Corresponding author: J. De Loor, Laboratory of Biochemistry, Department of Pharmacology, Toxicology and Biochemistry, Faculty of Veterinary Medicine, Ghent University, Salisburylaan 133, 9820 Merelbeke, Belgium; e-mail: email@example.com
Routinely, kidney dysfunction and decreased glomerular filtration rate (GFR) are diagnosed by the evaluation of changes in the serum creatinine (SCr) and blood urea nitrogen (BUN) concentrations. However, neither of these tests is sensitive or specific enough for the early diagnosis of impaired kidney function because they are both affected by other renal and nonrenal factors. Furthermore, kidney injury can be present in the absence of kidney dysfunction. Renal reserve enables normal GFR even when nephrons are damaged. Renal biomarkers, especially those present in urine, may be useful for the study of both acute and chronic nephropathies. The aim of this review is to describe the current status of urinary biomarkers as diagnostic tools for kidney injury in dogs with particular focus on acute kidney injury (AKI). The International Renal Interest Society (IRIS) canine AKI grading system and the implementation of urinary biomarkers in this system also are discussed. The discovery of novel urinary biomarkers has emerged from hypotheses about the pathophysiology of kidney injury, but few proteomic urine screening approaches have been described in dogs. Lack of standardization of biomarker assays further complicates the comparison of novel canine urinary biomarker validation results among studies. Future research should focus on novel biomarkers of renal origin and evaluate promising biomarkers in different clinical conditions. Validation of selected urinary biomarkers in the diagnosis of canine kidney diseases must include dogs with both renal and nonrenal diseases to evaluate their sensitivity, specificity as well as their negative and positive predictive values.
Despite its questionable sensitivity for the diagnosis of early acute kidney injury (AKI), glomerular filtration rate (GFR) is still considered a sensitive index of functional nephron mass. It is also the main determinant of the excretion rate of plasma components that are nonanionic and small enough to pass the glomerular barrier and undergo minimal reabsorption and secretion in the more distal nephron segments. Serum creatinine (SCr) and blood urea nitrogen (BUN) concentrations function as proxies of GFR, although their production rate, extrarenal elimination, and renal handling are not constant. A major disadvantage of the use of serum biomarkers to estimate GFR is that extrarenal determinants such as renal plasma flow (eg, hypovolemia) or hydraulic pressure in Bowman's space (eg, urinary obstruction) that decrease GFR without renal cause also can lead to nonspecific increases in the concentrations of these markers. In this context, urinalysis can add important information on the possible presence of kidney injury, dysfunction, or both in a patient. Examination of urinary sediment, especially urinary casts, and determination of urine specific gravity (USG) can be helpful in the diagnosis of AKI, but these tests lack sensitivity and specificity. Interpretation of SCr in combination with USG can help to determine whether azotemia is renal or nonrenal. Nevertheless, there are many variables such as fluid therapy before sampling, and other conditions that lead to dilute urine that limit its usefulness. When interpreting laboratory results, the clinician must understand that the relationship between serum biomarkers that estimate GFR and “measured GFR” is curvilinear. Clinically, this implies that if the kidneys undergo an insult when GFR is relatively normal, serum biomarkers may not reflect the resulting change in GFR. Furthermore, broad reference intervals in the concentrations of serum biomarkers and assay imprecision can mask clinically relevant decreases in GFR. At least 75% of nephrons must be nonfunctional before azotemia is present. This definition of azotemia is based on the traditional evaluation of SCr with reference ranges that may be too large, and without evaluation of smaller changes in SCr. Finally, serum biomarkers only provide a good estimation of the GFR when steady-state equilibrium has been achieved. This may not be the case with AKI when GFR can change rapidly. Direct determination of GFR using clearance measurements (“measured GFR”) in a well hydrated patient is a more accurate and sensitive strategy to evaluate renal function because it allows detection of decreased GFR before azotemia is present. Unfortunately, this method requires multiple samples, is time consuming, and requires complex analytical and calculation methods that limit its use in a clinical setting.[3, 4] Analysis of urinary biomarkers therefore may represent the best option for evaluation of kidney injury in human and veterinary medical practice. However, a better understanding of smaller changes in SCr also will help detect AKI at an earlier stage, and such changes are now included in the IRIS AKI grading system.
In this review, the underlying etiology of AKI in dogs is discussed, providing the relevant context on the most promising urinary biomarkers for early AKI detection. One retrospective study reported the prevalence of disorders previously associated with AKI in 99 dogs diagnosed with AKI. Isolated ischemic events were identified in 33 of these AKI dogs (33%), followed by isolated nephrotoxicant exposure in 21 dogs (20%). In the former group, pancreatitis, hypovolemic or hypotensive shock, or sepsis was identified in 67% of patients, whereas ethylene glycol intoxication was present in more than 50% of patients in the latter group. Multiple disorders were identified in 18 AKI dogs (17%), with disseminated intravascular coagulation in conjunction with another disease reported in more than 50% of these patients. Infectious disorders were reported in 4 AKI dogs (4%) and 1 dog (1%) was diagnosed with renal lymphosarcoma. Nevertheless, in the 22 remaining AKI dogs (20%) no AKI-associated disorder (or cause) could be identified. The renal cortex receives 90% of renal blood flow. This makes this region of the kidney particularly vulnerable to toxins originating from the blood. In addition, cells of the proximal tubule and thick ascending limb of the loop of Henle are most susceptible to ischemic injury because of their high metabolic activity. Therefore, it can be hypothesized that both low molecular weight (MW) and tubular proteins are likely to be the most promising urinary AKI biomarkers in dogs.
AKI Staging or Grading Systems in Humans and Dogs: Improving the Sensitivity of SCr as Surrogate Marker of GFR
In human medicine, the Acute Dialysis Quality Initiative group was the first to propose a graded definition of acute renal failure, the Risk of renal dysfunction-Injury to the kidney-Failure of kidney function-Loss of kidney function-End-stage kidney disease (RIFLE) criteria. This definition is based on increased SCr, decreased GFR, or decreased urine output. In addition to 3 AKI severity categories (Risk, Injury, Failure), 2 outcome groups (Loss, End-stage renal disease) also were defined. The Acute Kidney Injury Network (AKIN) further modified the SCr criteria to include less severe acute renal failure whereas eliminating the GFR criteria and both outcome groups of the RIFLE staging system. The AKIN also introduced the term AKI to better represent the entire spectrum of acute renal failure. Recently, the Kidney Disease: Improving Global Outcomes (KDIGO) organization proposed a consensus definition combining the best of both existing staging systems. The KDIGO staging system utilizes the 48-hour time frame from AKIN for the 0.3 mg/dL increase in SCr criterion and a 7-day time frame from RIFLE for the 50% increase in SCr criterion. The classification of 3 grades of increasing severity of AKI (Risk: stage 1, Injury: stage 2, Failure: stage 3) and the urine output criteria, as included in both RIFLE and AKIN, were retained.
Three similar staging or grading systems for dogs with AKI recently have been suggested.a,[[10, 11]] Although retrospective evaluation of two of these systems[10, 11] looks promising, prospective evaluation of a large population of dogs is needed to test the accuracy and clinical usefulness of AKI staging or grading systems in dogs. Moreover, these proposed systems require a multicenter approach and standardized inclusion and exclusion protocols. The retrospective analysis by Thoen and Kerl included nonazotemic dogs at admission, thus selectively investigating hospital-acquired AKI. Furthermore, both absolute and relative SCr changes from baseline, as defined by AKIN, were used to define different severities of AKI in dogs. In contrast, Lee and colleagues included only dogs that were azotemic at admission, thus investigating the community-acquired AKI subgroup. Dogs with chronic kidney disease (CKD), urinary obstruction, or bladder rupture were excluded. In this retrospective analysis, SCr reference ranges for each stage were used to stratify patients into AKI risk categories. A “normal canine GFR” was selected, and decreases of 25, 50, or 75% from this baseline were calculated as defined by RIFLE. Using an equation formula, the corresponding SCr concentrations were obtained. Urine output criteria and time frames were absent in both these studies. A prospective study defining AKI in intensive care unit dogs (n = 109) by 2 human staging schemes (ie, RIFLE and AKIN), and by the 3rd canine grading scheme of Cowgill,a reported an AKI incidence of approximately 10% at the interim analysis level. Furthermore, treatment with fluoroquinolones or diuretics was identified as a possible risk factor, and no association was found between AKI and either length of stay or survival.b These observations contrast markedly with reported data in humans. Classification schemes for AKI have proved to be valuable in predicting mortality in affected dogs[10, 11],c and cats.c
The canine AKI grading and subgrading scheme proposed by Cowgilla is now accepted by the IRIS group. It defines five AKI grades based on SCr concentration. Each grade is further subdivided on the basis of urine production (oliguric or nonoliguric) and initiation of (or requirement for) renal replacement therapy (RRT). Historical, clinical, laboratory (eg, azotemia), or imaging evidence must confirm the diagnosis of AKI before grading on the basis of SCr can be performed. Grade 1 defines AKI patients that are nonazotemic or those with SCr concentrations that are immediately responsive to adequate volume therapy. Dogs with AKI that are azotemic at admission but respond well to fluid therapy with normalization of the SCr concentration (nonazotemic concentration) should be reclassified as AKI grade 1. An increase in SCr concentration by 0.3 mg/dL within 48 hours can be considered confirmatory laboratory evidence of AKI. This criterion allows the diagnosis of acute on CKD and includes nonazotemic patients with insufficient evidence for AKI diagnosis based on other findings.
Definition of an Ideal Biomarker
Biomarkers are defined as biological variables that can be objectively measured and act as indicators of normal processes, pathological processes, or responses to intervention. The ideal biomarker for kidney injury should be able to (1) detect kidney injury at an early stage, (2) localize kidney injury (ie, at the glomerular level, tubular level, or both), (3) differentiate renal injury from pre-, post-, and nonrenal injury, (4) predict severity of renal injury, and (5) monitor the effects of intervention. To be clinically applicable, the biomarker should be accurate, easy to measure, and noninvasive. Serum biomarkers only correctly reflect kidney dysfunction if they are increased as a consequence of actual damage to the kidneys and not by extrarenal factors. Because of its proximity to the kidney and because it is easy to obtain, urine is a more promising biological fluid with which to identify the earliest biomarkers of kidney injury. Nevertheless, at different stages of AKI, it may be beneficial to evaluate different types of urinary biomarkers. Within the spectrum of kidney injury/dysfunction/failure, the early stage of kidney injury may be detected by early injury biomarkers whereas the later stages of kidney dysfunction/failure as diagnosed by routine functional biomarkers (eg, SCr, BUN) also may be detected by late injury biomarkers.
Specific detection of glomerular injury, tubular injury, or both at an early stage (ie, before GFR is decreased or before decreased GFR is detected by routine serum biomarkers) might permit earlier therapeutic intervention. Therefore, evaluation of urinary proteins is a promising strategy for detecting kidney injury.[16-19] Normal urine should contain only a small amount of protein because of the mechanical barrier of the glomerulus and the reabsorptive capacity of the proximal tubules. Rather than measuring overall proteinuria, the analysis of specific proteins in urine may provide better insight into the pathogenesis of the renal disease or lead to the discovery of proteins with early diagnostic potential for the disease. Evaluation of multiple candidate biomarkers to identify renal diseases might lead to additional insight into the pathophysiology of these diseases.
The ratio of the concentration of any urinary biomarker to urinary creatinine concentration is used to account for differences in USG between samples. Alternatively, urine can be diluted to a standard USG of 1.010 before biomarker measurement.
Urinary total protein is a mixture of filtered plasma proteins, kidney-derived proteins, and proteins originating from the lower urinary tract. The discovery of candidate urinary protein biomarkers for kidney injury is essentially a hypothesis-generating process and generally is based on 2 approaches.
In the 1st approach, candidate novel urinary biomarkers are proposed based on the pathophysiology of the disease. Altered functioning of the nephron can result in the presence of large amounts of proteins in urine. The glomerular filtration barrier normally excludes most proteins as large as or larger than albumin (ie, MW >69 kDa; Fig 1A). The charge of a protein also influences its filtration: positively charged proteins pass the glomerular barrier more easily than negatively charged proteins. Changes in the structure or composition of this barrier or in the hemodynamic state of the patient can lead to decreased glomerular permselectivity. This primary glomerular dysfunction results in the presence of high amounts of proteins with intermediate or high MW in the ultrafiltrate (Fig 1B). Proteins smaller than albumin (ie, low MW proteins) are freely filtered by the glomerulus. However, these proteins subsequently are efficiently reabsorbed by the proximal tubules in a normally functioning kidney (Fig 1A). Both primary and secondary tubular dysfunction represent an inability of the tubules to completely reabsorb filtered proteins and can result in proteinuria (Fig 1B and C). Affinity of the tubular receptors megalin and cubilin for these filtered proteins differs, and competition for binding has been reported in cases of protein overload.
In summary, both tubular and glomerular dysfunction can lead to proteinuria, albeit by different mechanisms. All of these proteins are detected simultaneously during evaluation by routine urinalysis and urinary protein to creatinine ratio (UPC) measurements, whereas more sophisticated testing allows their differentiation. However, as the mechanisms of tubular and glomerular dysfunction are sometimes associated, which was recently demonstrated in dogs with progressive glomerular damage and proteinuria, their presence may not be mutually exclusive, limiting the benefits of evaluating for high versus low MW proteinuria, respectively. Abnormal passage of proteins and dysfunction of tubular cells potentially can initiate a cascade of progressive loss of renal function.[24-26]
The 2nd approach can be considered a “proteomic assessment” with the goal of gaining a global overview of proteins and peptides in urine. Multiple techniques are available for this approach. Mass spectrometric analysis of the urinary proteome creates an extensive map of the proteins present in urine, which can promote the discovery of biomarkers for kidney injury. This strategy has been intensively applied in human nephrology[18, 19] but is rarely performed in veterinary medicine.[27-29] Two of these proteomic studies have identified retinol-binding protein (RBP) as a urinary biomarker for proximal tubular dysfunction.[28, 29] One study compared the protein profiles of urine from dogs with naturally occurring renal disease with those of urine from healthy dogs, whereas the other study compared the protein profiles of urine obtained at the onset of “glomerular” proteinuria with those of urine obtained at the onset of azotemia in dogs with X-linked hereditary nephropathy (XLHN). Decreased excretion of Tamm-Horsfall protein also was identified as indicative of distal tubular dysfunction in the former study. Urinary protein patterns in dogs have been evaluated by gel electrophoresis separation followed by either total protein staining,d,[30-34] or more specific protein immuno-assay with Western blot analysis.[30-33] Both of these methods can determine whether glomerular dysfunction, tubular dysfunction, or both contributed to the observed proteinuria. However, antibody specificity has been found to be low in dogs[31-33] and, in contrast to mass spectrometry, simultaneous identification of multiple individual proteins is not possible with Western blot analysis. Furthermore, gel electrophoresis separation identifies only a small part of the urinary proteome. The cited gel electrophoresis studies either compared urinary protein profiles of dogs with naturally occurring renal disease with some combination of renal pathological, immunohistological, and histopathologicala,[30, 34] findings, or compared protein profiles of urine from dogs with leishmaniasis, leptospirosis, or pyometra with those of urine from healthy dogs.
Most studies in dogs have selected urinary protein biomarkers based either on the pathophysiology of the investigated renal disease or on extrapolation from human medicine. Few studies have reported the use of canine proteomic urine screening approaches to discover novel renal protein biomarkers.d,[27-34] Mass spectrometric analysis of canine urine[27-29] in particular rarely has been performed.
Urinary Protein Biomarkers for Kidney Injury in Dogs
Next, an overview is given of studies describing the potential use of proteins originating from tubular cells and of specific intermediate MW, high MW, and low MW proteins as biomarkers for kidney injury in dogs. References to databases describing the detailed structure and biochemical characteristics of these urinary proteins also are provided. The majority of the referenced studies evaluated only a single specific protein as a candidate novel urinary biomarker; few studies have proposed a panel of biomarkers.[35-37] Nevertheless, both types of study are useful and enhance the likelihood of selecting a biomarker that is of clinical relevance. For many of the urinary biomarkers discussed, information for dogs with AKI currently is limited and therefore, where appropriate, information available from studies evaluating dogs with CKD or preazotemic CKD is emphasized.
Albumin and Other Urinary Intermediate MW Proteins
Albumin (Alb, MW 69 kDae) is synthesized in hepatocytes and is the most abundant protein found in plasma. Its major functions are to provide oncotic support and carrier capacity. The glomerular barrier almost completely limits Alb filtration because of its intermediate MW and negative charge, and in normal physiological situations healthy proximal tubular cells reabsorb most of any Alb that is filtered. Nevertheless, the capacity for reabsorption can be overloaded with excessive protein load. Results from the first large international prospective observational study investigating biomarkers after adult and pediatric cardiac surgery (Translational Research Investigating Biomarker Endpoints—or TRIBE—in AKI study) show potential for urinary Alb-to-creatinine ratios in preoperative AKI risk assessment in adults, whereas urinary Alb (evaluated as urinary Alb-to-creatinine ratio in children) may predict AKI in the early postoperative period when measured in adults and children.
Albumin was one of the first candidate urinary biomarkers extensively studied in dogs. When urine samples are diluted to a standard USG of 1.010, microalbuminuria is defined as a urinary Alb concentration >1 and ≤30 mg/dL. Albumin concentrations >30 mg/dL can be detected by conventional dipstick urinary protein screening methodology, and patients with these concentrations of urinary Alb are defined as overtly albuminuric or proteinuric. The IRIS group (http://www.iris-kidney.com/) recommends determination of UPC values to guide clinical decision-making and to monitor trends in CKD dogs. UPC values >0.5 define proteinuric dogs, UPC values <0.2 define nonproteinuric dogs, and UPC values ranging from 0.2 to 0.5 define borderline proteinuric dogs. When measuring urinary Alb in the latter 2 groups of dogs, microalbuminuria may be present, but the clinical relevance of this finding in guiding decision-making presently is not understood. When using urinary Alb-to-creatinine ratios, a cutoff of >30 mg/g can be used to define microalbuminuria.f In humans, urinary Alb-to-creatinine ratios ≥30 mg/g define microalbuminuria, whereas ratios ≥300 mg/g define overt albuminuria or proteinuria.
Microalbuminuria was shown to be a good indicator of early glomerular dysfunction in dogs with XLHN,f in dogs experimentally infected with Dirofilaria immitis L3 larvaeg and in Soft-coated Wheaten Terriers genetically at risk for protein-losing enteropathy and nephropathy.h Persistent microalbuminuria or proteinuria with inactive urine sediment strongly suggests the presence of canine CKD. However, some studies also observed microalbuminuria in nonrenal diseases, thereby questioning the specificity of this condition for diagnosis of renal diseases in dogs. Systemic injury can activate cytokine cascades and cause endothelial dysfunction, which can induce increased capillary permeability to plasma proteins. Nevertheless, there is consensus that dogs with at least 3 consecutive UPC values ≥0.5 (collected over a period of at least 2 weeks) that cannot be attributed to either a pre- or a postrenal cause (reflecting persistent renal proteinuria) most likely have glomerular or tubulointerstitial CKD. In general, urinary Alb should not be used for the screening of CKD. Indeed, although this renal biomarker has higher sensitivity than UPC, the latter has higher specificity.
Although urinary Alb remains useful in evaluating the renal handling of plasma proteins, its potential role in the early diagnosis of kidney dysfunction in nonazotemic dogs recently has been questioned. In addition, the presence of urinary Alb is not site-specific, because both glomerular and tubular dysfunction can result in passage of Alb into the urine (Fig 1B and C). Although large amounts of urinary Alb generally reflect altered glomerular permselectivity, they often do not permit localization of the affected compartment within the kidney. In the case of AKI, the renal tubules typically are damaged. Acute onset of tubular microalbuminuria may therefore be indicative of AKI and can be monitored if baseline values are available. However, this approach has not been evaluated in clinical studies.
Other intermediate MW proteins such as vitamin D-binding protein (MW 56 kDa), transthyretin (MW 55 kDa), and urinary transferrin (MW 76 kDa) have been reported to reflect glomerular proteinuria in humans. Increased amounts of vitamin D-binding protein and transferrin were observed by Western blot analysis in urine of dogs with XLHN compared to urine from healthy littermates. The presence of both proteins also was confirmed by Western blot analysis in 1 dog with CKD resulting from juvenile nephropathy. Nevertheless, their quantification and the evaluation of their individual potential as urinary biomarkers for kidney injury or dysfunction remain to be performed in dogs, as well as in humans.
Immunoglobulin G and Other Urinary High MW Proteins
Immunoglobulin G (IgG, MW 150 kDai) is produced by activated B lymphocytes (plasma cells) and circulates in the plasma as part of the humoral immune system. The appearance of this high MW protein in urine has been linked to altered glomerular permselectivity and to severity of glomerular lesions in humans. Acute kidney injury is almost always associated with primary tubular damage, thus limiting the usefulness of IgG and other high MW proteins in the diagnosis of canine AKI.
Immunoglobulin G was found to be increased in the urine of dogs with chronic renal failure, leishmaniasis,[31, 45] pyometra, leptospirosis, Cushing's syndrome, and XLHN.[23, 47] Quantitative results obtained with ELISA are available for canine leishmaniasis (not normalized for urine volume), for dogs with Cushing's syndrome, and for dogs with XLHN (normalized to urinary creatinine), whereas semiquantitative results generated with Western blot analysis were reported in the remaining studies.[23, 30-33]
The glomerular basement membrane of dogs with XLHN has altered structure and function that leads to progressive proteinuria. Immunoglobulin G was found in the urine of these dogs before the onset of azotemia and at higher concentrations than were found in the urine of age-matched healthy control dogs.[23, 47] Although these findings suggest that IgG may be useful for the early diagnosis of altered glomerular permselectivity, clinicians should be aware that tubular damage also may be present. In general, the identified glomerular damage should alert clinicians to carefully monitor renal function because glomerular diseases are initiation type risk factors for the development of CKD in dogs (http://www.iris-kidney.com/), whereas acute glomerulonephritis may sometimes lead to AKI too. More importantly, because CKD may be considered a susceptibility type risk factor for the development of AKI, this identification allows adaptation of patients' therapies according to their sensitivity to AKI (eg, choice of alternatives for nephrotoxicant drugs in patients at increased risk for AKI). On the other hand, AKI episodes are initiation type risk factors for CKD development in dogs (http://www.iris-kidney.com/).
Leptospirosis, a well-known cause of AKI in dogs, can result in urinary protein patterns indicating tubular damage and histopathological evidence of interstitial nephritis. This is not surprising because the causative microorganism can replicate in tubular epithelial cells. Nevertheless, in some dogs with leptospirosis, urinary IgG also was detected by Western blot analysis, suggesting possible glomerular involvement in a minority of affected dogs.
Similar to IgG, IgA concentrations were found to be higher in the urine of dogs with leishmaniasis, leptospirosis, and pyometra than in urine of healthy dogs. However, quantification of IgA was not performed, and results were based on gel electrophoresis of urine proteins. Increased urinary C-reactive protein (CRP, MW 115 kDaj) concentrations quantified by ELISA and normalized to urinary creatinine were measured in dogs with renal damage secondary to pyometra and in dogs with CKD.
Serum CRP concentrations also were found to be significantly increased in dogs with naturally occurring renal disease when compared to serum concentrations of healthy dogs. The observation of increased serum CRP concentrations in the nonazotemic, nonproteinuric group with decreased exogenous plasma creatinine clearance suggests that it is a promising early biomarker for kidney injury or dysfunction. In the uremic, overtly proteinuric group, all dogs had UPC ≥2 indicating glomerular damage. A higher serum CRP concentration also was reported in these dogs compared to the healthy dogs. Several hypotheses have been put forth to explain this observation, but further research is needed to determine the diagnostic and prognostic value of serum CRP concentration in canine renal disease.
No other high MW plasma proteins filtered by the glomeruli have been studied in the urine of dogs. Evaluation of the ability of such high MW proteins to detect renal disease (primarly CKD) before the onset of azotemia, as was performed for IgG in urine of dogs with XLHN, therefore is needed.
Retinol-binding Protein and Other Urinary Low MW Proteins
Retinol-binding protein (RBP, MW 21 kDak) is a low MW plasma protein synthesized in the liver and is the major carrier of retinol (vitamin A). The RBP-retinol complex is bound to transthyretin (which transports thyroxine and retinol) in plasma, and this binding precludes its glomerular filtration. Once retinol has been delivered to its target tissues, the affinity of RBP for transthyretin decreases. Uncomplexed RBP then is freely filtered by the glomeruli and efficiently reabsorbed after being catabolized in the proximal tubular cells. Consequently, urinary RBP has been proposed as a marker for proximal tubular dysfunction in humans and dogs.[30, 43, 54-56]
Urinary RBP may be useful in predicting the severity and outcomes of AKI in humans. Increased excretion of this protein in urine may predict subsequent development of microalbuminuria and diabetic nephropathy in normoalbuminuric diabetic human patients.[58-61]
Increased concentrations of urinary RBP were observed in dogs with CKD[30, 62] (not normalized for urine volume), urolithiasis (normalized to urinary creatinine), and XLHN by Western blot analysis, ELISA, or both compared with healthy dogs.[23, 30, 55, 62] The ability of RBP to identify kidney dysfunction at an early stage recently was evaluated in 3 independent studies. In the 1st and 2nd study, RBP was detected in urine before the onset of azotemia in dogs with XLHN,[29, 47] and additional data also support the correlation between RBP concentrations and CKD progression. A 3rd study investigated the influence of kidney function on urinary excretion of RBP in dogs with naturally occurring renal disease. In contrast to the first 2 studies, urinary RBP concentrations in the 3rd study were not different between normal dogs and nonazotemic dogs with decreased exogenous plasma creatinine clearance. Consequently, the authors concluded that urinary RBP was not diagnostically useful for the detection of mildly decreased GFR in the early stages of renal disease. These apparently contradictory data indicate that although RBP is a promising candidate urinary biomarker for tubular dysfunction in dogs, additional studies are warranted to evaluate its utility in the diagnosis of canine AKI.
Cystatin C (cys C, MW 13 kDa) is a cysteine protease inhibitor that is constantly produced by nucleated cells and released into the blood, where it is freely filtered at the glomerulus and, in normal physiological situations, reabsorbed and catabolized by the renal tubules without re-entering the bloodstream or being excreted in urine. The appearance of cys C in urine reflects tubular cell dysfunction. In human patients, urinary cys C (not normalized for urine volume) can accurately detect tubular dysfunction in both pure tubular and mixed tubulo-glomerular nephropathies. However, urinary cys C concentrations also may increase when excess Alb appears in urine because of competition for tubular reabsorption, even though concomitant tubular dysfunction is absent. Urinary cys C may be useful in predicting the severity and outcomes of AKI in human patients. Indeed, biomarker performance results from multicenter studies reported the ability of urinary cys C to diagnose AKI early and predict the need for RRT in critical illness or sepsis. In contrast, in the setting of cardiac surgery (TRIBE AKI study), urinary cys C did not either diagnose AKI early postoperatively or forecast AKI progression.
Recently, urinary cys C was evaluated as a marker for proximal tubular dysfunction in dogs with naturally occurring renal disease. In this study, urinary cys C (normalized to urinary creatinine) appeared to be a promising novel marker for canine renal disease, but the presence of proteinuria was not assessed in the renal and nonrenal disease groups.
The evaluation of urinary cys C and RBP in canine AKI patients has not yet been performed. Indeed, the study by Monti et al on urinary cys C included only 3 dogs with acute renal failure. Moreover, for their statistical analysis, the authors grouped the 3 dogs with acute renal failure with a larger number of CKD dogs into the “renal disease” group, which may not allow a valid comparison.
Because of the inherent lack of sensitivity and specificity of SCr and BUN for detecting renal dysfunction, serum and plasma cys C also have been investigated as alternative markers of renal function in dogs.[65, 66] Although not yet used as routine biomarkers, serum or plasma cys C concentrations are attractive because they are mainly determined by glomerular filtration. Therefore, according to the human medical literature, serum or plasma cys C may be useful indices of GFR. In the TRIBE AKI study, preoperative serum cys C concentrations could predict AKI in adults, suggesting a role for this biomarker in preoperative AKI risk assessment in cardiac surgery in adults. A preliminary study in dogs reported that nonrenal diseases minimally influenced serum cys C concentrations. Serum cys C has been proposed as an early marker for kidney dysfunction in critically ill dogs and dogs with leishmaniasis,[69-71] chronic nephritis, and diabetes mellitus. Although results are promising, additional studies are needed to confirm the value of serum cys C compared with SCr, especially in light of the reported biological variance of plasma cys C concentrations in dogs.
Other urinary low MW proteins such as α1-microglobulin and β2-microglobulin have been intensively studied as tubular dysfunction biomarkers in human kidney diseases.[73, 74] Increased urinary excretion of β2-microglobulin at baseline predicted deterioration of kidney function in adult patients with membr-anous nephropathy; however, the instability of β2-microglobulin in urine may limit its utility in detecting tubular dysfunction. Urinary α1-microglobulin, in contrast, may be useful for predicting the severity and outcomes of AKI.
Only 2 studies have evaluated these proteins in canine urine. One of these studies found increased concentrations of urinary β2-microglobulin with Western blot analysis during renal disease progression of dogs with XLHN, whereas the other study found that urinary α1-microglobulin and β2-microglobulin both were increased by Western blot analysis in dogs with XLHN and progressive renal disease.
N-acetyl-β-D-glucosaminidase (NAG, MW 150 kDal), alkaline phosphatase (AP), γ-glutamyl transferase (GGT), alanine aminopeptidase (AAP), and lactate dehydrogenase (LDH) are enzymes primarily located either in the lysosomes (NAG, AP), on the brush border (luminal side; GGT, AAP) or in the cytoplasm (LDH) of proximal tubular cells. Their size precludes glomerular filtration, and increased urinary concentrations of these proteins are believed to be related to tubular injury.[16, 17] Lysosomal enzymes normally are excreted in urine as the secondary lysosome remnants fuse with the cell membrane. Therefore, increased urinary activity of lysosomal enzymes may be because of enhanced lysosomal turnover as a result of stimulation of the endocytic lysosomal system rather than tubular injury. Nevertheless, all of these tubular enzymes are increasingly excreted in urine after tubular injury because of brush border shedding and cell membrane disruption. In humans, urinary NAG may be useful for confirming established AKI and predicting its severity and outcome.
The magnitude of urinary NAG-, urinary AP-, and urinary GGT-to-creatinine ratios was found to be associated with the severity of lesions in proximal tubular cells of dogs with pyometra. Together with the finding that the ratios decreased after ovariohysterectomy, this may indicate that in dogs with pyometra enzymuria reflects acute lesions associated with the disease rather than lesions associated with pre-existing CKD. The urinary NAG index (ie, the urinary NAG-to-creatinine ratio) was increased in dogs with pyelonephritis compared to patients with uncomplicated lower urinary tract infections. Urinary AP- and urinary GGT-to-creatinine ratios might be used to detect established AKI, but not CKD, in dogs with naturally occurring renal disease. Both of these indices could differentiate AKI dogs from healthy dogs, but none could differentiate between CKD dogs and healthy dogs. The urinary AP index additionally could differentiate AKI dogs from CKD dogs. The large overlap in urinary GGT indices between the AKI group and the control group partially could be explained by the within-day variation in urinary GGT. When using urine spot samples, normalization of urinary enzyme activity-to-urinary creatinine does not preclude within-day variation. Otherwise, diuresis also was found to have a marked influence on urinary enzyme excretion. Likewise, the early onset of acute renal damage was accurately reflected by a 2- to 3-fold increase in either baseline urinary NAG- and urinary GGT-to-creatinine ratios, baseline 24-hour urinary GGT activity, or baseline urinary GGT-to-creatinine ratio in dogs with gentamicin-induced nephrotoxicosis. In contrast, urinary NAG-, urinary GGT- and urinary LDH-to-creatinine ratios were slightly increased after cisplatin-induced AKI in dogs, despite the presence of azotemia and histological evidence of tubular injury. No explanation of these findings was given, but the variation in age and body weight among the dogs was high. When measuring enzyme activities in urine to detect AKI, the ability to monitor trends in experimental versus naturally occurring AKI precludes false positive results. This is because of the fact that enzymuria may be an excessively sensitive test to assess kidney injury when measured at only 1 time point. Therefore, the increase in activity of a single enzyme at 1 time point has limited diagnostic value. Urinary NAG is, to date, one of the most studied tubular protein biomarkers in dogs.
Tamm-Horsfall protein (THP, MW 100 kDa) is synthesized in the cells of the thick ascending loop of Henle and is located principally on the surface of the luminal cell membrane. In humans, vitamin A is secreted in urine as a water-soluble metabolite. In contrast, carnivores excrete vitamin A in urine as lipophilic retinol and retinyl esters. Therefore, THP is released in the urine of dogs to facilitate the excretion of retinol and retinyl esters. Tamm-Horsfall protein may have some immunomodulatory activity,[89, 90] but it is clinically important because it represents the matrix of all urinary casts.
Occasional reports have described urinary THP as biomarker for distal tubular dysfunction in dogs. Its potential use is illustrated by decreased urinary THP-to-creatinine ratios in dogs with CKD and in dogs with urolithiasis. The high metabolic activity of the cells of the thick ascending limb of Henle's loop makes them particularly vulnerable to ischemic injury. Consequently, in AKI caused by ischemic insults, evaluation of THP as an early AKI biomarker is potentially valuable (eg, in hospital monitoring of sepsis patients).
Other Tubular Proteins
Kidney injury molecule-1 (KIM-1) is a transmembrane protein that is expressed on the luminal surface of the proximal tubules during acute or chronic tubular injury or dysfunction. It is believed to play a role in the subsequent regeneration process of tubular structures. Urinary KIM-1 is a potentially useful biomarker for the diagnosis of established AKI in humans. More specifically, it is a promising renal biomarker for the early diagnosis of AKI in the early postoperative period after cardiac surgery (TRIBE AKI data for KIM-1 are still under review), but not for the prediction of AKI progression. In the setting of critical illness or sepsis, results from multicenter studies showed the ability of urinary KIM-1 to diagnose AKI early, but being unable to forecast the need for RRT. Liver-type fatty acid-binding protein (LFABP) binds selectively to intracellular free unsaturated fatty acids and lipid peroxidation products during hypoxic tissue injury, thereby limiting further oxidative damage. It is found in urine in response to hypoxia. Urine concentration of LFABP has been described as a sensitive indicator of acute and chronic tubulointerstitial injury. Urinary LFABP is one of the most promising biomarkers for early prediction of AKI in humans. Although both of these tubular proteins have potential for diagnostic purposes, they have not yet been evaluated in dogs.
Urinary interleukin 2 (IL-2), IL-8, monocyte chemoattractant protein-1 (MCP-1), granulocyte-macrophage colony-stimulating factor (GM-CSF), and keratinocyte-derived chemokine (KC), normalized to urinary creatinine, were found to be increased earlier than SCr in dogs with acute tubular injury after cisplatin administration. This study was the first to propose inflammation-associated proteins in the urine of dogs as candidate biomarkers for AKI. Urinary neutrophil gelatinase-associated lipocalin (NGAL) and IL-18 are novel candidate biomarkers currently under intensive evaluation in humans. The urinary concentrations of both of these biomarkers may be useful for early prediction and diagnosis of established AKI in humans. In the TRIBE AKI study, urinary IL-18 performed well in the early diagnosis of AKI postoperatively, both in children and adults, whereas its ability to predict AKI progression was limited to adults. Urinary NGAL was useful in the early postoperative period to diagnose AKI in children. Both biomarkers also performed well based on multicenter data in the setting of critical illness or sepsis for the early diagnosis of AKI and the prediction of the need for RRT. Similar to KIM-1 and LFABP, with the exception of 2 proof-of-concept studies on urinary NGAL, no studies examining these proteins in dogs have been performed. Indeed, urinary NGAL (not normalized for urine volume) recently was investigated in 6 dogs treated with gentamicin. The data from 2 dogs showed that this biomarker was detected 9 and 7 days earlier than an increase in SCr.m Urinary NGAL also was monitored in 25 puppies with XLHN. In these dogs, before azotemia developed, urinary NGAL-to-creatinine ratios were found to be increased compared with concentrations in unaffected dogs, with changes approaching statistical significance.
An overview of the urinary biomarkers examined in clinical conditions and the location of the renal damage they indicate is provided in Table 1.
Table 1. Overview of urinary protein biomarkers studied in dogs
Nephron Segment Involved
Clinical Conditions in Which the Biomarker Has Been Studied in Dogs
Evaluation of urinary biomarkers for early detection of canine AKI requires prospective analysis in a sufficient number of dogs at risk for the development of AKI. In the hospital setting, serial monitoring is possible (hospital-acquired AKI) and risk factors may be known. Monitoring for AKI in the intensive care setting with sensitive and specific injury biomarkers may allow earlier therapeutic intervention and prognostic information. Moreover, long-term follow-up studies that measure clinical outcome in relation to injury biomarker concentration trends may provide valuable information when evaluating recovery from either hospital- or community-acquired AKI.
Limitations When Measuring Urinary Biomarkers
The validation of assays for the species-specific measurement of urinary Alb, urinary RBP, and urinary enzymes in dogs has been performed.[62, 94-96] However, these studies have not uniformly used the same validated assays, thus complicating direct comparison of the results. Urinary biomarker concentrations mostly were compared with those from healthy control populations, and reference intervals for urinary NAG and urinary GGT have been described in healthy dogs.[96, 98] Nevertheless, studies evaluating biomarker specificity are largely lacking. In addition, the day-to-day variability and diurnal variations of biomarker concentrations must be taken into account before changes in urinary biomarker concentrations can correctly be attributed to changes in renal function. Most studies have used quantitative immunoassays to measure urinary concentrations of biomarkers. However, such assays are relatively time-consuming and labor-intensive and therefore not ideal for clinical settings.
Reporting absolute concentrations of urinary biomarkers rarely is performed because of the major influence of diuresis. Normalization to urinary creatinine corrects for differences in urinary flow rate but comes with some caveats. First, this normalization implies that the rate of renal creatinine excretion is stable, which may not be the case in dogs with AKI. Changes in creatinine production (eg, caused by changing muscle mass) or excretion must be taken into account when comparing urinary biomarker-to-creatinine ratios of the same dog over time. Second, interindividual differences in urinary creatinine have been described in dogs, which makes it possible to over- or underestimate renal impairment. Finally, correct normalization to urinary creatinine requires that the excretion of creatinine is similar to that of the biomarker.
The ongoing development of canine AKI staging or grading systems holds great promise for the uniform characterization of AKI in dogs over different settings in the near future. The implementation of such a validated AKI staging or grading system in dogs will not only improve the recognition of canine AKI but also its management, as was shown for the IRIS CKD staging system. The discovery of novel urinary biomarkers indicating kidney injury provides important therapeutic prospects. First, injury biomarkers may diagnose AKI or (deteriorating) CKD earlier than the routinely used blood test SCr. Second, novel therapeutics that protect or repair tubular integrity and function can be evaluated in a population with actual tubular damage. Urinary biomarkers also can help differentiate between glomerular and tubular involvement, thus improving the diagnosis and staging or grading of specific renal diseases. The heterogeneous and typically multicausal nature of AKI complicates its diagnosis if only 1 urinary biomarker is measured at 1 selected time point. Evaluation of multiple urinary biomarkers over time will lead to more specific urinary biomarker profiles in different AKI settings, and likely to novel insights into the pathophysiology of this complex syndrome. These urinary biomarker concentrations should be interpreted together with the imperfect gold standard, SCr. Evaluation of the implementation of multiple urinary biomarkers in an AKI staging or grading system will likely further improve the sensitivity and specificity of the latter system allowing more appropriate redefinition of the different AKI stages or grades. Hypothetically, a selected panel of urinary biomarkers could add complementary information to routinely evaluated clinical tests. However, validation of both the specific and timely diagnostic value as well as the predictive value (eg, complete recovery, evolution to CKD, death) of the emerging AKI biomarkers in standardized settings currently is still largely lacking, even in humans.[102, 103] Future research therefore should focus on the in-depth evaluation of these novel and yet-to-be-discovered biomarkers in different renal diseases of dogs.
This research was funded by the Bijzonder Onderzoeksfonds of the University of Ghent (BOF grant to B. Maddens) and the Fonds voor Wetenschappelijk Onderzoek (FWO grant to J. De Loor).
Conflict of Interest: Authors disclose no conflict of interest.
Cowgill LD. Staging patients with acute kidney injury: a new paradigm. Proceedings of the 2010 ACVIM forum; 2010 June 9–12; Anaheim, CA
Eatroff A, Langston C. Acute kidney injury in dogs hospitalized in the intensive care unit: a prospective study (interim analysis) (2012 ACVIM Forum Abstracts). J Vet Intern Med 2012;26:798–798
Harison E, Langston C, Palma D, et al. Acute kidney injury (AKI) as a predictor of mortality in dogs and cats (2011 ACVIM Forum Abstracts). J Vet Intern Med 2011;25:723–724
Nabity MB, Brock R, Brown JS, et al. Comparison of urine sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) with renal histological findings in dogs with renal disease (2010 ACVIM Forum Abstracts). J Vet Intern Med 2010;24:769–769