Traditional Renal Biomarkers and New Approaches to Diagnostics

Abstract

Traditional biomarkers of renal disease have a number of limitations, whether evaluating veterinary patients or performing preclinical toxicity studies. Serum creatinine and urea nitrogen are affected by nonrenal influences that limit their usefulness for detecting small but significant decreases in glomerular filtration rate (GFR) in veterinary patients. These nonrenal influences can be more controlled in preclinical studies than in clinical patients; however, because of its high functional reserve, these estimates of GFR are insensitive for detecting kidney injury prior to loss of a substantial proportion of functioning nephrons. Urine biomarkers can be highly sensitive for tubular or glomerular injury that might lead to irreversible damage to the nephron. Several proteins are qualified by the Food and Drug Administration for nonclinical application as urinary biomarkers of drug-induced nephrotoxicity, and many of these also have preliminary data supporting their usefulness for kidney injury in dogs and cats. In addition to these relatively recently identified biomarkers, efforts are underway to discover new renal biomarkers using a variety of techniques including liquid chromatography–mass spectrometry and small RNA sequencing. Ultimately, the interplay between preclinical studies and clinical patients in discovery and validation of renal biomarkers is critical to their successful implementation.

Traditional biomarkers of renal disease are useful in the diagnosis and monitoring of disease; however, they all have a number of limitations, whether evaluating veterinary patients or performing preclinical toxicity studies. They can be generally divided into markers of glomerular filtration rate (GFR) or markers of glomerular or tubular damage and/or dysfunction. The most commonly used markers of GFR are creatinine and urea nitrogen. Creatinine is generally preferred over urea nitrogen, as it has fewer nonrenal influences. Still, creatinine’s nonrenal influences hamper its use for early detection of decreased GFR. Since creatinine is produced by the muscle, muscle mass is the most important factor determining basal creatinine concentration in health. Because of this, reference intervals for a widely disparate patient population are insensitive to early decreases in GFR. This is particularly a problem for veterinary patients, where age- and breed-specific reference intervals do not yet exist. To help counteract this limitation, estimated GFR formulas are used in human medicine to help indirectly account for factors that can affect muscle mass as well as inherent GFR, although these formulas also have significant limitations (Levey, Inker, and Coresh 2014). More preferred in an individual patient is to “trend” serum creatinine over time, meaning that serial determinations are made and compared to previous values. The rationale for this method is that in a healthy, adult animal, little intraindividual variation exists for creatinine. However, muscle wasting due to aging and/or disease progression decreases creatinine production and can interfere with trending, resulting in overestimation of GFR. Analytical variability has also been demonstrated to be a significant problem in veterinary medicine (Ulleberg et al. 2011; Braun et al. 2008), whereas in human medicine, efforts to reduce analytical variability through the Creatinine Standardization Program have been helpful (Delanaye, Cavalier, and Pottel 2017). Creatinine’s dependence on muscle mass, the wide reference interval used clinically, and high analytical variability affect veterinary patients more than human patients, where population-based reference intervals are more stratified and measurement of creatinine is more uniformly performed. Research animals are also more immune to these factors, as nonrenal influences can be more controlled in preclinical studies, age- and strain-matched control groups are standardly included for comparison, and samples are typically batched for analysis, limiting analytical variability. Regardless, a major limitation for all species is that the high functional reserve of the kidneys limits the sensitivity of markers of GFR for the detection of mild or early kidney damage. Therefore, the focus for early detection of kidney disease has largely shifted to markers of kidney injury in both patients and preclinical studies.

Traditional markers of glomerular and tubular damage include renal proteinuria, urine specific gravity, normoglycemic glucosuria, and casts. Proteinuria is the most commonly used marker of kidney damage and dysfunction, although it first requires that prerenal (e.g., hemoglobinuria) and postrenal (e.g., urinary tract infection) causes of proteinuria be excluded. Urine protein can then be quantified and normalized to urine creatinine concentration to provide a urine protein–creatinine ratio (UPC). In species that normally excrete very little protein in their urine (e.g., humans, dogs, cats), a UPC > 0.2 is considered abnormal or borderline abnormal. The higher the UPC, the more likely the protein loss is caused by glomerular damage rather than (or in addition to) tubular damage. However, in rodents, physiological proteinuria is common, particularly in males. This interferes with the ability to detect mild glomerular damage. For example, in a variety of mice strains, the UPC can be >10 in healthy male mice (Cheetham et al. 2008). In healthy Wistar rats, a UPC > 2 was commonly observed (Imafidon, Akomolafe, and Oladele 2016). Certain types of casts and normoglycemic glucosuria can be highly specific for tubular damage, but they are not commonly observed, and they may not be present in early disease.

The limitations of traditional biomarkers necessitate the discovery and use of new biomarkers to overcome these limitations in the clinics as well as preclinical studies. In veterinary medicine, a new marker of GFR has recently been made widely available to veterinarians (IDEXX Laboratories, Inc., Westbrook, ME, USA). Symmetric dimethylarginine (SDMA) is a methylated arginine that is released during proteolysis in all cells and is predominantly excreted in the kidney. It was identified as a marker of GFR in dogs and cats using liquid chromatography–mass spectrometry, but a high-throughput, antibody-based test is now commercially available through IDEXX. Studies in dogs and cats demonstrate a strong correlation with GFR and creatinine (Tatematsu et al. 2007; Jepson et al. 2008; Braff et al. 2014; Hall et al. 2014a, 2014b; Nabity et al. 2015). Several studies demonstrate that SDMA can increase above its reference interval as much as 2 and 4 years earlier than creatinine in naturally occurring disease in dogs and cats, respectively (Hall et al. 2014a, 2016). Based on clinical data from a large number of canine and feline samples with both creatinine and SDMA measured by IDEXX Laboratories, concordant results were obtained in approximately 90% and 80% of samples, respectively. The most frequent category of nonconcordant results was samples where SDMA was increased but creatinine was within the reference interval, accounting for 8% and 17% of samples in dogs and cats, respectively. The proportion of this particular category increased in older dogs and cats, composing up to 1/3 of the samples from patients 15 years of age or older (vs. 5–10% in animals < 10 years of age). A similar trend was observed in hyperthyroid cats. It is likely that the majority of these discordant samples come from muscle-wasted patients in which creatinine is overestimating GFR (i.e., creatinine is lower than would be expected for the degree of renal function). However, for some of the cases in which SDMA is increased and creatinine is within the reference interval, elevations of SDMA could be caused by nonrenal factors or, if SDMA is only mildly increased, by analytical or biological variability. The most intriguing discordant category is where creatinine is increased and SDMA is within the reference interval. This particular category accounted for approximately 1% of samples for both dogs and cats. An example was highlighted in a recently reported study, where a dog with leptospirosis had a moderately increased creatinine due to acute kidney injury but SDMA was well below the reference limit (Dahlem et al. 2017). No explanation for this discrepancy was found. SDMA has recently been validated for use in rodents (Sprague-Dawley rats; Cross et al. 2017), and validation in other species is in progress. In summary, SDMA appears to be a more reliable indicator of decreased GFR than creatinine in patients with poor and/or decreasing muscle mass, and it increases before creatinine in many dogs and cats with chronic kidney disease. However, more investigation is needed regarding nonrenal influences on SDMA, and variability in its measurement must be considered when interpreting results (Kopke et al. 2018).

Discovery and investigation of potential biomarkers of kidney injury represents a major area of investigation in translational research, with “-omic” technologies playing a prominent role in discovery. While a number of proteomic techniques exist, a comprehensive proteomic sample fractionation method that creates thousands of native protein subfractions has been recently investigated for kidney disease (Muckova et al. 2015). This technique was used to determine the proteome in plasma from mice with Alport syndrome, a glomerular disease that leads to end-stage renal disease. With this technique, investigators identified substantial heterogeneity in subfractions from a single protein (Muckova et al. 2015). For instance, for transferrin, most subfractions were unexpectedly higher in Alport mice than in controls at the preclinical stage, but by the clinical stage, most transferrin subfractions were lower in Alport mice, as expected for a negative acute phase response secondary to disease (Muckova et al. 2015). Preliminary data have demonstrated similar results in dogs with Alport syndrome. Furthermore, in Alport mice, some of the peptides identified in preclinical and clinical stages of disease persist at normal levels with treatment with an angiotensin converting enzyme inhibitor (Muckova et al. 2015). These proteins/proteoforms that indicate therapeutic effect could be particularly valuable for therapeutic targeting and monitoring as well as for early diagnostics. Overall, these results have shed light into the large number of modifications present in circulating proteins and their alteration with disease, and this represents an exciting area for future research.

MicroRNAs (miRNAs) in both urine and serum are also being explored as markers of renal disease. They are short noncoding RNAs (21–25 nucleotides in length) that posttranscriptionally regulate gene expression to influence a variety of physiological and pathological processes through complex miRNA-mRNA interactions. They have been found to be stable in biofluids, thereby representing a promising area of discovery that can influence disease detection as well as treatment. Preliminary data in dogs using small RNA sequencing for both urine and serum samples revealed a number of miRNAs that are differentially expressed in dogs with glomerular diseases compared with clinically healthy dogs. For instance, we found miRNA-21 to be significantly increased in the urine of dogs with kidney disease, similar to a recent study (Ichii et al. 2017). Several miRNAs also appeared to differentiate among the glomerular diseases included in our study.

In summary, traditional biomarkers will likely always have a place in the diagnosis of kidney disease, with proteinuria currently serving as the earliest indicator of kidney disease due to glomerular damage. However, new biomarkers are critically needed to improve early diagnosis and for therapeutic monitoring. SDMA is promising as an earlier marker of decreased GFR than creatinine in dogs and cats, particularly in those patients where decreased muscle mass interferes with creatinine interpretation. miRNAs are another promising source of biomarkers for kidney disease that can also reflect underlying disease pathogenesis. Many additional promising biomarkers have been discovered based on preclinical studies, and new techniques are being developed for discovery of even more. Ultimately, however, these new biomarkers will need extensive evaluation in the clinical setting to determine their pros and cons in a diverse population. This interplay between preclinical studies and clinical patients in discovery and validation of renal biomarkers is critical to their successful implementation in both arenas. Preclinical studies evaluating kidney biomarkers provide a controlled environment to narrow the field for clinical studies. Likewise, studies in human and veterinary patients can provide evidence for the use of preclinical biomarkers in the field.