated autophagic response, a developmental state-specific increase in α1-AT-containing globules, and increased mortality.
- protein degradation
- endoplasmic reticulum
the classical form of α1-antitrypsin (α1-AT) deficiency, homozygous PIZZ α1-AT deficiency, is caused by a point mutation encoding a substitution of lysine for glutamate-342 (23). This substitution confers polymerogenic properties on the mutant α1-ATZ molecule (2). Aggregated mutant α1-ATZ is retained in the endoplasmic reticulum (ER) rather than secreted in the blood and body fluids where its function is to inhibit neutrophil proteases. Individuals with this deficiency have a markedly increased risk of developing emphysema by a loss-of-function mechanism, i.e., reduced levels of α1-AT in the lung to inhibit connective tissue breakdown by neutrophil elastase, cathepsin G, and proteinase 3. A subgroup of PIZZ individuals develops liver injury and hepatocellular carcinoma by a gain-of-function mechanism, i.e., accumulation of aggregated mutant α1-ATZ within the ER is toxic to liver cells. The “accumulation” mechanism is best demonstrated by transgenic mice engineered for expression of the human α1-ATZ gene (1, 4). In addition to periodic acid-Schiff (PAS)-positive, diastase-resistant intrahepatic globules that represent ER dilated with the aggregated mutant protein, these mice develop liver injury and hepatocellular carcinoma. Because there are normal levels of antiproteases in these animals, as directed by endogenous genes, the liver injury cannot be attributed to a loss-of-function mechanism. In fact, detailed histological characterization of the liver in one transgenic mouse model by Geller and colleagues (5) has shown that there are focal areas of inflammatory infiltration and regenerative activity in the form of multicellular liver plates (5). As these mice reach 18 mo of age, close to 80% of them have adenomas and carcinomas in the liver.
There is marked variability in the phenotypic expression of liver disease among affected PIZZ individuals. Some patients have severe liver disease and need liver transplantation surgery early in life, whereas others do not develop clinical signs of liver disease until late in adult life, if ever. A prospective nationwide screening study in Sweden has shown 80–90% of PIZZ individuals have no clinical evidence of liver disease at 18 years of age, their age at the time of the last report on the population (20, 21). One study (26) with genetically engineered fibroblast cell lines from PIZZ individuals carefully characterized for the absence or presence of liver disease indicates that protection from liver disease is correlated with efficient degradation of α1-ATZ in the ER and, therein, presumably with a reduced burden of aggregated protein in the ER. However, there is relatively limited information in the literature about the genetic and environmental mechanisms for inefficient degradation or for the effect of physiological stressors on susceptibility to liver injury. There is also very little information about the mechanisms responsible for the effect of developmental stage on liver injury in α1-AT-deficient patients, particularly the observation of exacerbated hepatic inflammation and dysfunction in the newborn followed by a honeymoon period until late childhood/early adolescence in many patients.
Recent studies in our laboratories have demonstrated that ER retention of mutant α1-ATZ induces a marked autophagic response in cell culture and PiZ transgenic mouse models of α1-AT deficiency as well as in the liver of PIZZ patients (24). The autophagic response is thought to be a general mechanism whereby cytosol and intracellular organelles, such as ER, are first sequestered from the rest of the cytoplasm within unique vacuoles and then degraded by fusion with lysosomes to clear the cell of senescent constituents. It occurs in many cell types during the cellular remodeling that accompanies differentiation, morphogenesis, and aging and is induced by stress states such as nutrient deprivation. In this study, we examined the effect of fasting on the liver of the PiZ mouse, because it is a well-defined physiological stimulus of autophagy and a known environmental stressor of liver disease in infants and children.
MATERIALS AND METHODS
Antibodies against α1-AT included rabbit anti-human α1-AT from DAKO (Santa Barbara, CA) and goat anti-human α1-AT from Cappel (Durham, NC). Antibodies against calnexin SPA-865 and antibodies against BiP/GRP78 were purchased from StressGen (Victoria, BC). Antibodies against lysosome-associated membrane protein-1 (LAMP-1) were purchased from Santa Cruz Biotechnology (Santa Cruz, CA), and antibodies against cathepsin were from Upstate Biotechnology (Lake Placid, NY). Rhodamine and Cy2 conjugated anti-Ig antibodies were purchased from Jackson Immunoresearch (West Grove, PA).
Liver tissue was removed by conventional surgical techniques from PiZ and C57 black mice that had been killed with anesthesia. Standardized sections were obtained from the same region of the left median lobe and fixed for light microscopic, immunofluorescent, or electron microscopic analysis as previously described (24). During fasting experiments, mice were kept in their usual cages with water available ad libitum but without any source of solid food or nutrients. An 18-h interval of fasting to study autophagy was selected on the basis of previous studies of autophagy in rodent liver (3, 8, 9, 17, 18). All studies were approved by the Washington University Animal Studies Committee.
At the time of death, liver sections were fixed in formalin, paraffin embedded, and sectioned for light and immunofluorescent microscopy using previously described techniques (24). Well-established techniques were used for immunostaining and hematoxylin and eosin (H & E), PAS/digestion, and oil red-O staining. All PAS studies were performed in parallel with nondigested controls. All double-label immunostains were repeated using different secondary antibodies with different fluorophores to confirm colocalization results. Samples for transmission electron microscopy were fixed at the time of death in 1% glutaraldehyde-0.1 M sodium cacodylate and embedded in polybed for ultrathin section transmission electron microscopy, as described previously (24). All light and fluorescent photomicrographs were obtained by viewing with a Zeiss Axioskope microscope with the Axiocam digital image capture system. Electron microscopy specimens were viewed and photographed using a Zeiss 902 transmission electron microscope. Quantification of PAS-positive globules in liver was performed by averaging the total globules counted in five randomly selected fields with a ×20 objective lens for each specimen. Quantification of autophagy was performed with grids superimposed on 10 photomicrographs of each specimen at ×3,000 and showing a complete hepatocyte with a nucleus. The area of cytoplasm occupied by autophagic vacuoles was then determined as a percentage of the total area of cytoplasm (not including the area occupied by fat droplets). Immune label electron microscopy was performed exactly as previously described (24). All studies involving human tissue were approved by the Washington University Human Studies Committee.
Histological features of PiZ transgenic mouse liver at baseline and after fasting.
First, we examined sections of liver from 3-mo-old PiZ mice at baseline and after an 18-h fast (Fig. 1). Representative photomicrographs of liver stained with H & E are shown in Fig. 1 A, left, and demonstrate the focal lymphocytic infiltration of hepatic lobules, as previously described in these mice (1). The hepatic architecture is otherwise preserved, and the hepatocytes appear intact except for round, eosinophillic bodies identifiable within the cytoplasm of some hepatocytes. These bodies were more fully examined with the PAS stain described below. Examination of the fasted liver specimens by H & E stain revealed no significant differences in the inflammatory infiltrate or overall hepatic architecture, although diffuse vacuolization was observed within the cytoplasm of all hepatocytes, suggesting the possibility of microvesicular fat deposition. Eosinophillic bodies within the cytoplasm of some hepatocytes were observed again.
Next, we examined these same specimens with PAS staining followed by digestion (Fig. 1 A, middle). This technique stains glycoprotein red and has been previously shown to label accumulation of mutant α1-ATZ within dilated ER membranes in hepatocytes (6, 12, 15, 19, 24, 28, 11). These structures have been termed “globules” when described in α1-AT-deficient liver and in the PiZ mouse liver. The presence of mutant α1-ATZ in these globules has been confirmed by immunostaining in this (see Immunofluorescence analysis for α1-AT in the PiZ mouse liver before and after fasting) and in many previous studies (6, 11, 12, 15,19, 28). The globules also immunostain positive for known ER resident proteins and immunostain negative for lysosomal markers, consistent with their identification as ER membranes and not lysosomes or autophagosomes (see Immunofluorescence analysis for α1-AT in the PiZ mouse liver before and after fasting). The results here show obvious PAS-positive globules within the cytoplasm of many but not all hepatocytes. However, examination of the liver after fasting revealed a marked increase in the number of PAS-positive globules. Quantitation of the number of globules per microscopic field in the liver of five 3-mo-old PiZ mice after 18 h of fasting revealed a 2.5-fold increase over the number observed at baseline (P < 0.05 by ANOVA). Control, C57 black mouse livers stained with PAS/digestion showed no globular accumulations of glycoprotein either at baseline or after fasting (Fig.1 A, right).
Because the H & E-stained liver sections suggested the presence of increased microvesicular fat, we also examined these same liver specimens with oil red-O stain to label fat accumulation within hepatocytes (Fig. 1 B). Nontransgenic mice of the same genetic background (C57 black) were used as control. The results show that there is an increase in fat deposition in C57 black mice after fasting for 18 h. For the PiZ mouse, there is a similar amount of fat at baseline but a much more dramatic increase after fasting. Taken together, these results indicate that there is a specific alteration in the response of the PiZ mouse liver to fasting characterized by increased PAS-positive, diastase-resistant globules, and increased steatosis.
Effect of developmental stage on the response of the PiZ mouse liver to fasting.
Next, we examined PiZ mouse liver specimens for changes in histology and α1-ATZ accumulation associated with aging. Liver specimens from at least three individual mice at ages of 1, 2, 3, 4, 6, 9, 12, and 16 mo were examined by PAS stain with digestion with representative specimens from 1-, 4-, 9-, and 16-mo-old mice shown in Fig. 2. The results show a progressive increase in the size of the globules during aging. Mean globule diameter was 1.7 ± 0.8 (SD) μm at 1 mo of age vs. 6.3 ± 3 μm at 16 mo of age (P < 0.001). There was no change in the number of globules, as determined by quantitative analysis. There was no change in the amount of fat deposition during aging.
Next, we examined the effect of fasting at different ages (Fig.3). The results show the marked 2.5-fold increase in the number of globules in 3-mo-old mice as described above but no change in 1- or 10-mo-old mice. Analysis of three individual mice at ages 1, 2, 3, 4, 6, 9, 10, and 12 mo showed that the increase in number of globules induced by fasting was only present in mice between 2 and 6 mo and peaked at 3 mo of age (data not shown). In contrast, aging had no effect on the fat deposition induced by fasting (data not shown). These results indicate that there is at least one developmental stage-specific alteration in the PiZ mouse liver at baseline, characterized by a progressive increase in the size of globules, and a developmental stage-specific alteration in the response to fasting, characterized by an increased number of globules between 2 and 6 mo of age.
We also examined 3-mo-old PiZ and C57 black mice during prolonged fasting. Groups of five mice in each category were monitored during a 72-h fast. The results showed that none of the PiZ mice (0%) were able to survive the 72-h period but that all (100%) of the C57 black mice survived without difficulty. The cause of death was not apparent. None of the mice that died during the prolonged fast had massive hepatic necrosis by routine histological examination of the livers. However, the livers were completely devoid of fat at death (data not shown), despite the marked fat accumulation noted in the PiZ mice after the overnight fasting intervals (Fig. 1.)
Immunofluorescence analysis for α1-AT in the PiZ mouse liver before and after fasting.
Next we used immunofluorescence for α1-AT to analyze the increase in globules in fasted 3-mo-old PiZ mice (Fig.4 a). The results show, at baseline, that antibody to α1-AT stains globules intensely in a manner identical to PAS staining. Moreover, antibody to α1-AT diffusely stains the cytoplasm of many, but not all, hepatocytes, in a fine reticular pattern. Positive staining with antibody to α1-AT of some, but not all, hepatocytes is a characteristic of human PIZZ and mouse PiZ liver (6, 11, 12, 15,19, 28). After fasting for 18 h, there is an increase in α1-AT-positive globules, exactly as shown with PAS staining above. Interestingly, fasting was also associated with a complete disappearance of the diffuse, reticular staining in the cytoplasm.
Next, we examined the same specimens for staining with antibody to BiP, a well-characterized resident protein of the ER (Fig. 4 A,middle). The results show, at baseline, staining of globules and diffuse staining of the cytoplasm of all hepatocytes in a fine, reticular pattern. After fasting, anti-BiP antibody exclusively stains globules. The diffuse reticular cytoplasmic staining completely disappears. This change in distribution of α1-AT and BiP staining induced by fasting was specific for the PiZ mouse, as shown by immunostaining for BiP in nontransgenic C57 black mice (Fig.4 A, bottom). Although there was a slight decrease in intensity after fasting, the pattern of diffuse, reticular staining by anti-BiP antibody was the same at baseline and after fasting in the C57 black mouse. These data indicate that there is a specific increase in globules that stain positively for α1-AT and BiP and a specific change in distribution of α1-AT and BiP staining in the PiZ mouse induced by fasting.
To be assured that the globule structures were, in fact, dilated ER and not themselves autophagosomes, we first performed double-label immunofluorescence on these same baseline and fasted PiZ liver specimens for α1-AT and cathepsin D, which is found in lysosomes and autophagosomes (Fig. 4 B, top). The results show that the globules stain positively for α1-AT but not for cathepsin D. However, many small, punctate structures consistent with lysosomes scattered throughout the cytoplasm are labeled positively for cathepsin D and not α1-AT. Occasional clusters of very small, punctate structures are double labeled for α1-AT and cathepsin D, which are consistent with previously published labeling patterns of nests of autophagosomes under immunofluorescence and immune electron microscopy (3, 18,24). Identical staining of control C57 black baseline and fasted liver specimens (Fig. 4 B, bottom) demonstrated punctate structures positive for cathepsin D but, as expected, no structures immunoreactive for human α1-AT.
Next, to determine if antibodies to α1-AT and BiP are staining the same structures, double-label immunostaining was applied to the same specimens (Fig.5 A). The results show that, at baseline, α1-AT and BiP are colocalized in globules, but there are some hepatocytes with diffuse reticular staining of the cytoplasm with anti-BiP but not anti-α1-AT, as previously described above (Fig. 5 A). After fasting, α1-AT and BiP staining is shifted to and colocalized within the globules.
We also used antibody to calnexin, another resident protein of the ER, on these liver specimens (Fig. 5 B). The results were identical. At baseline, α1-AT and calnexin are colocalized in globules, but there are some hepatocytes with diffuse reticular staining of the cytoplasm with anti-calnexin but not anti-α1-AT (Fig. 5 B). After fasting, α1-AT and calnexin are exclusively colocalized within the increased numbers of globules.
The change in distribution of α1-AT, BiP, and calnexin immunoreactivity was seen in PiZ mice at all ages from 1 to 16 mo (data not shown).
Ultrastructural analysis of the PiZ mouse liver before and after fasting.
To determine if the disappearance of the diffuse reticular cytoplasmic immunostaining with anti-α1-AT, anti-BiP, and anti-calnexin antibodies in PiZ mouse liver after fasting is because of a disappearance of nondilated ER, we examined the liver of the fasted PiZ mouse by electron microscopy (Fig.6). The results show that there are abundant nondilated ER cisternae throughout the cytoplasm in addition to massively dilated globules and many fat droplets. These results indicate that the disappearance of the diffuse reticular cytoplasmic immunostaining for α1-AT, BiP, and calnexin is explained by a change in the distribution of these molecules rather than a complete transformation of ER membranes.
Finally, we examined the ultrastructure of hepatocytes in PiZ mice using transmission electron microscopy to determine the effect of fasting on autophagy. We examined the same baseline and fasted C57 black and PiZ liver specimens as described above by transmission electron microscopy. Representative photomicrographs are shown in Fig. 7. At baseline, the nontransgenic C57 black mouse hepatocytes show normal structures, including a nucleus, mitochondria, and ER membranes. Some of these hepatocytes have one or two double or multilammelar vacuoles containing electron-dense material, which are the classical morphological characteristics of autophagosomes (Fig. 7, Aand B). When the C57 black mouse is fasted 18 h, there is a considerable increase in the autophagic vacuoles visible in the cytoplasm, consistent with previous studies that have demonstrated increased autophagy during fasting (3, 8, 9, 17, 18). Fasting is also associated with an increase in scattered, small, simple cytoplasmic vacuoles containing material that has the density of lipid, consistent with mild-to-moderate microvesicular fat accumulation described above in the C57 black mouse. There is little if any change in the architecture of the ER detectable after fasting.
In the PiZ mouse liver, there are already a large number of autophagosomes at baseline, which can be identified by the identical ultrastructural criteria as autophagosomes in the C57 mice. In fact, the electron photomicrographs shown in Fig. 7 A suggest that there are significantly more autophagosomes in the PiZ mouse liver at baseline than in the C57 black mouse liver after fasting. Interestingly, there is no increase in the number of autophagosomes when comparing PiZ mouse liver after fasting with that at baseline. The increase in globules and fat vesicles is readily apparent and intervenes between autophagosomes, allowing fewer to be visible in any single field of view. At higher magnification in Fig. 7 B, the morphology of the autophagosomes is even more clearly evident. Again, there is an increase in the number of autophagosomes after fasting in the C57 black mouse. There are many more autophagosomes in the PiZ mouse at baseline without any increase after fasting. To provide further assurance that these multilamellar vacuoles containing electron-dense debris were indeed autophagosomes, we performed immune-label electron microscopy on the PiZ liver. Antibody against the lysosomal membrane protein, LAMP-1, and the lysosomal hydrolase, cathepsin D, both of which are thought to be present within autophagosomes, were used, followed by secondary anti-Ig antibody bearing immunogold beads (3, 8, 17, 18, 24). The result in Fig. 7 C shows that the multilamellar, electron-dense structures within PiZ mouse liver label positively for both of these proteins.
Quantitative morphometric analyses of the autophagosomes in the liver cells at this steady state are shown in Fig.8. The area of cytoplasm, not including fat droplets, occupied by autophagosomes was determined in 10 hepatocytes from four different mice in each category. The results show that there is a significant increase in autophagic vacuoles at steady state mediated by fasting in the C57 black mouse, reaching 1.5% of the area of the cytoplasm. Autophagosomes occupy 2.4% of the cytoplasm in the PiZ mouse at baseline, with no significant change during fasting, indicating that there is a marked constitutive increase in steady-state autophagic vacuoles in the PiZ mouse and an inability to mount an augmented autophagic response to fasting. Because there is no way to separately determine the rate of formation of autophagosomes and the rate at which autophagosomes fuse with lysosomes, it is not possible to exclude the possibility that fasting causes an increase in both formation and clearance of autophagosomes in the α1-AT-deficient liver in such a way that there is no net change in steady-state levels of autophagosomes morphologically.
In homozygous PIZZ α1-AT deficiency, a point mutation confers polymerogenic properties on the mutant α1-ATZ molecule (2). This mutant protein is retained in the ER rather than secreted. Most of the evidence in the literature suggests that chronic liver injury and hepatocellular carcinoma in some PIZZ individuals are the result of hepatotoxic effects of ER α1-ATZ retention. However, it has been difficult to explain why as many as 80–90% of PIZZ individuals escape liver injury or have much milder liver injury detected incidentally at autopsy. It has also been difficult to explain why the liver disease can become apparent early in life in some patients and then enter a honeymoon period compared with other patients in which liver disease is only discovered much later. There have been very few studies of potential genetic traits and/or environmental factors that provide an explanation for this variability in liver disease phenotype and in natural history. In one study using a genetic complementation approach to express the mutant α1-ATZ protein in cell lines from PIZZ individuals with or without known liver disease, we found that inefficient degradation of α1-ATZ in the ER correlated with susceptibility to liver disease (26). However, it is still unknown whether inefficient ER degradation is determined by genetic or environmental mechanisms, and there is no information in the literature about how environmental factors may interact with genetic determinants of susceptibility to liver disease.
To begin to examine these issues and the effect of environmental conditions and developmental stage on susceptibility to liver injury in α1-AT deficiency, in this study we used the genetically engineered PiZ mouse as an in vivo model and fasting as a model environmental stressor. Fasting is a well-characterized stressor of liver disease in infants and in experimental models of liver disease and is particularly informative for α1-AT deficiency because it is one of the best studied stimuli of the autophagic response. Many previous studies have shown an increase in autophagic activity in the liver during nutrient deprivation (3, 8, 9, 17,18). Our recent morphological studies have shown that retention of mutant α1-ATZ in the ER is associated with an autophagic response in cell culture and transgenic mouse model systems as well as in human α1-AT-deficient liver (24).
The results show that autophagy is constitutively activated, as indicated by a steady-state increase in autophogic vacuoles, in the PiZ mouse at baseline. Morphometric analysis suggests that the number of autophagosomes in liver cells of the PiZ mouse at baseline is >50% higher than that in liver cells of the C57 black mouse after stimulation by fasting. In contrast to the C57 black mouse, fasting does not lead to an increase in autophagosomes in the liver of the PiZ mouse. These results therefore provide further evidence for the integral relationship between the pathological effects of α1-AT deficiency and the autophagic response of the host. If the autophagic response is intended to serve a protective role, by clearing ER distended with aggregated mutant α1-ATZ molecules or by suppressing tumorigenesis (7, 10), then our data indicate that the α1-AT-deficient liver has little reserve and would probably be easily overwhelmed by physiological and pathological stressors. Furthermore, the consequences of constitutive high-level activation of autophagy on the liver are entirely unknown. From our search of the literature, the only other condition in which there is accumulation of autophagic vacuoles under homeostatic conditions is Danon disease (14, 22). In contrast to α1-AT deficiency, however, autophagosomes accumulate in Danon disease because of a genetic defect in the terminal phases of autophagy, i.e., the fusion of autophagic vacuoles with lysosomes and subsequent degradation within autolysosomes (14,22). It should be noted, however, that currently available methodology does not permit us to exclude the possibility that there is no net change in steady-state levels of autophagosomes in the α1-AT-deficient liver during fasting because there are increases in both formation and clearance of these vacuoles. The results of the current studies indicate that the α1-AT-deficient liver is also susceptible to physiological stress, as evidenced by increased mortality and dysregulation of lipid metabolism. It is well known that hepatic steatosis often follows fasting in wild-type mice, but there was a massive increase in fat deposition in the fasting PiZ mouse noted in this study. Recent studies have shown that there is increased hepatic biogenesis, uptake of cholesterol and triglycerides, and hepatic steatosis as a result of the increased sterol regulatory element-binding protein signaling that accompanies the unfolded protein response (16, 25). It is not uncommon to see hepatic steatosis in patients with α1-AT deficiency being evaluated for exacerbation of liver disease.
One of the other unexpected series of findings reported here is the increase in size of α1-AT-containing globules during aging and the increase in the number of α1-AT-containing globules induced by fasting in the PiZ mouse at 2–6 mo of age. Electron microscopic examination shows that these globules represent portions of the ER because they are studded with ribosomes and contiguous with normal-appearing ER. Moreover, these globules stain positively for the resident ER proteins BiP and calnexin but not for lysosomal enzymes. It is completely unknown at this time why the size of these globules increases during aging and why the number of globules increases during fasting only in PiZ mice at the 2- to 6-mo period of development. Immunofluorescent studies shown in Figs. 4 and 5 also show that there is a change in distribution of α1-AT, BiP, and calnexin exclusively in the globules after fasting. The disappearance of BiP and calnexin immunoreactivity from the diffuse reticular network of the cytoplasm after fasting could not be attributed to a disappearance of nondilated ER, because there was still abundant nondilated ER in liver cells evident by electron microscopy. Perhaps these globules represent specialized subdomains of the ER in which polymerized α1-ATZ, together with chaperones such as BiP and calnexin, accumulates with an exceptionally long half-life. In fact, the redistribution of BiP observed here during fasting bears a remarkable resemblance to that described when ER-to-Golgi transport is inhibited experimentally in yeast (13, 27) and attributed to the need for chaperones to localize at ER exit sites where cargo molecules or misfolded substrates accumulate.
Current address for D. Perlmutter: Dept. of Pediatrics, University of Pittsburgh School of Medicine and Children's Hospital of Pittsburgh, Pittsburgh, PA 15213.
Address for reprint requests and other correspondence: J. H. Teckman, Dept. of Pediatrics, Washington Univ. School of Medicine, 660 S. Euclid Blvd., Campus Box 8208, St. Louis, MO 63110 (E-mail:).
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.
May 29, 2002;10.1152/ajpgi.00041.2002
- Copyright © 2002 the American Physiological Society