Renal biomarkers in cats: A review of the current status in chronic kidney disease

Abstract Serum creatinine concentration, the classical biomarker of chronic kidney disease (CKD) in cats, has important limitations that decrease its value as a biomarker of early CKD. Recently, serum symmetric dimethylarginine concentration was introduced as a novel glomerular filtration rate biomarker for the early detection of CKD in cats. However, data on its specificity are still limited. The limitations of conventional biomarkers and the desire for early therapeutic intervention in cats with CKD to improve outcomes have prompted the discovery and validation of novel renal biomarkers to detect glomerular or tubular dysfunction. Changes in the serum or urinary concentrations of these biomarkers may indicate early kidney damage or predict the progression of kidney before changes in conventional biomarkers are detectable. This review summarizes current knowledge on renal biomarkers in CKD in cats, a field that has progressed substantially over the last 5 years.

regarded as abnormal in cats with dehydration or azotemia, urinary concentrating ability is occasionally intact in cats with CKD. 9,11 Furthermore, inappropriately dilute urine also can be non-renal in origin, for instance secondary to certain drugs (eg, diuretics), glucosuria, electrolyte imbalances, or liver disease. 12 Measurement of glomerular filtration rate (GFR) is the gold standard and most sensitive test for impaired renal function. 13 The GFR can be determined directly by measuring the clearance of an endogenous or exogenous filtration marker. 13 Although limited sampling techniques have been developed to improve the practical use of GFR measurement in cats, none of these current techniques is feasible for routine veterinary practice, and direct GFR measurement is still mostly a research tool. [14][15][16][17] Proteinuria as assessed by the urinary protein : creatinine ratio (UPC) is a routine renal biomarker and traditional hallmark of glomerular disease but also increases with tubular dysfunction in cats. 8,18,19 Persistent renal proteinuria without azotemia may indicate early glomerular disease in cats. 9 Epidemiologic studies have suggested that UPC is prognostic for survival, progression and the development of CKD in cats. 3,[20][21][22] Results of the UPC values in cats commonly are influenced by non-renal diseases such as hyperthyroidism, viral infections, and lower urinary tract diseases, which substantially decrease the specificity of UPC for kidney disease. 6,23,24 To delay CKD progression, treatment ideally should be instituted as soon as the diagnosis is achieved. 25 The search for serum or urine renal biomarkers that are more sensitive for detecting early kidney damage or small decreases in kidney function is driven by the limitations and poor sensitivities of traditional biomarkers such as serum creatinine (sCr) concentration. 26 The innovative concept that acute kidney injury (AKI) and CKD in cats and dogs are interconnected processes recently has been proposed, analogous to the situation in humans. This concept states that sustained or serious kidney injury, such as that occurring during an episode of AKI, could lead to the development of CKD and vice versa.
This ongoing kidney injury generally occurs before a noticeable decrease in GFR. 6,27 Furthermore, "acute kidney stress" was introduced as a concept in humans and dogs to describe a predisposition to AKI or a very early insult to the kidney before AKI. 28 Combining these concepts, it can be hypothesized that kidney stress or early kidney injury may be associated with the development and progression of CKD. Therefore, renal biomarkers that can identify active kidney stress or injury have the potential to detect CKD earlier than biomarkers that are surrogates of decreased GFR. Moreover, these injury biomarkers, especially those in urine, can also specifically localize injury to glomeruli or tubules. Furthermore, these biomarkers might have the potential to predict the development of CKD, monitor recovery, and facilitate prognostication. 6,29 In this review, we categorize and discuss renal biomarkers measured in feline blood and urine based on their origin and role in kidney disease. In particular, we describe biomarkers of GFR, metabolic derangement, glomerular vs tubular injury or dysfunction, and renal fibrosis, focusing on biomarkers described using new data obtained the last 5 years.

| Systemic biomarkers of GFR
The serum or plasma biomarkers under investigation in cats with CKD are summarized in Table 1.
Because of logistical drawbacks, GFR is usually indirectly estimated by measuring serum biomarker concentrations. The traditional GFR biomarker is sCr, which is widely used to diagnose CKD because it is economical, readily available, easy to measure, and shows low intra-individual variability. 9,40 Also when within reference range, serial monitoring of sCr can be used to detect early change in renal function. 10,40 However, sCr has important limitations. First, it is influenced by non-renal factors, especially muscle mass. 30,41,42 Second, sCr has a low sensitivity to detect an early decrease in GFR and cannot detect kidney damage that does not affect GFR. 6,29 Third, its wide reference interval because of high inter-individual variability and its analytical variability can lead to misinterpretation in clinical practice. 43,44 These major limitations hamper the utility of sCr to detect early CKD in cats, resulting in a search for other indirect GFR biomarkers such as symmetric dimethylarginine (SDMA).
Besides the conventional functional biomarker, sCr, which is the main representative of this group, SDMA concentration recently has been described as a novel GFR biomarker for the diagnosis of CKD in cats. 45,46 Its main advantage is that it is more sensitive than sCr concentration because it can detect a smaller decrease in GFR, increasing when the GFR decreases by approximately 40% in cats. 25,30 Another advantage is that SDMA concentration is not affected by muscle mass. [47][48][49] However, breed-related differences in SDMA concentration have been noted and require further exploration. 48 A more important limitation of SDMA concentration is that any increase must be interpreted with caution because data on its specificity are scarce. 46,50 International Renal Interest Society (IRIS) has recently incorporated SDMA concentration into guidelines for CKD staging. 51 According to IRIS, persistently increased SDMA concentration may indicate early CKD (IRIS stage 1). 51 Hence, multiple blood samplings are required when interpreting SDMA concentration. Because extensive reviews on the current knowledge of SDMA in dogs and cats have been published recently, 29,46,52,53 SDMA will not be discussed in depth in the present review. To date, the only other surrogate biomarker of GFR that has received attention in cats is serum cystatin C (sCysC). This marker will be reviewed below, but despite extensive analytical validation sCysC has failed to show clinical utility as a marker of GFR. 32,36 2.1.1 | Serum cystatin C Cystatin C is a 13 kilodalton (kDa) non-glycosylated protein produced by all nucleated cells at a constant rate. It is involved in intracellular protein catabolism as a proteinase inhibitor. 54 Circulating CysC generally passes through the glomerular barrier without restriction. 55 In the proximal tubules, CysC is reabsorbed via megalin receptors and then is metabolized completely. 56 There is no proof that CysC is secreted by tubular cells into urine. 57 In humans and dogs, sCysC concentration is used as GFR biomarker and is highly correlated with sCr and GFR. 34,57-60 Also, a GFR equation based on the combination of sCysC and sCr has been shown to be the optimal equation to estimate GFR in humans. 61 No cat-specific assay is available for CysC measurement. Although particle-enhanced nephelometric (PENIA) and particle-enhanced turbidimetric (PETIA) sCysC immunoassays used in humans had acceptable precision and reproducibility in cats, [31][32][33] the cross-reactivity between the anti-human CysC antibody and feline CysC was relatively low based on Western blot analysis. 31,32 This questions the reliability of the PETIA and PENIA used in humans for measurement of sCysC in cats. Moreover, sCysC concentrations significantly decrease after storage for only 5 months at both À20 C and À72 C. 62 Age, sex, breed, and body weight do not affect sCysC concentrations in cats, 35,63 in contrast to humans and dogs. 64,65 No significant differences were found between pre-and post-prandial sCysC concentrations in cats. 62 Hyperthyroidism increases sCysC concentrations in cats independent of decreased renal function. 24,36 However, feline immunodeficiency virus infection has no effect on sCysC concentrations. 24 Although initial studies indicated that CKD cats had significantly higher sCysC concentrations than healthy cats, there was overlap between groups. 31,35,58 Also, despite a significant negative correlation with GFR, sCysC concentrations were not different for different IRIS stages within the group of CKD cats. 35 The sensitivity of sCysC for the detection of decreased GFR was only 22%, and, in recent studies, sCysC concentrations could not differentiate CKD cats and healthy cats. 32,36 The correlation between GFR and sCysC was significantly weaker than that between GFR and sCr. 32 Additionally, sCysC could not predict the development of azotemia after treatment of hyperthyroid cats. 36 Overall, despite promising data in humans, sCysC is not a useful biomarker for renal function in cats.

| BIOMARKERS OF METABOLIC DERANGEMENT
The decrease in phosphate excretion from diseased kidneys results in disturbances in the systemic calcium-phosphate homeostasis characterized by increasing serum phosphate concentrations, stimulation of parathyroid hormone (PTH) secretion, and osteodystrophy. This systemic condition of calcium-phosphate homeostasis disturbance caused by CKD is called chronic kidney disease-mineral and bone disorder (CKD-MBD). 66 Recently, knowledge of this metabolic derangement in cats with CKD has been improved by studying how the phosphaturic hormone, fibroblast growth factor-23 (FGF-23), behaves in CKD cats and geriatric cats (Table 1).

| Fibroblast growth factor-23
Fibroblast growth factor-23 is a hormone responsible for the regulation of phosphorus and calcitriol produced by osteocytes and osteoblasts. 67,68 Both hyperphosphatemia and increased blood calcitriol (active vitamin D) concentrations stimulate FGF-23 production, which promotes urinary phosphorus excretion and decreases intestinal phosphorus reabsorption. Urinary phosphorus excretion is increased by the inhibition of proximal tubular sodium-phosphorus cotransporters, whereas intestinal phosphorus reabsorption is decreased by impaired renal calcitriol production. [69][70][71] The concentration of FGF-23 also has been investigated as an early biomarker of CKD-MBD in humans. [72][73][74] An increase in FGF-23 concentration is an early abnormality detected in people with early CKD, preceding increased serum phosphate and  77 Furthermore, pFGF-23 was increased in non-azotemic cats with SDMA concentrations >14 μg/dL despite an absence of hyperphosphatemia. 78 These findings indicate that FGF-23 might have diagnostic potential for early detection of CKD and early phosphate derangement in cats and that phosphate dysregulation may be ongoing in the early stage of CKD before azotemia and hyperphosphatemia occur. Plasma FGF-23 concentrations were also significantly higher in cats with azotemic CKD than in healthy cats, and these concentrations significantly increased with the severity of CKD. 37,38 Also, pFGF-23 concentration was identified as independent predictor of CKD progression (>25% increase in sCr) within 12 months of diagnosis of CKD on a population basis. 79,81 However, the studies described are population-based studies and overlap identified between groups indicates that it is difficult to extrapolate to an individual cat basis. 79 Tubulointerstitial nephritis commonly is found in CKD cats. 8 This explains why CKD cats usually have mild proteinuria, and mostly changes in tubular biomarkers are expected. On the other hand, marked proteinuria or markedly increased glomerular biomarkers are expected in primary glomerular disease, which is uncommon in cats, except for renal amyloidosis in predisposed cat breeds. 8,82,83 Other biomarkers also can be present in urine as a consequence of fibrosis or oxidative injury, or both reflecting the pathogenesis of CKD. 84 Urinary biomarker concentrations usually are corrected by determining their ratio to urinary creatinine concentrations to adjust for variation in urinary volume and concentration. 85 An overview of urinary renal biomarkers evaluated in CKD in cats based on their origin in the kidney is presented in Figure 1. Knowledge about urinary renal biomarkers in cats is summarized in Table 2.
Other glomerular biomarkers such as immunoglobulins G, M, and A as well as C-reactive protein and thromboxane B2 have been studied in dogs, but have not yet been investigated in cats. 29,105 Also, no recent information of other tubular biomarkers previously investigated in cats, namely urinary N-acetyl-b-D-glucosaminidase (NAG) and retinal-binding proteins (RBP) have been published since previous reviews on this topic. 29,105 Therefore, these biomarkers will not be discussed in the present review.
In general, circulating ALB cannot freely pass through an intact glomerular barrier because its size exceeds the glomerular permeability threshold. Small quantities of ALB that pass through the glomerulus into tubular fluid are completely reabsorbed by proximal tubular cells.
Consequently, healthy cats have <1 mg/dL of ALB in their urine (uALB). Thus, both glomerular and tubular dysfunction can result in albuminuria. 106 Microalbuminuria and overt albuminuria are defined as uALB concentrations between 1 to30 mg/dL and >30 mg/dL, respectively. 106,107 Albuminuria and increased uALB/Cr have been evaluated in AKI and CKD in dogs, 87,108,109 and microalbuminuria was associated with increased risk of death in dogs with critical conditions. 87 In the past, semi-quantitative techniques including urine dipstick colorimetric testing and sulfosalicylic acid turbidity testing were used as screening tools for albuminuria in cats. However, their poor specificity for the detection of albuminuria in cats decreases the utility of these tests. 86,110,111 In recent years, more accurate techniques have been validated to quantify uALB concentration in cats, including a feline-specific sandwich ELISA, a PETIA used in humans, and electrophoresis, all with acceptable precision and reproducibility. 20,33,86,90 However, the availability of these tests to general veterinary practitioners remains limited.
Several studies have demonstrated the diagnostic and prognostic value of uALB concentration for CKD in cats. Correlations between albuminuria and proteinuria were reported to be significant. 20,23,33,110 Similar to proteinuria, albuminuria was negatively associated with survival in a population of client-owned cats with or without CKD. 20 However, guidelines to predict survival in an individual cat based on severity of albuminuria currently do not exist. Cats with CKD and higher uALB/Cr ratios had shorter survival times than those with lower uALB/Cr ratios. 20 Urinary ALB/Cr ratios in cats with stage 1 CKD were significantly higher (2-fold) than in healthy cats, but unfortunately the study groups were not age-matched with a younger healthy group being compared to the CKD cats. 90 The other study showed that uALB/Cr had poor diagnostic value when differentiating cats with azotemic CKD from healthy cats, but it tended to be higher in cats with azotemic CKD. 33 In a more recent study, uALB/Cr was shown to have high sensitivity and specificity for classifying the severity of proteinuria in cats with various diseases. This ratio could distinguish healthy from diseased cats, but it could not differentiate between CKD cats and cats with non-renal diseases. 86 This result was supported by several studies in cats that detected microalbuminuria in a wide variety of non-renal diseases, suggesting that albuminuria is not specific for renal diseases in cats. 23,[87][88][89]110 Moreover, it has been suggested that the clinical importance of the uALB/Cr ratio has not yet surpassed that of the UPC, which is routinely available in commercial laboratories. 9,18 Therefore, the clinical implication of albuminuria in cats with renal diseases is unclear.

| Transferrin
Transferrin is an ion-binding glycoprotein synthesized in the liver and other tissues. Because its MW (77 kDa) is similar to that of ALB, transferrinuria is expected to be present when there is glomerular damage. In a study in humans, transferrinuria showed potential as a sensitive indicator of early diabetic nephropathy and a predictor of microalbuminuria development. 112 However, urinary transferrin concentration was not associated with the progression of renal disease in another study in humans. 113 A single study has investigated urinary transferrin concentrations in cats using sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). This study showed that cats with IRIS stage 1 CKD had higher urinary transferrin concentrations than did healthy cats.
With a cutoff of 0.93 mg/dL, urinary transferrin concentration was more sensitive and specific than plasma creatinine concentration for the detection of early-stage CKD. 90 Because of limited data availability, it is currently unknown whether increased urinary transferrin concentration reflects primarily tubulointerstitial nephritis or primary glomerular disease, taking into account that tubulointerstitial nephritis is the most common lesion in CKD cats.

| Neutrophil gelatinase-associated lipocalin
Neutrophil gelatinase-associated lipocalin (NGAL) is produced by neutrophils and epithelial cells in various tissues, including renal tubular cells. 114 It contributes to the innate immune response as a bacteriostatic agent and also aids in limiting damage after renal tubular injury. 115 Neutrophil gelatinase-associated lipocalin is expressed in response to infection, inflammation, and neoplasia. 116,117 Circulating NGAL passes freely through the glomerular barrier and is reabsorbed in the proximal tubules. 115,118 Renal injury impairs tubular reabsorption of NGAL and NGAL is released from the damaged renal tubular cells. Consequently, an increased NGAL concentration in the serum or urine may reflect tubular injury. [119][120][121] In humans, both sNGAL and uNGAL concentrations appear to be promising biomarkers for the detection and prediction of AKI as well as CKD. [122][123][124] In dogs, sNGAL is prognostic for survival in dogs with CKD and the urinary NGAL : creatinine ratio (uNGAL/Cr) has been suggested to be a useful biomarker for the detection of early-stage CKD and predicting CKD progression in dogs. [125][126][127][128] An in-house sandwich ELISA using anti-canine NGAL antibodies has been established for measuring pNGAL and uNGAL concentrations in cats, with acceptable precision and repeatability. 94 Recently, a commercial ELISA using anti-human NGAL antibodies has been validated for sNGAL and uNGAL measurements in cats. 91 In humans, uNGAL remains stable for at least 96 hours at 4 C and at least 6 months at À80 C, slowly decreasing with storage at À80 C over 5 to 8 years. 129 Urinary NGAL concentrations in both humans and dogs decrease only slightly after 3 freeze-thaw cycles. 6,130 The effect of biological status on NGAL concentrations has not been reported in cats, except for age, which does not influence uNGAL/Cr. 94  To date, studies that evaluate NGAL as early renal biomarker in CKD cats are limited, and conflicting results have been found. In a study evaluating uNGAL concentration in 80 cats with azotemic CKD and 18 healthy cats, uNGAL/Cr ratio was significantly increased in the former group, particularly in cats with IRIS stages 3 and 4 CKD. 94 In contrast, a recent study evaluating 9 CKD cats and 44 healthy cats reported that uNGAL/Cr was not different between the groups. The reason for the discordant findings between these studies is unclear, but might be because of the different assays used. 91 The former study also assessed the prognostic utility of uNGAL concentrations in CKD cats and found higher baseline uNGAL/Cr concentrations in the group with progression compared to the group without progression. 94 Unfortunately, this finding must be interpreted cautiously because of important study limitations such as the different severity of azotemia between both groups at baseline and the short duration of follow-up (30 days 95 Its presence may reflect advanced tubular injury or impairment, which may be useful as a prognostic factor in cats with AKI or CKD. The monomer was also the predominant form of uNGAL detected in a group of non-pyuric cats with non-renal diseases. 95 The monomeric form is valuable primarily to investigate the diagnostic potential of NGAL in AKI and as biomarker for progression of CKD. The dimeric form is predominantly present in cats with UTI and pyuria, emphasizing that UTI should be ruled out when interpreting uNGAL concentrations in cats. 95 Interestingly, the NGAL/ MMP9 complex is commonly found in the urine of cats independent of their health status, which is a notable difference compared to humans and dogs. 95,[132][133][134] This finding may infer that healthy cats physiologically express NGAL/MMP9 complex, or unknown pathological conditions are present in these cats.
Based on results of the available studies, uNGAL may relate more to the severity of azotemia or renal impairment rather than the type and progression of kidney disease. Also, the limitation of using nonfeline specific assays for NGAL measurements in cats must be considered.

| Liver-type fatty acid-binding protein
Liver-type fatty acid-binding protein (L-FABP) is a 14-kDa protein located in the cytoplasm of hepatocytes and proximal tubular epithelial cells. 135,136 It also is expressed in the intestine, lung, and pancreas. 137 Circulating L-FABP normally passes through the glomerular barrier and is reabsorbed in the proximal tubules. In the kidneys, L-FABP plays a major role in fatty acid homeostasis, resulting in fatty acid catabolism that provides energy to proximal tubular cells. 136,138 Moreover, it has protective effects on the proximal tubules by transporting harmful lipid peroxidation products from the cytoplasm into the tubular lumen. 136,139 Ischemic injury and oxidative stress of the proximal tubules result in L-FABP excretion into the urine in small laboratory animals and humans. 135,140 In the clinical setting, uL-FABP/ Cr ratio is a biomarker of early tubular damage that can reflect the severity of tubulointerstitial injury, detect AKI, and predict CKD progression in humans. [141][142][143][144][145][146][147] Furthermore, L-FABP has been shown to be renoprotective in small laboratory animals. 148,149 A commercial feline L-FABP ELISA based on cross-reactivity with anti-human L-FABP has been used in research on cats and recently has been validated for feline serum and urine. [91][92][93] Human uL-FABP was stable at 4 C for 48 hours and at À70 C for at least 18 months. 150,151 The stability of L-FABP in samples from cats has not been reported.
In humans, other non-renal factors are known to influence uL-FABP concentrations. 141,143,[152][153][154][155][156] In non-diabetic human patients, age, sex, and serum cholesterol concentration are not associated with uL-FABP/Cr, whereas this biomarker significantly increases in anemic patients. 154 Proteinuria, albuminuria, and hematuria are strongly associated with uL-FABP/Cr, whereas pyuria has very little impact. 141,143,153,156 Liver disease significantly influences sL-FABP concentrations but not uL-FABP/Cr. Also, sL-FABP concentration did not correlate with uL-FABP/Cr in humans with CKD and liver disease. 152 Nevertheless, uL-FABP/Cr was correlated with various liver enzymes in critically ill humans. 155 In cats, the current information about the effect of non-renal factors on L-FABP is limited. It is only known that uL-FABP/Cr is significantly increased in hyperthyroid cats, and resolved after euthyroidism was established. 91 It remains unknown whether uL-FABP concentration can predict post-treatment azotemia in hyperthyroid cats because of a low number of post-treatment azotemic cats in this study. 91 Currently, only 3 studies have evaluated L-FABP in cats. [91][92][93] It was detected in the cytoplasm of proximal tubular cells in healthy feline kidneys using immunohistochemistry with antibodies targeting human L-FABP, and it was expressed in the tubular lumen in response to ischemic-reperfusion injury. 92

| Kidney injury molecule-1
Kidney injury molecule-1 (KIM-1) is a type-1 membrane glycoprotein. 157 It is a cell surface receptor in epithelial, lymphoid, and myeloid cells that scavenges oxidized circulating low-density lipoproteins and membrane-associated phosphatidylserine. 158 No KIM-1 protein was detectable in the urine of healthy humans. 159 With renal tubular damage, KIM-1 is upregulated in proximal tubular cells. 160,161 Several studies in humans have reported an association between uKIM-1 and both AKI severity and rapid CKD progression. [159][160][161][162][163][164] In dogs, uKIM-1 has been suggested as a promising biomarker for the detection of both naturally occurring AKI and CKD. 165,166 Feline KIM-1 has been detected in urine using the commercially available anti-rat KIM-1 lateral flow immunoassay. 167 More recently, the same research group also developed a specific lateral flow assay for the detection of feline uKIM-1. 96 In mice, uKIM-1 is stable for 5 days at room temperature and at least for 1 year at À80 C, and multiple freeze-thaw cycles do not significantly alter uKIM-1 concentrations. 168 The effects of storage time and temperature on uKIM-1 concentrations in cat samples remain unknown.
Regarding the influence of non-renal factors, age may positively affect uKIM-1 concentrations in humans, 169,170 as might ethnic origin, with those of white European ancestry having higher uKIM-1 concentrations than African Americans. 171 A difference in uKIM-1 concentrations between the sexes in humans is debatable. 169 In their second study, the authors found that healthy cats had no positive KIM-1 immunohistochemical staining, whereas cats with experimental and naturally occurring AKI showed increased expression of KIM-1 in the proximal tubules. 173 Their recent third study showed an overlap between uKIM-1 concentrations in cats with suspected AKI and healthy cats using an in-house feline-specific KIM-1 immunoassay. The concentrations in cats with suspected AKI were highly variable, whereas healthy cats had low and non-variable uKIM-1 concentrations. Some cats (31%) with conditions associated with AKI had a substantial but transient increase in uKIM-1. Only 1 CKD cat expressed uKIM-1 at lower levels than healthy cats. 96 Overall, uKIM-1 is highly expressed after early severe tubular injury but further tubular loss may result in negative results. Therefore, kinetic monitoring of this biomarker using serial sample analysis might be important in practice. Finally, KIM-1 seems to have greater potential in the AKI setting compared to the CKD setting in cats, but further evaluation of uKIM-1 in CKD cats is warranted.

| Vascular endothelial growth factor
Vascular endothelial growth factor (VEGF), also known as vascular permeability factor, is a signaling protein affecting angiogenesis. 174 It is expressed in renal proximal tubular cells in response to hypoxia. 175 In normal human kidneys, VEGF is constantly produced to maintain the glomerular and peritubular vasculature. 176 Urinary VEGF expression is downregulated in progressive CKD in humans. [176][177][178][179] Thus, uVEGF is a potential prognostic indicator for kidney function loss.
Very few studies on VEGF in CKD have been performed in cats. 97,98,180 The commercially available human VEGF ELISA crossreacts with its feline homolog and has been validated for the detection of feline uVEGF. 180 In CKD cats, uVEGF concentrations were significantly lower than in healthy cats, with considerable overlap between groups. The association between uVEGF/Cr and sCr was not significant in these CKD patients. 97 Hyperthyroid cats also had significantly higher uVEGF/Cr ratios than healthy cats. Upon treatment, uVEGF significantly decreased but it remained significantly higher than in healthy cats. Moreover, hyperthyroid cats with post-treatment renal azotemia had lower uVEGF/Cr both pre-and post-treatment compared to those that remained non-azotemic after treatment. Additionally, uVEGF/Cr was negatively correlated with sCr but positively correlated with plasma total thyroxine concentrations, plasma renin activity, and UPC in hyperthyroid cats. 98 Consequently, increased uVEGF/Cr in hyperthyroid cats may not reflect renal dysfunction, whereas abnormal uVEGF/ Cr may indicate decreased renal mass or function in hyperthyroid cats. 98 Overall, the higher uVEGF/Cr in non-azotemic compared to CKD cats may indicate a renoprotective effect of VEGF in healthy kidneys, and a low uVEGF/Cr in healthy cats may indicate early CKD. 97 Based on these findings, the exact role of VEGF in the feline kidney remains unclear and additional studies are required to determine the diagnostic and prognostic value of uVEGF in early CKD in cats.

| Urinary cystatin C
Serum CysC concentration was discussed in Section 2.1.1 as a GFR biomarker. However, tubular renal damage decreases tubular reabsorption of CysC, resulting in increased excretion. 181 Therefore, uCysC concentration has been suggested as a promising biomarker for the assessment of renal tubular function in humans and dogs. In humans, uCysC/ Cr was found to be useful as a tubular damage biomarker in patients with CKD. 181 In dogs, uCysC/Cr can be used to distinguish dogs with renal disease from dogs with non-renal disease. 182 The anti-human-based PETIA and PENIA kits have been validated for uCysC in cats, although the former has poor repeatability in the lower concentration range. [31][32][33] In humans, uCysC is stable at both À20 and 4 C for 7 days and even at 20 C for 48 hours. 183 Urinary CysC concentration is increased in humans with proteinuria, but the effect of proteinuria on uCysC concentration has not been studied in cats. 184 Urinary CysC/Cr ratios were significantly increased in hyperthyroid cats but not in cats with diabetic mellitus. 24,185 One study found that uCysC/Cr was significantly higher in CKD cats compared to healthy cats. Nevertheless, 34% of CKD cats had undetectable uCysC concentrations vs 88% of healthy cats. 32 A more recent study observed that uCysC/Cr was significantly lower in CKD cats compared to healthy cats. 33 These different findings among studies might be a consequence of different CKD stages of the cats included or different analytical methods used. In the former study, CKD cats had more advanced disease and higher median UPC than those in the latter study. 32,33 Moreover, the usefulness of uCysC/Cr as a screening test to distinguish azotemic CKD cats from healthy cats was not better than USG. 33 These poor clinical correlations may be a result of the disadvantage of the current assays such as sCysC. Therefore, with the currently available assays, uCysC does not seem to be a reliable biomarker for detection of early CKD in cats.

| Heat shock protein-72
Heat shock protein-72 (HSP-72) is a stress-induced HSP isoform, which plays a role in protecting cell and protein structure upon cellular insult and stress conditions. 186,187 It is synthesized by a variety of cells including renal tubular cells. 188 In children undergoing renal transplantation, urinary HSP72 excretion was substantially increased during the early post-transplant period. 189 In dogs, uHSP72/Cr was shown to be a potential renal biomarker for AKI. 190 A commercial HSP72 ELISA kit for cats has been validated by its manufacturer and used in a recent study in cats. 99  These lipids are reliable indicators of oxidative injury and lipid peroxidation. 192 Oxidative stress can induce tubulointerstitial injury and fibrosis. 193,194 Plasma F 2 -isoprostane concentrations were increased in human CKD patients, [195][196][197][198] whereas uF 2 -isoprostane concentrations in humans with CKD were lower than in healthy humans. 199 Another study demonstrated that uF 2 -isoprostane concentrations correlated positively with eGFR in elderly men. 200 Affinity column purification followed by a competitive ELISA has been used to quantify uF 2 -isoprostane concentrations in cats. This method weakly correlates with gas chromatography/mass spectrometry, which remains the gold standard for isoprostane measurement. 201 Thus, the validity of the data assessed using the ELISA method remains doubtful. Hypertension and proteinuria were not associated with uF 2 -isoprostane concentrations in cats. 100

| Others
Novel biomarkers in veterinary medicine are cystatin B (CysB) and urinary clusterin. Cystatin B is a low-molecular weight protein (11 kDA) that functions to inhibit members of the cysteine proteases (family 1).
Cystatin B is mainly an intracellular protein with limited concentration in the circulation. 27 Clusterin is a glycoprotein (70-80 kDA) produced in various tissues during several physiological and pathological processes. 27,202,203 Normally, clusterin is found in very low concentration in urine. Urinary clusterin (uClust) concentration is increased during tubular injury in dogs. 204,205 In people, uClust is part of renal biomarker panels for drug development and toxicity. 27 For both markers, concentrations can be determined in dogs and cats using sandwich immunoassays. 27 To accurately detect active tubular injury in samples from dogs and cats, the kidney-specific isoform of clusterin is measured. 27 Studies have focused on the diagnostic potential of these markers in dogs and cats with AKI. Urinary CysB is increased in dogs with AKI and is shown not to be affected by UTI. 27,206 Similarly, an abstract reported that cats with AKI had significantly higher uCysB concentrations than healthy cats and that uCysB can predict AKI with 90% sensitivity and 92% specificity. 207 In dogs, uClust concentration is increased in AKI. 27,206,208 The only study evaluating uCysB and uClust concentrations in CKD cats did not find a difference in the concentrations of those and other renal biomarkers between cats treated with low-dose meloxicam and cats treated with placebo. 209 Current scientific information in cats is too limited to draw conclusions on the clinical utility of CysB and clusterin to diagnose CKD in cats.

| Biomarkers of renal fibrosis
The most common histopathological diagnosis in CKD in cats is interstitial inflammation and fibrosis. 210 Among the histopathological findings in CKD in cats, interstitial fibrosis is most closely associated with azotemia severity, hyperphosphatemia, anemia, and proteinuria. 8 These findings suggest that renal fibrosis and its mediators play crucial roles in CKD pathogenesis. Renal biomarkers indicating renal fibrosis therefore may be of benefit for the early detection and treatment of CKD in cats.

| Transforming growth factor-β1
Transforming growth factor-β1 (TGF-β1) is a multifunctional cytokine that potentially plays a role as a profibrotic mediator activating tissue fibrosis. 211 It is produced by parenchymal and inflammatory cells in several organs and especially the kidneys. [212][213][214][215][216] In the kidneys, TGF-β1 initially is released in an inactive form in response to various renal insults such as oxidative stress, hypoxia, proteinuria, and products of the renin-angiotensin-aldosterone system. [217][218][219][220] Inactive TGF-β1 then must be altered to its active form by proteolytic cleavage before promoting renal fibrosis. 221,222 The profibrotic effect of TGF-β1 consists of myofibroblast formation, which promotes extracellular matrix production, decreased extracellular matrix degradation, and tubular injury and cellular apoptosis. 84,[223][224][225][226] It also has an important role in tissue homeostasis, recruiting stem cells in tissue reparative processes. 227 Increased urinary TGF-β1 (uTGF-β1) expression is found in rodents and humans with kidney disease related to interstitial fibrosis. 228,229 Overall, studies in humans have shown significant upregulation of uTGF-β1 expression in patients with glomerular diseases. [228][229][230] Urinary TGF-β1 concentrations were significantly correlated with the severity of interstitial fibrosis in renal biopsy samples but did not correlate with sCr and GFR. 228,230 In CKD cats, a commercial multispecies total TGF-β1 ELISA has been used to measure total urinary TGF-β1 concentrations in cats.
However, no validation data have been reported for feline urine using this ELISA. 97,102 A commercial active TGF-β1 (aTGF-β1) ELISA for humans, which detected an active form of TGF-β1 in urine, later was validated for cats with acceptable precision and reproducibility. 103 In rodent models, renal interstitial fibrosis was associated with aging, resulting in increased TGF-β1 expression in the kidneys. 231 However, it is unknown whether this phenomenon occurs in cats.
Proteinuria may influence uaTGF-β1 concentrations, because significant correlation was found between log UPC and uaTGF-β1 : creatinine ratio (uaTGF-β1/Cr). 103 Few studies have evaluated TGF-β1 for assessing renal fibrosis in CKD cats. Cats with CKD had higher urinary total TGF-β1 concentrations than did healthy cats, whereas serum TGF-β1 concentration was not different between groups. 97,102 Urinary total TGF-β1/Cr also was positively correlated with sCr concentrations. 97 In contrast, another study was unable to measure total TGF-β1 concentrations in feline urine using the same ELISA. 103 Moreover, baseline uaTGF-β1/Cr ratios were not different between CKD cats and healthy cats, suggesting that CKD cats might not express more uaTGF-β1 than healthy cats. 103 However, non-azotemic cats that later developed renal azotemia had a 2.6-fold increase in uaTGF-β1/ Cr from baseline approximately 6 months before azotemia was detected. These results suggest that rather than sampling at a single time point, serial measurements in non-azotemic geriatric cats of uaTGF-β1/Cr may be useful to predict azotemic CKD later in life. 103 Additionally, uaTGF-β1/Cr moderately correlated with the severity of renal interstitial fibrosis in this histopathologic study. 103 In vitro, TGF-β1 induced the expression of genes associated with the pathogenesis of renal fibrosis in the proximal tubular epithelial cells of cats. 232 These findings support that TGF-β1 might be involved in the pathogenesis of CKD in cats by inducing pro-fibrotic factors related to renal fibrosis, and uaTGF-β1 expression may reflect renal fibrosis severity in cats. Overall, TGF-β1 may be promising to identify the likelihood of development of azotemic CKD in non-azotemic geriatric cats.

| Procollagen type III amino-terminal propeptide
Procollagen type III amino-terminal propeptide (PIIINP), a LMW (44 kDa) peptide, is cleaved and released from collagen into extracellular fluids including blood during the synthesis and deposition of type III collagen. Circulating PIIINPs can pass through glomeruli and are reabsorbed by proximal tubular cells. 233 In the absence of renal damage, only a small amount of PIIINP can be detected in human urine.
However, urinary PIIINP expression is increased in renal fibrosis. 234 This molecule has been studied for its utility as a novel biomarker of fibrosis in various organs such as lung, heart, liver, and kidney. Many studies in humans and dogs have shown that plasma PIIINP concentrations are increased in patients with diseases characterized by fibrosis. [235][236][237][238][239] The urinary PIIINP : creatinine ratio (uPIIINP/Cr) is negatively correlated with eGFR and positively associated with the progression of CKD and severity of renal fibrosis in humans. [240][241][242][243] A single study has evaluated PIIINP in small groups of cats using a sandwich ELISA. No validation data of the ELISA was reported. The study showed significantly increased uPIIINP/Cr in cats with azotemic CKD compared with healthy cats. The uPIIINP/ Cr ratio also was highly correlated with renal parenchyma stiffness as determined by ultrasonic shear-wave ultrasonography. 104 Currently, information of PIIINP in cats remains scarce. Additional studies using validated assays are required to establish whether PIIINP is a promising biomarker for CKD in cats. ACKNOWLEDGMENT Thirawut Kongtasai is supported by a grant from Mahidol University, Thailand, for studying the doctoral program at Ghent University.

CONFLICT OF INTEREST DECLARATION
Authors declare no conflict of interest.

OFF-LABEL ANTIMICROBIAL DECLARATION
Authors declare no off-label use of antimicrobials.

INSTITUTIONAL ANIMAL CARE AND USE COMMITTEE (IACUC) OR OTHER APPROVAL DECLARATION
Authors declare no IACUC or other approval was needed.

HUMAN ETHICS APPROVAL DECLARATION
Authors declare human ethics approval was not needed for this study.