| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Division of Pulmonary and Critical Care Medicine, The Johns Hopkins Medical School, Baltimore Maryland; and Division of Pulmonary, Allergy, and Critical Care Medicine, University of Pittsburgh Medical School, Pittsburgh, Pennsylvania
Correspondence and requests for reprints should be addressed to Christopher P. O'Donnell, Ph.D., Division of Pulmonary, Allergy, and Critical Care Medicine, University of Pittsburgh Medical School, NW 628 UPMC Montefiore, 3459 Fifth Avenue, Pittsburgh, PA 15213. E-mail: odonnellcp{at}upmc.edu
ABSTRACT
The technologies of genomics and proteomics are powerful tools for discovering novel gene and protein expression responses to disease. Considerable evidence indicates that a genetic basis exists to the causes of sleep-disordered breathing, in particular its most common form of obstructive sleep apnea (OSA), which is characterized by periods of intermittent hypoxia and disrupted sleep. However, the genetic contribution to the pathogenesis of OSA has largely been determined using traditional genetic approaches of family, twin, and linkage studies in clinical populations and quantitative trait loci and targeted gene procedures in animal models of OSA. In contrast to the pathogenesis of OSA, the consequences or sequelae of OSA are highly amenable to genomic and proteomic approaches. Animal studies have assessed changes in gene and protein expression in multiple organ systems in response to intermittent hypoxia and sleep deprivation and uncovered novel gene activation paradigms. The first tentative steps have been made toward applying proteomic analyses of blood and urine from patients with OSA as a potential screening tool for diagnosis in the clinical setting. It is anticipated that genomic and proteomic technologies will become increasingly used in the area of OSA with the unprecedented access to tissue in procedures such as bariatric surgery. OSA represents a severe insult to the oxygenation of tissues and the homeostasis of sleep, and genomic and proteomic approaches hold promise for defining previously unexplored mechanisms and pathways that lead to downstream pathologies, including hypertension, insulin resistance, and neurocognitive dysfunction.
Key Words: hypertension • insulin resistance • intermittent hypoxia • obstructive sleep apnea • proteomics
Sleep-disordered breathing covers a spectrum of syndromes in which ventilation is inappropriately reduced during sleep. These syndromes include hypoventilation, Cheyne-Stokes respiration, central sleep apnea, and obstructive sleep apnea (OSA), and are characterized by a range of etiologies and physiologic consequences. For the purposes of this review, we will focus on the most common and well-studied syndrome of OSA.
In its simplest form, OSA is a reduction (hypopnea) or complete cessation (apnea) in airflow due to obstruction of the upper airway after the initiation of sleep (1). Because the obstruction is most often relieved by arousal from sleep, the syndrome is characterized by repetitive brief episodes of airway obstruction during sleep, separated by periods of unobstructed breathing that are initiated by arousal (2). However, it is possible for neural reflex responses to restore airway patency in the absence of a discernable change in sleep/wake state. The reflex resolution of airway obstruction without arousal is more common in pediatric compared with adult OSA populations (3). A major limitation to the systematic study of OSA is the determination of what constitutes a clinically significant degree of apnea. The most common metric used to characterize the severity of OSA is based on nasal flow assessments of the number of apneas and hypopneas that occur per hour (apnea–hypopnea index). The degree of hypoxic stress resulting from OSA can also be determined by the changes in arterial oxygen desaturation that occur throughout the night. Consequently, with regard to studies of the genetics, genomics, and proteomics of OSA, either the apnea–hypopnea index or hypoxic stress is the predominant outcome used to define the presence versus absence, or the severity, of exposure.
OSA is prevalent in the general population, with estimates ranging from 4 to 24% for men and 2 to 9% for women (4). The two major risk factors for OSA are male sex and adiposity (4–6), with the incidence of OSA paralleling the recent rise in obesity in Western society (7). Research in OSA has also increased dramatically in the last decade as scientists struggle to define the syndrome, grapple with its underlying etiology, and determine the nature and scope of its pathologic consequences. It is perhaps not surprising that given OSA's impact on two of the most fundamental physiologic needs—oxygen and sleep—that a diverse range of pathophysiologic outcomes have been independently linked to the syndrome. In recent years, the techniques of genetics, genomics, and proteomics have been applied to the study of OSA, and in the following sections, we review what has been learned from these studies.
For the purposes of this review, we will determine how genetic, genomic, and proteomic studies have provided new insights into the (1) pathogenesis, (2) expression, and (3) sequelae of OSA (Figure 1). The pathogenesis of OSA includes factors such as obesity, craniofacial abnormalities, and deficits in central respiratory control of upper airway muscles that cause airway obstruction during sleep. The expression of OSA relates to factors that impact on the rate and duration at which a collapsible airway intermittently obstructs, including arousal thresholds and hypoxic and hypercapnic sensitivities. The sequelae of OSA relate to cardiovascular, metabolic, neurologic, and other outcomes, and are the most studied aspect of OSA and also the most amenable to genomic and proteomic analyses.
|
PATHOGENESIS OF OSA
Several genetic epidemiologic studies have reported a familial aggregation with OSA (8–13), although the genetic basis for airway collapse during sleep is not clearly defined. The pathogenesis of OSA is likely dependent on independent or interactive effects of mechanical and neural deficits in airway control (2, 14–16). This is based on the fact that obesity in adults and tonsils in children are major risk factors for OSA (4, 17), yet not all obese people have OSA and not all normal-weight people can sleep with unobstructed airways. Craniofacial defects provide the potential for a mechanically based pathogenesis. Evidence from twin studies suggests a significant genetic component to cephalometry. Familial studies in the United Kingdom and the United States have also demonstrated that OSA is associated with structural features of the upper airway (11, 18, 19), with soft tissue structures being particularly important (20). In addition, differences in upper airway structural properties may account for sex differences that make males more susceptible to OSA than females (21).
Much less is known about the genetic basis for impaired central control of upper airway muscle tone during sleep, presumably because of the greater difficulty defining the functional properties of the airway during sleep compared with the structural properties. Studies in rats and dogs demonstrate that central serotinergic and adrenergic pathways are important in the maintenance of upper airway and diaphragmatic control during sleep (22–26). Genetic studies on inbred mice indicate that strain can significantly influence the pattern of central ventilatory control during wakefulness and sleep (27–29) and affect lung volume (30), both of which may have the potential to impact on the collapsibility of the upper airway. In addition, individual genes have been identified that impact on central control of ventilation. The homeobox gene PHOX2B has been shown to play an essential role in the normal patterning of breathing, and defects are associated with the congenital central hypoventilation syndrome (31). In mice, leptin deficiency causes obesity hypoventilation, and the defect can be corrected with acute leptin replacement, independent of weight changes (32). Leptin provides an interesting example of the potential for a single gene to influence the pathogenesis of OSA through both the mechanical stresses of adiposity and direct action on neural pathways controlling ventilation. The concept of "leptin resistance," the inability of leptin to appropriately activate leptin receptor signaling, is associated with obesity (33–35) and impaired ventilatory control (32). There is also evidence that leptin resistance in the obese Zucker rat is associated with a more collapsible upper airway, although the relative impact of adiposity versus central respiratory control in mediating the increased collapsibility has not been delineated (36). Several other obesity-related genes, including POMC, COH1, SLC6A14, PPARG, UCP, and APOe, have been recently proposed as playing a potential role in OSA through regulation of appetite, fat deposition, and thermogenesis (37). Thus, there are a number of obesity-related genes that have the potential to impact on the pathogenesis of OSA.
The techniques of genomics and proteomics have not been systematically applied to the investigation of the pathogenesis of OSA (Figure 1). The most obvious limitations arise from determining what tissue to sample for mRNA or protein expression and how to interpret the findings. Possible tissue sources include blood, urine, fat/muscle/liver tissue (bariatric surgery patients), or potentially upper airway structures removed during surgery. Apart from blood and urine, it is difficult to obtain serial samples on these or other tissue depots in patients with OSA. Finally, there is the question of whether any changes in mRNA or protein expression between patients versus controls, or in patients over time or after treatment, represent a cause or consequence of OSA. As a result, many factors have limited the utility of genomic/proteomic approaches to the study of the pathogenesis of OSA.
EXPRESSION OF OSA
The physiologic response to collapse of the upper airway during sleep is dependent on multiple factors, several of which are likely to be under genetic control. For example, the duration over which an apnea lasts after the airway has collapsed is in part dependent on the sensitivity to the rising levels of hypoxia and hypercapnia and their ability to trigger arousal and restore a wakefulness tone and patency to the upper airway. Studies in inbred mice with genetic differences in the structural and functional properties of the carotid body (major sensor of systemic hypoxia) demonstrate that the duration and the rhythmicity at which sleep-induced hypoxic events occur spontaneously is under genetic control (38). The A/J strain of mice, which has smaller, less functional carotid bodies than the DBA/2J strain of mice, exhibits higher rates and longer durations of sleep-induced hypoxia. Moreover, this genetic difference in expression of sleep-induced hypoxia between the strains was not due to an overall difference in arousability but rather was specific to hypoxia-induced arousal. Similarly, others have shown in inbred mice that a genetic component exists to the ventilatory response that occurs with hypoxia, hypercapnia, and hyperoxia (27–29). In fact, specific loci for hypoxic and hypercapnic responsiveness have been mapped to chromosomes 1, 5, and 9 in mice, and potential candidate genes, including NOS1 and the ß2-adrenergic receptor, have been proposed (39–41). Thus, animal studies provide evidence that genetic factors impact on chemoresponsiveness and the ability of hypoxia to induce arousal from sleep.
Humans exhibit a marked variability in their ventilatory responsiveness to hypoxia and hypercapnia (42). There are data from studies in high-altitude native populations (43), adult twin studies (44), newborn twin studies (45), and a three-generation family study (46) that demonstrate that a significant genetic component may contribute to the variability in ventilatory responsiveness to hypoxia. Although several studies have suggested a genetic basis also exists for ventilatory responsiveness to hypercapnia (44, 47, 48), the evidence is less compelling than for hypoxia. To our knowledge, there are no human studies that have determined whether arousability or, more importantly, arousal thresholds to hypoxia or hypercapnia have a heritable or genetic component. Consequently, the evidence that genetic factors can influence the expression of sleep apnea in humans is considerably more limited than from studies in animals. Similar to the pathogenesis of sleep apnea described above, the techniques of genomics or proteomics have not been applied to differential phenotypes of hypoxic or hypercapnic sensitivity or arousal thresholds (Figure 1).
SEQUELAE OF OSA
Although the use of proteomics/genomics is limited in studying the pathogenesis and expression of OSA, the same is not true for the sequelae of OSA (Figure 1). Although the pathogenesis of OSA is dependent on collapsible structures in the upper airway, the consequences or sequelae of OSA are likely to impact on multiple organ systems. As mentioned above, OSA results in disturbances of two major homeostatic systems—oxygenation and sleep—and is known to cause disruption of cardiovascular, metabolic, and endocrine function (6, 49–56). However, very few studies have attempted to harness genomic/proteomic technology to study the sequelae of OSA, and most studies have been conducted in animal models, as detailed below.
Polotsky and coworkers (57) conducted the first use of high-throughput microarray mRNA expression analyses on subcutaneous white adipose tissue from lean and genetically obese ob/ob mice in an animal model of exposure to intermittent hypoxia (IH). The 5-d period of IH was achieved by lowering inspired oxygen levels from room air to 5% over 30 s and reoxygenating back to room air over the subsequent 30 s, providing a rate of one IH event per minute, which was sustained for 12 h throughout the light or sleeping phase only. The study used an earlier generation of a murine Affymetrix chip (U74A; Affymetrix, Santa Clara, CA) that contained 6,000 sequences, and the study was limited by the pooling of mRNA samples from five mice in the IH-exposed group for comparison to five samples pooled from the control or intermittent air group. Using a simple twofold cutoff criterion, lean mice exhibited up-regulation of 6 genes and down-regulation of 47 genes, and obese mice exhibited up-regulation of 12 genes and down-regulation of 3 genes. Given that the focus of the study was on glucose metabolism, the authors reported that, in lean mice, leptin was significantly up-regulated, whereas phosphotidylinositol-3-kinase (PI3K), caveolin 3, and insulin-degrading enzyme were among genes that were down-regulated. In obese mice, caveolin 3 was up-regulated and uncoupling protein 1 was down-regulated. Interestingly, the study did not show changes in the expression of other genes controlling insulin signaling or insulin resistance. However, because the mRNA was pooled across animals, the absence of expression in insulin-related pathways should be interpreted with caution.
In a later study (58), the same authors used a more sophisticated approach to assess the effects of a chronic 12-wk exposure of IH on hepatic gene transcription in genetically obese ob/ob mice. Microarray analysis was performed on individual animals (five for IH and five for control) using the Mouse Genome 430A 2.0 Affymetrix chip with 22,626 sequences. Using Significance Analysis of Microarrays (SAM) software and fold changes of more than 1.2 and less than 0.8, 135 genes were identified as up-regulated and 11 genes were down-regulated. Clusters of genes were linked to gene ontologies, and the predominance of up-regulated genes represented pathways of lipid and carbohydrate metabolism, whereas genes of proteolysis and peptidolysis were significantly down-regulated. Using an even more restrictive SAM analysis (> 2.0, < 0.5), there was still consistent up-regulation of key genes controlling lipid biosynthesis, including acetyl–coenzyme A carboxylase, fatty acid desaturase, and mitochondrial glycerol-3-phosphate acyltransferase (GPAT), all of which are regulated by the nuclear transcription factor sterol regulatory element binding protein-1 (SREBP-1). A subsequent study in lean mice exposed to IH determined that elevated protein levels of SREBP-1 in the liver were associated with increased serum cholesterol, phospholipids, and triglycerides, as well as increased liver triglyceride content (59). The most recent work from the same group suggests that up-regulation of the SREBP-1 pathway in response to hypoxia is mediated via hypoxia-inducible factor 1 (60). Thus, microarray studies have identified a discrete number of genes important in lipid and carbohydrate metabolism that are up-regulated in response to IH, laying the framework for future, more targeted interventional studies.
A recent study by Fan and colleagues (61) assessed mRNA expression by microarray analyses in hearts of mice exposed to IH for periods of 1, 2, and 4 wk using four individual arrays per group and per time point. An interesting aspect of the study was that hearts from animals exposed to IH were not only compared with control mice but also with mice exposed to continuous hypoxia (CH; 11% inspired oxygen). The IH cycled between room air and 11% over 4 min and then returned to room air over 4 min, providing one hypoxic event per 8 min. In addition to the slow cycling time, the IH stimulus was continued for 24 h/d, making the IH paradigm difficult to compare with previous studies. Last, the studies were conducted in neonatal mice on the second day after birth and therefore the results are unlikely to be applicable to adult animals.
The mRNA expression was determined for 7,455 distinct genes and detection of significantly regulated genes was determined on the basis of fold changes in expression combined with t test Bonferroni-type adjustment (61). CH caused an up-regulation of 549 genes and down-regulation of 375 genes, whereas IH caused up-regulation of 294 genes and down-regulation of 440 genes, and the largest number of gene expression changes occurred at the 2-wk time point. The most notable phenotypic finding of the study was that exposure to CH caused a far greater increase in heart weight than IH. The heart weight only increased in IH animals compared with control animals if heart weight was corrected for body weight, which increased at a slower rate in IH mice compared with control or CH mice. The authors highlight the expression profile of the eurkaryotic translation initiation factor-4E, an enhancer of protein synthesis, because it was significantly up-regulated in CH, but down-regulated in IH, putatively impacting on the hypertrophic response of the heart to CH, but not IH. Signaling pathways implicated in cardiac hypertrophy, such as members of the mitogen-activated protein kinase (MAPK) family, were consistently up-regulated only in response to CH. The authors' postulate that the lack of a robust hypertrophic response in hearts of animals exposed to IH may be related to down-regulation of subunits in the mitochondrial complex I, potentially impairing mitochondrial function and limiting energy supply to the heart. The important conclusion from the study is that exposure to CH and IH have very distinct, yet diverse, effects on the profile of gene expression in the neonatal heart.
OSA is characterized by fragmentation of normal sleep, potentially impacting on gene expression in the brain. Although no study has specifically characterized the effects of OSA-related sleep fragmentation on the brain, a series of studies from the laboratory of Cirelli and Tononi (62–64) have detailed the changes in gene expression from approximately 10,000 transcripts in the cerebral cortex of rats during sleep and wakefulness and in response to sleep deprivation. In response to 3 h of wakefulness or sleep deprivation, there was activation of nuclear immediate early genes or transcription factors, such as c-fos, and mitochondrial genes, including subunit 1 of cytochrome C oxidase, which potentially are responding to the increased metabolic demand of wakefulness or sleep deprivation. During more extended periods of wakefulness or sleep deprivation (8 h), a completely separate group of genes related to metabolism (e.g., Glut 1) and molecular chaperones or heat shock proteins (e.g., HSP60) were up-regulated. Interestingly, the changes in gene expression were dependent on an intact noradrenergic system, because unilateral lesioning of the left locus coeruleus abolished changes in gene expression in the left cerebral cortex, whereas overall sleep architecture and elevated gene expression in the right cerebral cortex were unaffected (63). It remains to be seen whether the sleep fragmentation of OSA can also impact on the expression of early immediate genes, mitochondrial genes, or genes involved in molecular chaperoning and metabolism.
The technique of proteomics has also been applied to determining what proteins demonstrate increased rates of expression in response to sleep deprivation in the basal forebrain of rats (65). Two-dimensional gel electrophoresis and matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI TOF MS) were used to separate and identify proteins. Total protein samples were pooled for the control and sleep-deprived animals and differences were determined by fold changes in the integrated density of protein spots. A total of 969 protein spots were identified, of which 89 spots showed a more than twofold difference in response to sleep deprivation. Eleven proteins that were altered by threefold or more in response to sleep deprivation were categorized as cytoskeletal or synaptic proteins, including elevated protein levels of 
-tubulin and Rho A GTPase and reduced protein levels of neuromodulin and vesicle-associated membrane protein-2. The authors concluded that impaired learning and memory associated with sleep deprivation might be dependent on altered protein levels impacting on structural/synaptic plasticity in the basal forebrain.
OSA, similar to sleep deprivation, causes deficits in cognition, memory, and learning (66, 67). A rat model of 10% nadir IH also exhibits neurocognitive defects (68–72), with cellular changes occurring in hypoxic-sensitive areas in the cortex and the CA1 region of the hypothalamus, but not the hypoxic-insensitive CA3 region of the hippocampus. Klein and colleagues and Gozal and colleagues performed a proteomic assessment of the CA1 and CA3 hippocampal regions of rats exposed to 6 h of IH (73, 74). Similar procedures of 2D gel electrophoresis and MALDI TOF MS analysis as described above were used to determine that, of 99 identified proteins, 32 were up-regulated by IH in the CA1 region compared with the CA3 region of the hippocampus. Data were analyzed using a combination of fold changes and t tests from paired animals exposed to room air or IH and in CA1 or CA3 regions from the same animal. Proteins up-regulated by IH in the CA1 region fell into three distinct functional groups of cytoskeleton-related, metabolic, and molecular chaperone survival proteins. This pattern of protein expression from the CA1 hippocampal region in rats exposed to IH is consistent with the genomic and proteomic expression profiles described above in sleep-restricted animals, which showed increased cytoskeletal protein levels in the basal forebrain (64) and increased mRNA expression profiles of metabolic and molecular chaperone proteins in the cerebral cortex (65).
Changes in renal protein levels in response to IH were also investigated by Klein and colleagues (75) using a similar proteomic approach as described above. Studies were conducted in rats exposed to IH (nadir, 10% inspired oxygen for 12 h/d during light phase) or chronic hypoxia (constant 10% inspired oxygen) for either 14 or 30 d, with significant hypertension resulting only in animals exposed to IH for 30 d. Of 248 proteins analyzed, IH at 30 d was shown to significantly increase all five forms of kallistatin, a potent vasodilator, whereas all five forms of 1-antitrypsin precursor, an inhibitor of kallikrein activation, exhibited significantly lower protein levels. The authors went on to demonstrate that transgenic hKLK1 rats that overexpress human kallikrein were protected from the development of hypertension after 30 d of exposure to IH. Thus, a proteomic approach was able to identify putative proteins associated with the development of hypertension in response to IH, and subsequent targeted interventional studies using transgenic animals verified the causative role of the kallistatin–kallikrein pathway in IH-mediated hypertension.
To our knowledge, only two studies have used a genomic or proteomic approach in the clinical arena (76, 77). Both studies come from the laboratory of Gozal and colleagues and involve serum and urine proteomic approaches in pediatric OSA with a focus on exploring potential disease protein fingerprints for diagnostic purposes. The first study compared morning serum protein expression profiles using surface-enhanced laser desorption/ionization TOF MS (SELDI-TOF MS) in 20 children with OSA with 20 children who snored but did not have OSA (76). Three proteins were identified based on molecular mass that predicted the presence of OSA with 93% sensitivity and 90% specificity. In a second study, morning urinary protein expression profiles were compared between 11 children with OSA and 11 control children using 2D gel electrophoresis and MALDI TOF MS analysis (77). Of 100 potential proteins, a group of four differentially expressed proteins were positively identified as albumin, immunoglobulin, gelsolin (acting binding protein), and perlecan (heparin sulfate proteoglycan). These studies represent the first tentative step in determining, from a relatively small pool of proteins, whether pediatric OSA has a distinct proteomic profile. Future studies expanded to larger populations, using increased numbers of proteins, and with potentially more severe forms of obesity-associated adult OSA may provide unique OSA proteomic fingerprints. The ability to screen for the presence of OSA based on a proteomic profile obtained from blood or urine would represent a major clinical advance in a field where diagnostic studies are expensive and time-consuming and large numbers of patients remain undiagnosed.
In addition to future studies using genomics and proteomics in the clinical OSA setting, there is also the possibility to apply these techniques using more basic science approaches. For example, correlates of sleep and wake have been determined in Drosophila as well as methods developed for exposure to IH (78, 79). In cell culture, technical challenges have been overcome and cells can be exposed to IH at rates relevant to clinical OSA (80), allowing detailed interrogation of intracellular signaling pathways (81, 82). Thus, considerable opportunities exist to apply proteomics and genomic technologies across multiple paradigms of OSA from the clinical arena on one hand to cell culture models of IH on the other (Figure 1).
CONCLUSIONS
Genomic and proteomic approaches in OSA have been limited to a few studies in animal models and the efforts of one group to characterize protein expression in a pediatric OSA population. It is unlikely that the genetic basis underlying the pathogenesis and expression of OSA will be significantly advanced by the technologies of genomics and proteomics. Rather, more traditional approaches of human and animal genetics and targeted gene technologies hold the most promise for exploring the "causes" of OSA. However, the consequences or sequelae of OSA are highly amenable to genomic and proteomic approaches, and their application in animal models of IH or sleep deprivation has provided novel information about gene activation in the brain, heart, liver, fat tissue, and kidney. In the clinical arena, genomic and proteomic approaches hold promise as a screening tool for diagnosis of OSA using blood or urine. Studies in patients undergoing bariatric surgery provide the prospect of unprecedented access to fat, liver, and muscle tissue and the opportunity to assess changes in gene and protein expression resulting from both obesity and OSA.
FOOTNOTES
Conflict of Interest Statement: Neither author has a financial relationship with a commercial entity that has an interest in the subject of this manuscript.
(Received in original form June 13, 2006; accepted in final form July 7, 2006)
REFERENCES
This article has been cited by other articles:
|
|
 |
|
  R. L. Riha, T. Gislasson, and K. Diefenbach The phenotype and genotype of adult obstructive sleep apnoea/hypopnoea syndrome Eur. Respir. J., March 1, 2009; 33(3): 646 - 655. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 W. Wu, N. B. Dave, G. Yu, P. J. Strollo, E. Kovkarova-Naumovski, S. W. Ryter, S. R. Reeves, E. Dayyat, Y. Wang, A. M. K. Choi, et al. Network analysis of temporal effects of intermittent and sustained hypoxia on rat lungs Physiol Genomics, December 12, 2008; 36(1): 24 - 34. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. J. Eckert and A. Malhotra Pathophysiology of Adult Obstructive Sleep Apnea Proceedings of the ATS, February 15, 2008; 5(2): 144 - 153. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L.-M. Chen, I. Choi, G. G. Haddad, and W. F. Boron Chronic continuous hypoxia decreases the expression of SLC4A7 (NBCn1) and SLC4A10 (NCBE) in mouse brain Am J Physiol Regulatory Integrative Comp Physiol, December 1, 2007; 293(6): R2412 - R2420. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. Parati, C. Lombardi, and K. Narkiewicz Sleep apnea: epidemiology, pathophysiology, and relation to cardiovascular risk Am J Physiol Regulatory Integrative Comp Physiol, October 1, 2007; 293(4): R1671 - R1683. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Barratt MB ChB (Ho and P. Topham MB ChB MD Urine proteomics: the present and future of measuring urinary protein components in disease Can. Med. Assoc. J., August 14, 2007; 177(4): 361 - 368. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
