Expression of osteogenic proteins in kidneys of cats with nephrocalcinosis

Abstract

Background

Nephrocalcinosis is a common pathological finding in cats with chronic kidney disease and nephrolithiasis. Understanding its pathogenesis may identify future therapeutic targets.

Hypothesis

Nephrocalcinosis is associated with expression of an osteogenic phenotype.

Animals

Kidneys with medullary mineralization were obtained from 18 cats (10 with and 8 without nephroliths) undergoing necropsy.

Methods

Cross‐sectional study. Microradiography and histopathology (modified von Kossa stain) were used to confirm parenchymal mineralization. Immunohistochemistry for 5 osteogenic markers was performed to determine their co‐localization with nephrocalcinosis. The proportion of kidneys with stronger immunointensity in mineralized versus non‐mineralized regions was analyzed using 1‐tailed sign tests. The proportion of kidneys with co‐localization of nephrocalcinosis and each marker was compared between kidneys with and without nephroliths using Fisher's exact tests.

Results

Nephrocalcinosis co‐localized with osteopontin immunoreactivity in all 18 cats (100%) and with osteocalcin in 12 cats (67%). Both osteogenic markers had stronger immunointensity in mineralized regions compared with non‐mineralized regions. Limited co‐localization was observed with other markers: bone morphogenic protein‐2 in 2 kidneys (both with nephroliths) and tissue non‐specific alkaline phosphatase in 1 kidney (without nephroliths); runt‐related transcription factor‐2 was undetected. No statistically significant differences were found in the co‐localization of nephrocalcinosis with osteogenic proteins between kidneys with and without nephroliths.

Conclusions and Clinical Importance

Expression of osteogenic proteins in areas of nephrocalcinosis indicates that nephrocalcinosis is associated with the development of an osteogenic phenotype. Targeting these processes could offer a novel approach to prevent nephrolithiasis at its origin.

Abbreviations

BMP2

bone‐morphogenic protein‐2

CaOx

calcium oxalate

CKD

chronic kidney disease

OCN

osteocalcin

OPN

osteopontin

RUNX2

runt‐related transcription factor‐2

TNAP

tissue non‐specific alkaline phosphatase

1. INTRODUCTION

Nephrocalcinosis, the deposition of minerals in the renal parenchyma, is a common pathological finding in cats with chronic kidney disease (CKD) and nephrolithiasis. ¹ , ² , ³ , ⁴ Over time, cats with CKD and nephrocalcinosis develop increased plasma concentrations of creatinine, phosphate, and fibroblast growth factor 23, suggesting that nephrocalcinosis might be a marker, a cause, or a consequence of CKD progression. ³ In humans, the extension of parenchymal mineralization beyond the urothelium and into the urinary space is associated with calcium oxalate (CaOx) urolith formation. ⁵ , ⁶ , ⁷ , ⁸ , ⁹ In cats, an association between renal mineralization and nephrolithiasis also has been observed. ¹⁰ These observations highlight the need to better understand the mechanisms underlying nephrocalcinosis as potential targets to mitigate CaOx urolith formation.

Extraosseous mineralization is classified as dystrophic or metastatic. ¹¹ Dystrophic mineralization is abnormal deposition of calcium salts associated with damaged or necrotic tissue. Metastatic mineralization is abnormal deposition of calcium salts in viable soft tissue as a consequence of disruption in calcium‐phosphate homeostasis. In cats with CKD, higher plasma and blood concentrations of total calcium and ionized calcium correlate with increased risk of nephrocalcinosis. ² , ³ The prevalence of nephrocalcinosis increases from 45% in normocalcemic to 81% in hypercalcemic CKD cats. ³ Higher blood ionized calcium, plasma phosphate, and plasma creatinine concentrations, and alanine aminotransferase activity are independent risk factors associated with nephrocalcinosis cats with CKD. ³ Despite recognizing these risk factors, the precise mechanisms of mineral deposition in the kidneys of these cats remain unexplored. One hypothesis suggests that medullary nephrocalcinosis might arise from transdifferentiated renal mesenchymal cells expressing an osteogenic phenotype. ¹² , ¹³ , ¹⁴ Human urolith‐forming kidneys with medullary nephrocalcinosis express osteogenic markers with membrane‐bound vesicles containing crystals associated with collagen matrix that resembles bone mineralization. ¹³ , ¹⁴

We hypothesize that kidneys of cats with nephrocalcinosis express osteogenic markers. The aim of our cross‐sectional study was to assess whether expression of osteogenic markers co‐localizes with medullary nephrocalcinosis and if this co‐localization is different between kidneys with and without nephroliths.

2. MATERIALS AND METHODS

2.1. Case selection

This cross‐sectional study was conducted on kidneys obtained from client‐owned cats undergoing necropsy at the University of Minnesota Veterinary Diagnostic Laboratory. Samples from cats were not specifically collected for the study but were residual samples obtained after medical evaluation or end‐of‐life procedures, with written permission from the owners. Kidneys were either harvested immediately after euthanasia or within 24 hours from carcasses promptly refrigerated after euthanasia. Once harvested, kidneys were decapsulated and preserved in 10% formalin.

2.2. Microradiography

Each kidney underwent high‐resolution microradiography to detect mineralization (Faxitron UltraFocus 26 kV, 0.3 mA, ×2‐×5 magnification). Two distinct forms of mineralization were identified. Medullary nephrocalcinosis was characterized by radiopaque striations in the renal medulla. Nephroliths were defined as radiopaque structures exhibiting globular shapes with well‐defined edges located within the renal pelvis or diverticula. Cats with bilateral kidney medullary mineralization were included. One kidney was randomly selected from each cat for further evaluation.

2.3. Histopathology

A 3‐5 mm transverse central slice from each formalin‐fixed kidney was paraffin‐embedded, and 4 μm thick serial sections were cut for staining and histological analysis. Adjacent sections were used for the assessment of histopathological diagnoses (International Veterinary Renal Pathology Service, College Station, Texas), mineralization, and osteogenic proteins. Kidneys exhibiting mineralization of tumors, cysts, or inflammatory masses on histological examination were excluded from the study. Modified von Kossa staining (SSK‐CALC, StatLab Medical Products, McKinney, Texas) was used to identify calcium deposition. Kidneys also were excluded if they had microradiographic mineralization but lacked histopathologic confirmation of medullary mineralization (ie, von Kossa negative).

2.4. Immunohistochemistry

Immunoreactivity for 5 osteoblast‐associated proteins (Runt‐related transcription factor‐2 [RUNX2], bone‐morphogenic protein‐2 [BMP2], tissue non‐specific alkaline phosphatase [TNAP], osteopontin [OPN], and osteocalcin [OCN]) was evaluated on mineralized kidneys using the following primary antibodies: mouse anti‐Runx2 (cat #SC‐390715 e, Santa Cruz Biotechnology, Dallas, Texas) at 1:50 dilution, rabbit anti‐BMP2 (cat #PB9687, Boster Biological Technology, Pleasanton, California) at 1:500 dilution, rabbit anti‐TNAP (cat #ab108337, Abcam, Boston, Massachusetts) at 1:250 dilution, rabbit anti‐OPN (cat #ab8448, Abcam, Boston, Massachusetts) at 1:500 dilution, and mouse anti‐OCN (cat #ab13418, Abcam, Boston, Massachusetts) at 1:150 dilution. The epitopes recognized by each antibody were obtained from manufacturer information and aligned to the feline proteins using the “Align” tool in the UniProt database and had alignments of 100%, 100%, 95%, 80%, and 64% with the homologous regions of feline RUNX2 (UniProt: A0A337SSZ3), BMP2 (UniProt: A0A2I2UEZ4), TNAP (UniProt: Q29486), OCN (UniProt: A0A5F5XF81), and OPN (UniProt: A0A337RZL8), respectively.

After deparaffinization and rehydration, all samples, except slides used for mouse anti‐OCN immunoreactivity, underwent heat‐induced antigen retrieval in a microwave oven for 10 minutes in citrate buffer (cat #S1699, Agilent Technologies, Inc, Santa Clara, California) and were allowed to cool to room temperature. Endogenous peroxidase was inactivated using 3% hydrogen peroxide incubation for 15 minutes at room temperature. Before exposure to primary antibodies, the samples were blocked for non‐specific binding by incubating them with 10% normal goat serum protein for 15 minutes at room temperature. Antibodies were incubated for 1 hour at room temperature, followed by overnight incubation at 4°C. Peroxidase 3‐Amino‐9‐Ethylcarbazole (cat #K3469, Agilent Technologies, Inc, Santa Clara, California) was utilized to detect mouse/rabbit primary antibodies (cat #K4001/K4003, Agilent Technologies, Inc, Santa Clara, California). Washing steps were carried out using Tris Buffered Saline (0.05 M, pH 7.6) with Tween 20. All samples were nuclear counterstained with Mayer's hematoxylin and coverslipped using a glycergel mounting medium (cat #C0563, Dako North America, Carpinteria, California). Immunoreactivity for each osteogenic protein was tested separately on formalin‐fixed paraffin‐embedded slides and on von Kossa‐stained slides in each kidney. Positive controls for RUNX2, BMP2, TNAP, OPN, and OCN consisted of sections obtained from osteosarcomas from dogs and cats, an established model known to express these osteogenic markers. ¹⁵ , ¹⁶ , ¹⁷ In addition to the mineralized kidneys, 5 macroscopically normal kidneys were obtained at necropsy to evaluate immunohistochemical patterns.

2.5. Digital image analysis

For microscopic assessment (HuronViewer, version 1.3.1.), slides were imaged using a slide scanner (Huron TissueScope LE, St. Jacobs, Ontario). The location of mineralization first was categorized as inner medulla, outer medulla, or cortex. Medullary mineralization was subcategorized as tubular (intraepithelial, intratubular, basement membrane) or interstitial in each slide stained with von Kossa.

The percentage of von Kossa‐stained area on whole slide images (%vk area) was measured using QuPath software (version 0.5.1‐arm64). To detect and measure the area of the kidney on the slide, a threshold was created at very low resolution (7.26 μm/pixel, Gaussian, 2 smoothing sigma) that included all kidney structures. Stain vectors for von Kossa were created for each slide using the deconvoluted image. The threshold then was adjusted at high resolution (0.88 μm/pixel, Gaussian, 2 smoothing sigma) to detect the von Kossa‐stained areas. The von Kossa staining score in the papillary area also was graded as follows: 0, normal; 2, <25%; 3, 50%‐75%; and 4, >75% affected.

Co‐localization of mineralization and osteogenic markers was assessed by visually comparing the sites of von Kossa staining and immunoreactivity in a series of slides with single and double staining. The immunoreactivity was semi‐quantitatively scored on immunohistochemical slides using H AEC color deconvolution 2 (ImageJ2 version 2.14.0). Three images (2040 × 1450 pixels) per zone (inner medulla, outer medulla, or cortex) were analyzed for mean gray values, resulting in 4 categories: negative (>181), weak (121‐180), moderate (61‐120), and strong (0‐60). ¹⁸ The evaluator was unaware of which kidneys were controls or cases and whether they had uroliths.

2.6. Statistical analysis

Clinicopathological data, when available, were presented as median (minimum, maximum). Categorical data were presented as percentages. The comparisons of continuous clinicopathological data and the percentage of mineralization between groups were assessed using the Wilcoxon Rank‐Sum test.

Kendall's Tau‐b correlation coefficient was calculated to assess whether the von Kassa mineralization scores correlated with the degree of immunoreactivity for each marker. Differences in immunoreactivity intensity between mineralized and non‐mineralized regions of the same kidney (paired samples) were calculated to determine if the intensity was stronger in the mineralized regions. This data was analyzed using a 1‐tailed sign test to assess whether the proportion of kidneys with stronger immunoreactivity intensity in mineralized regions exceeded 0.5. For each osteogenic biomarker, the number of kidneys with co‐localization of mineralization (von Kossa staining) and the biomarker (any category of positive staining) were determined for the groups with and without nephroliths. The proportions of samples with co‐localization were compared between these 2 groups using Fisher's exact tests. Analyses were performed using the “stats” package (version 4.3.2) in R software for statistical computing (version 4.3.2 GUI 1.8, R Foundation for Statistical Computing, Vienna, Austria). Statistical significance was defined as P < .05.

3. RESULTS

3.1. Microradiographic pattern of mineralization

Nineteen kidneys with microradiographic mineralization were evaluated for potential study inclusion: 11 with nephroliths and 8 without nephroliths. Radiopaque mineral striations extended from the cortex inwardly toward the renal papilla in all kidneys (Figure 1A‐D). Diffuse radiopaque punctate minerals were identified in 10 kidneys: 6 with nephroliths and 4 without nephroliths (Figure 1A,B).

FIGURE 1.

Faxitron microradiographs of kidneys before and after colorization (red) to identify parenchymal mineralization. Mineralization appeared as radiating striations extending from the cortex inward toward the renal papilla in the kidney with nephroliths (A and B) and without nephroliths (C and D). Punctate mineralization is represented in blue box (B). The radiopaque nephroliths were colored yellow.

3.2. Microscopic pattern of mineralization

In 1 kidney, mineralization was not detected histopathologically in any sections; this kidney was excluded from the study, resulting in 10 kidneys with nephroliths and 8 kidneys without nephroliths. In another kidney, histological preparation and microtome sectioning resulted in a loss of a portion of the inner medulla; this kidney was not excluded, but only the outer medulla and cortex were evaluated. Signalment and recent clinical data within 2 weeks of euthanasia were available in some of the cats (Table 1).

TABLE 1.

Descriptive statistics of clinicopathological variables in nephrocalcinosis cats, categorized by the presence of stones detected on microradiography.

Variables	Mineralized kidneys with nephroliths (n = 10)		Mineralized kidneys without nephroliths (n = 8)		P‐value
	Median [min, max]	n	Median [min, max]	n
Age (years)	15 [15, 17]	5	18 [4, 18]	3	.54
Female spayed (%)	4 [80%]	5	2 [66.7%]	3
Creatinine	3.3 [1.7, 20]	5	7.3 [5.4, 9.1]	2	.43
Total calcium	9.0 [8.1, 9.1]	3	9.9 [9.7, 10.1]	2	.2
Phosphorous	5.4 [5.3, 10.8]	3	11.6 [8.0, 15.1]	2	.4
von Kossa stained (%vk area)	0.06 [0.02, 1.66]	8	0.12 [0.001, 1.28]	7	.78

Note: Age and creatinine were compared using the Wilcoxon rank sum test with continuity correction.

In the inner medulla of all 17 case kidneys evaluated, mineralization was observed within renal tubules (tubular epithelia, lumens, and basement membranes) and within the interstitium (Figure 2A,B). In the outer medulla of 18 case kidneys, mineralization was observed within tubular lumens in 9 kidneys (Figure 2C,D) and in the interstitium in 2 different kidneys. In the cortex of the 18 kidneys, mineralization was observed infrequently (n = 1 kidney). Of the 18 case kidneys, 15 were suitable for quantification of %vk area (ie, the entire cross‐section of the kidney was mounted; Table S1).

FIGURE 2.

Photomicrographs of renal medulla stained with von Kossa for detecting minerals and mouse anti‐Runx2 for RUNX2 immunoreactivity. At the renal papilla, minerals were abundant (A; ×10 magnification). Panel (B) is the higher magnification (×40 magnification) of the insert in (A) showing mineralization primarily located at the basement membrane of renal tubules (arrow). At the outer medulla, fewer von Kossa‐positive regions were observed (C; ×10 magnification). Panel (D) is the higher magnification (×40 magnification) of the insert in (C), showing von Kossa‐stained minerals within tubules (arrow). The immunoreactivity of RUNX2 is negative.

Kidneys from 5 cats without clinical kidney disease were selected as normal controls. In 1 cat, the kidney was von Kossa‐negative with no histopathological abnormalities. In 2 cats, the kidneys were von Kossa‐positive with no additional histopathological abnormalities (Table S1). The remaining 2 cats had von Kossa‐positive kidneys with histopathological abnormalities consistent with CKD; these 2 controls were removed from the study.

3.3. Immunoreactivity for osteogenic proteins and their co‐localization with mineralization

The immunoreactivity of OPN was weak to moderate in the tubular epithelia of the cortex, outer medulla, and inner medulla of the 3 control kidneys (Table S2). In the 2 control kidneys with mineralization, moderate to strong OPN immunoreactivity co‐localized with mineralized regions. In all 18 case kidneys, OPN was weak to moderate in the tubular epithelia in non‐mineralized regions and was moderate to strong at sites co‐localized with von Kossa‐stained minerals (Figure 3A,B; Table S2). The immunoreactivity of OPN was significantly stronger in the mineralized region than in the non‐mineralized region (P < .001; 95% confidence interval [CI], 0.69‐1.00). No significant correlation was found between the area of mineralization (von Kossa scores) and the intensity of OPN immunoreactivity at the mineralized region (𝜏b = .03, P = .89). The proportion of kidneys showing co‐localization was not different between kidneys with and without nephroliths (Table 2).

FIGURE 3.

Photomicrographs of renal medulla immunoreactivity for individual osteogenic proteins (left panel) and double stained for the corresponding osteogenic protein and von Kossa (right panel). Osteopontin immunoreactivity co‐localized with von Kossa (A and B; ×20 magnification), osteocalcin immunoreactivity co‐localized with von Kossa (C and D; ×20 magnification). Bone morphogenic protein‐2 immunoreactivity co‐localized with von Kossa (E and F; ×20 magnification). Tissue non‐specific alkaline phosphatase immunoreactivity co‐localized with von Kossa (G and H; ×10 magnification). Runt‐related transcription factor 2 immunoreactivity not co‐localized with von Kossa (I and J; ×40 magnification).

TABLE 2.

Statistical analysis on the proportion of kidneys with co‐localization of osteogenic proteins and minerals between kidneys with and without nephroliths.

Osteogenic proteins	Proportion co‐localizing with mineral (#positive/total)
	Total	Nephroliths	No nephroliths	P‐value
Osteopontin	1.00 (18/18)	1.00 (10/10)	1.00 (8/8)	1.00
Osteocalcin	0.67 (12/18)	0.50 (5/10)	0.88 (7/8)	.15
Bone morphogenic protein‐2	0.11 (2/18)	0.20 (2/10)	0.00 (0/8)	.48
Tissue non‐specific‐alkaline phosphatase	0.06 (1/18)	0.00 (0/10)	0.13 (1/8)	.44
Runt‐related transcription factor‐2	0.00 (0/18)	0.00 (0/10)	0.00 (0/8)	1.00

The immunoreactivity of OCN was weak in the tubular epithelia of the cortex, outer medulla, and inner medulla of the 3 control kidneys (Table S2). In 1 of the 2 control kidneys with mineralization, moderate OCN immunoreactivity co‐localized with mineralized regions. Among the 18 case kidneys, OCN immunoreactivity was weak at tubular epithelium in non‐mineralized regions. In 12 of 18 case kidneys (5 kidneys with nephroliths and 7 kidneys without nephroliths), OCN immunoreactivity was moderate to strong when co‐localized with mineralized regions (Figure 3C,D; Table S2). Among kidneys with co‐localization, OCN immunoreactivity was significantly stronger in the mineralized regions compared with non‐mineralized areas (P = .003; 95% CI, 0.66‐1.00). The extent of mineralization (von Kossa scores) was not correlated with OCN intensity at the co‐localization sites (𝜏b = .02, P = .90). In 6 of 18 case kidneys, the mineralized regions showed negative OCN immunoreactivity. The proportion of kidneys showing colocalization was not different between kidneys with and without nephroliths (Table 2).

Three control kidney slides were lost during tissue processing for BMP‐2. In all 18 case kidneys, BMP2 immunoreactivity was weak to moderate in tubular epithelia of non‐mineralized regions. Bone morphogenic protein‐2 only colocalized with mineralized regions (moderate intensity) in 2 of 18 kidneys; both had nephroliths (Figure 3E,F; Table 2).

The immunoreactivity of TNAP was negative in the tubular epithelia of the cortex, outer medulla, and inner medulla of the non‐mineralized control kidney (Table S2). In the 2 mineralized control kidneys, 1 showed weak TNAP immunoreactivity in the tubular epithelia of the medulla; none showed TNAP co‐localization with mineralized regions. Among the 18 case kidneys, TNAP immunoreactivity varied from negative to moderate in the tubular epithelia of non‐mineralized regions; in only 1 case kidney (1 with nephroliths), TNAP immunoreactivity co‐localized with von Kossa‐stained minerals (Figure 3G,H).

Moderate RUNX2 immunoreactivity was observed in the perinuclear area of the renal vasculature in the cortex and outer medulla of the 3 control kidneys. Runt‐related transcription factor‐2 immunoreactivity was weak in the inner medullary vasa recta of the non‐mineralized and 2 mineralized control kidneys. Among the 18 case kidneys, RUNX2 immunoreactivity was weak in the cortex and outer medulla. Immunoreactivity of RUNX2 did not co‐localize with mineralized regions in case kidneys with and without nephroliths (Figure 3I,J).

4. DISCUSSION

Nephrocalcinosis is associated with nephrolithiasis and CKD in cats. ¹ , ² , ³ Its formation likely involves several physiologic and pathologic pathways and is incompletely understood. To determine if renal cells develop an osteogenic phenotype, we used immunohistochemistry to examine the expression of several osteogenic proteins and their co‐localization with mineralization in kidneys with and without nephroliths. In our study, osteopontin and osteocalcin had moderate to strong immunoreactivity that co‐localized with tissue mineralization in all (18/18) and most (12/18) kidneys, respectively. Co‐localization of nephrocalcinosis with the other osteogenic markers (BMP2, TNAP, and RUNX2) was infrequent to absent (Table 2).

Osteopontin, a non‐collagenous protein produced by osteoblasts, is important for mineralization, but its precise role is unclear. The negative charges of OPN aid in the recruitment of ionic groups (eg, calcium), and OPN's collagen‐binding properties allow it to serve as a nucleation scaffold during early bone mineralization. ¹⁹ Osteopontin knockout mice have a decrease in bone mineral crystal thickness, indicating its role in normal crystal growth during the production of mineralized matrix to form bone. ²⁰ Its co‐localization with mineral deposition in all 18 kidneys in our study suggests that it actively participates in renal mineralization in cats. Similar findings have been observed in the kidneys of humans with CaOx stones. ¹²

Interestingly, OPN also is associated with mitigating extraosseous tissue mineralization. Wild‐type mice fed a high phosphate diet develop less renal mineralization than OPN knockout mice fed the same diet. ²¹ Phosphorylation‐induced conformational changes in OPN may alter its function. ²¹ , ²² It is unknown whether the co‐localization of OPN expression with mineralization observed in our study is attributable to a causal role of OPN in the induction of mineralization or a compensatory role in the inhibition of mineralization. Additional studies are needed to determine the precise role of OPN in nephrocalcinosis in cats.

Osteocalcin, the most abundant non‐collagenous protein in bone, is produced solely by osteoblasts and is a biomarker of osteoblastic bone formation. ²³ It is primarily stored in bone in its carboxylated form, with an additional circulating uncarboxylated form. Carboxylated osteocalcin, distinguished by 3 γ‐carboxyglutamic acid residues, exhibits a high affinity for calcium. ²⁴ Positioned within intrafibrillar and interfibrillar collagen regions, OCN plays a crucial role in bone quality and strength by adjusting the alignment of biological apatite crystallites. ²⁵ In our study, the co‐localization of OCN with mineral depositions suggests its involvement in extraosseous renal mineralization. Similarly, OCN co‐localizes with mineralization of the renal medulla of human CaOx nephrolith formers. ²⁶ The exact role of OCN in extraosseous mineralization remains unknown. In bone, however, OCN synergistically works with OPN to form a structurally appropriate mineralized matrix. ²⁷ In our study, the co‐localization of both OPN and OCN supports that the expression of osteoblast‐like characteristics in the kidneys of cats may, in part, contribute to nephrocalcinosis. The absence of OCN immunoreactivity in mineralized regions in 6 of 18 case kidneys was unexpected. Variations in staining intensity or immunoreactivity may have decreased detection sensitivity, potentially resulting in amounts of OCN below the detection threshold. Although the spatial distribution of minerals was consistent across samples, ectopic mineralizationalso also can occur via alternative osteogenic and non‐osteogenic pathways, which may explain this unexpected finding. ²⁸

Osteogenic transdifferentiation previously has been proposed as a cause of renal mineralization. Renal interstitial cells, tubular epithelial cells, pericytes, and vascular smooth muscle cells can differentiate into osteoblast‐like cells in humans and rats. ¹² , ²⁹ , ³⁰ , ³¹ In these studies, protein expression and gene upregulation of osteogenic markers such as OPN and OCN have been observed. Mineralized renal papillae of humans with CaOx nephrolith overgrowth consist of collagen with morphologically preserved cell nuclei entrapped within mineralization, suggesting that renal cells undergo transdifferentiation and contribute to mineralization similar to osteocytes. ³²

Bone morphogenic protein‐2 belongs to the transforming growth factor‐beta superfamily. It stimulates osteoblast differentiation by activating canonical signaling pathways through Smad1/5/8 and non‐canonical signaling through p38 mitogen‐activated protein kinase pathways. These pathways ultimately regulate the expression of RUNX2, a critical transcription factor for osteoblast differentiation. ³³ , ³⁴ , ³⁵ The OLMALINC/OCT4/BMP2 axis enhances the osteogenic phenotype of renal interstitial fibroblasts in humans with renal mineralization. ³⁶ In our study, the reasons for the limited co‐localization of BMP2 and TNAP with mineralization and the absence of RUNX2 co‐localization are unknown. It could be attributed to these proteins functioning as markers of osteogenic differentiation primarily in the early and middle stages, with their expression decreasing in the late stage. ³⁷ , ³⁸ Alternatively, the protein expression levels might be beneath the detection threshold for immunohistochemistry labeling. False‐negative results attributable to primary antibody cross‐species detection are unlikely, given the high similarity between the sequence recognized by the antibody epitope and the homologous feline protein. Furthermore, antibodies successfully stained osteosarcoma‐positive controls from cats and non‐mineralized regions of kidney samples, supporting the reliability of the detection method.

Differences in semiquantitative immunoreactivity between mineralized and non‐mineralized regions suggest that osteogenic protein concentration or activity is different between these regions. The higher immunointensity of OPN and OCN suggests that these proteins are important for extraosseous mineralization. ³⁹ Additionally, their presence in both mineralized and non‐mineralized regions suggests they also may contribute to other physiological processes, potentially playing preparatory or regulatory roles before mineralization occurs.

We did not find significant differences in the expression of osteogenic proteins between kidneys with and without nephroliths. This observation may suggest that renal parenchymal mineralization in cats can occur from causes independent of nephrolithiasis. It is also possible that additional factors, which were not evaluated in our study, are necessary for initiating nephrolith formation over parenchymal mineralization. In humans, renal papillary mineralization, also known as Randall's plaque, is the initial site of nephrolith formation for many patients with idiopathic CaOx urolithiasis. ⁵ However, it is not the only process leading to nephrolithiasis. Although the severity of Randall's plaque is directly proportional to the incidence of CaOx nephroliths, not all patients with Randall's plaque develop nephroliths. ⁴⁰ , ⁴¹ , ⁴² It is also possible that some nephrolith‐forming cats may not yet have developed visible nephroliths. However, 2 of 3 cats with available signalment were 18 years old at the time of euthanasia. The highest risk for upper urinary tract urolithiasis occurred in cats between 4 and 8 years old, with a median age of 8 years old (interquartile range, 5‐11), suggesting that these 2 older cats would not have been latent nephrolith formers. ⁴³

Our findings resemble medullo‐papillary mineralization, which is known in humans as the initiating site for CaOx urolith growth through heterogeneous nucleation. ¹² , ¹³ , ¹⁴ In a high‐resolution X‐ray computed tomography study evaluating the medullo‐papillary complex in kidneys of humans, intratubular calcium phosphate aggregated in the outer medulla, along with simultaneous observation of interstitial spherical and fibrillar aggregates located distally near the papillary tip. ⁴⁴ , ⁴⁵ , ⁴⁶ , ⁴⁷ Ultrastructural examination of mineral particles in humans with CaOx nephroliths identified laminated particles in the basement membrane that formed aggregates in the interstitium of the papillary tip. ¹⁴ Consequently, papillary nephrocalcinosis in human CaOx nephrolith formers is suggested to originate from intratubular mineralization near the outer medulla, which precedes basement membrane mineralization and extends inward toward the renal papilla. ⁴⁷ However, our study did not yield direct evidence supporting this hypothesis. Although we observed intratubular mineralization in the outer medulla in some kidneys, determining a clear chronological sequence of events requires additional studies.

Our study had some limitations. Kidney specimens were obtained at necropsy. Although carcasses were immediately refrigerated and kidneys were retrieved within 24 hours of euthanasia, proteolysis of epitopes recognized by the antibodies used for immunohistochemistry was possible. However, all of our osteogenic markers were detected in almost all samples. Immunointensity can be affected by various factors throughout the process. ⁴⁸ , ⁴⁹ All samples were processed uniformly, and variability was further minimized by limiting the comparison of intensity to within individual slides. Consequently, comparisons of immunointensity between kidneys with nephroliths and those without were not performed to avoid potential confounding from variability among different slides. Another limitation is related to the examination of tissue sections and the inherent 3‐dimensional paraboloid shape of the medullo‐papillary complex. This complexity was highlighted when 1 kidney that appeared mineralized on microradiography had to be excluded because of negative testing for mineralization on microscopic examination. The remaining kidneys, however, showed centralized mineralization consistent with microradiography. Previous studies found that increased serum calcium or phosphate concentrations or increased serum calcium phosphate products were associated with renal mineralization in humans, cats, and rats. ² , ³ , ⁵⁰ , ⁵¹ , ⁵² , ⁵³ However, in our study, serum biochemistry data were not available for all cats at the time of euthanasia. The absence of clinical information made it difficult to determine whether mineralization was linked to specific clinicopathological abnormalities and their severity. Further research with larger sample sizes is needed to explore these underlying mechanisms of mineralization and clinical implications.

In conclusion, our findings suggest that nephrocalcinosis involves a cell‐mediated process characterized by an osteogenic phenotype with OPN and OCN expression. In humans, the extension of medullary mineralization beyond the urothelium is the site for nephrolith formation. Likewise, medullary mineralization in the kidneys of cats that extends to the renal papilla and breaches the urothelium into the urinary space also has the potential to become a nidus for nephrolith formation. Targeting the processes underlying renal mineralization could offer a novel approach to mitigating CKD progression and preventing nephrolith formation at its origin.

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

This study utilized organs from deceased animals as received and is declared exempt from IACUC review. All owners provided written informed consent for their animal's remains to be used for research purposes.

HUMAN ETHICS APPROVAL DECLARATION

Authors declare human ethics approval was not needed for this study.

Supporting information

Table S1: Histopathological findings in case and control kidneys ordered by increasing amount mineralization (vK score).

Table S2: The number of kidneys distributed by immunointensity of osteogenic markers and the presence of co‐localization with mineralization in control and case kidneys with and without nephroliths.

Hengtrakul N, Furrow E, Borofsky M, Toth F, Lulich JP. Expression of osteogenic proteins in kidneys of cats with nephrocalcinosis. J Vet Intern Med. 2025;39(1):e17278. doi: 10.1111/jvim.17278