Genome-Wide Gene Expression Profiling of Randall’s Plaques in Calcium Oxalate Stone Formers

Abstract

Randall plaques (RPs) can contribute to the formation of idiopathic calcium oxalate (CaOx) kidney stones; however, genes related to RP formation have not been identified. We previously reported the potential therapeutic role of osteopontin (OPN) and macrophages in CaOx kidney stone formation, discovered using genome-recombined mice and genome-wide analyses. Here, to characterize the genetic pathogenesis of RPs, we used microarrays and immunohistology to compare gene expression among renal papillary RP and non-RP tissues of 23 CaOx stone formers (SFs) (age- and sex-matched) and normal papillary tissue of seven controls. Transmission electron microscopy showed OPN and collagen expression inside and around RPs, respectively. Cluster analysis revealed that the papillary gene expression of CaOx SFs differed significantly from that of controls. Disease and function analysis of gene expression revealed activation of cellular hyperpolarization, reproductive development, and molecular transport in papillary tissue from RPs and non-RP regions of CaOx SFs. Compared with non-RP tissue, RP tissue showed upregulation (˃2-fold) of LCN2, IL11, PTGS1, GPX3, and MMD and downregulation (0.5-fold) of SLC12A1 and NALCN (P<0.01). In network and toxicity analyses, these genes associated with activated mitogen-activated protein kinase, the Akt/phosphatidylinositol 3-kinase pathway, and proinflammatory cytokines that cause renal injury and oxidative stress. Additionally, expression of proinflammatory cytokines, numbers of immune cells, and cellular apoptosis increased in RP tissue. This study establishes an association between genes related to renal dysfunction, proinflammation, oxidative stress, and ion transport and RP development in CaOx SFs.

The prevalence of kidney stone disease is nearly 9% in the adult population and continues to increase worldwide.^1,2 This condition has a medical and economic impact³ and is reported to be associated with complications such as metabolic syndrome (MetS)^4,5 and ESRD.⁶ The pathogenesis of kidney stone formation has been investigated, and there are two major theories for predicting lithogenesis.⁷ One pathway involves formation of intra-tubular crystals in the duct of Bellini, the so-called Randall plugs, as observed with both experimental hyperoxaluria-induced animal models⁸ and human primary hyperoxaluria and with calcium phosphate (CaP), struvite, in addition to some idiopathic calcium oxalate (CaOx) stone formers (SFs). The other pathway involves overgrowth on interstitial apatite plaques, the so-called Randall plaques (RPs),⁹ as observed in some idiopathic CaOx SFs.

In a recent study using genome-wide analysis and genome-recombined mice, we found OPN expression in renal tubular cells and macrophage (Mϕ) migration in the interstitial space around crystals to be essential for stone formation.^10–12 We found that the anti-inflammatory phenotype Mϕ played a suppressive role in kidney stone formation via renal crystal phagocytosis.¹³ Differentiation and induction of anti-inflammatory Mϕ are considered a potential therapeutic approach for kidney stone disease; however, this evidence is only applicable to Randall plugs, which have similarities with the hyperoxaluric mouse model. Understanding of RPs is also essential to clarify the potential of molecular therapies, such as OPN and Mϕ-related genes.

Regarding the origin of some idiopathic CaOx kidney stones, Evan et al. made major contributions to the study of the microscopic structure of RPs, which begin in the basement membranes of thin loops of Henle with calcium deposits.¹⁴ Despite numerous studies involving animal hyperoxaluric stone models and human samples,¹⁵ the exact role of RP in the formation of CaOx crystals remains unknown. Since morphologic, mineral, and matrix-based investigations provided the pathologic united theory of RP formation,¹⁶ molecular-level analysis of cellular function is necessary for better understanding of the role of RPs. The recently developed nephroureteroscopic technique permits more detailed analysis of RPs involving both microscopic and genomic analyses.¹⁷

Therefore, to establish molecular-targeted therapies for kidney stones, we investigated the gene expression profiles of RP sections from human papillary tissues and studied the factors controlling the development of RPs using microarray and immunohistochemical analyses.

Results

Patient Background

Patients who underwent percutaneous nephrolithotomy or retrograde intrarenal surgery for calcium-based stones were enrolled in this study. There were no statistical differences in the general background among the seven controls and 23 CaOx SFs such as age, gender, side of treatment, and body mass index. Based on the composition of the stone fragments obtained during lithotripsy, CaOx SF was defined as a patient with kidney and/or ureter stones that had >80% content of CaOx crystals. There were no significant differences in serum and urine parameters among the three groups (Table 1).

Table 1.

Patient backgrounds

Characteristics	Control (n=7)	CaOx SFs (n=23)	P Value
General
Age, y	56±18	59±13	NSa
Gender, M/F	4:3	14, 9	NSc
Side, Rt/Lt	3:4	13, 10	NSc
BMI, kg/m2	22.1±2.9	23.3±4.7	NSa
Stone
Stone composition
CaOx, %	—	91.4±9.1	—
CaP, %	—	5.5±7.5	—
Stone volume, mm3	—	761±1528	—
Stone density, HU	—	820±375	—
Hydronephrosis, grade 0,I,II	5,1,1	12, 6, 5	NSd
Serum
BUN, mg/dl	12.8±3.7	14.5±3.8	NSb
Cre, mg/dl	0.9±0.4	0.9±0.2	NSb
Ca, mg/dl	9.4±0.7	9.4±0.6	NSb
P, mg/dl	3.2±1.0	3.1±0.3	NSb
UA, mg/dl	4.9±1.2	5.8±1.6	NSb
WBC, cells/μl	6514±1748	6783±1853	NSb
CRP, mg/dl	1.9±1.7	0.3±0.1	NSb
Urine			NSb
pH	6.8±0.6	6.6±0.7	NSb
WBC, cells/HPF	9±10	19±28	NSb
RBC, cells/HPF	16±37	19±29	NSb

Data are presented as the mean±SD. Grade of hydronephrosis was categorized by Society for Fetal Urology. NS, not significant; M, male; F, female; Rt, right; Lt, left; BMI, body mass index; —, inapplicable data; HU, Hounsfield unit; BUN, blood urea nitrogen; Cre, creatinine; Ca, calcium; P, phosphorus; UA, uric acid; WBC, white blood cell; CRP, c-reactive protein; HPF, high power field; RBC, red blood cell.

^aStatistical analyses performed by t test.

^bStatistical analyses performed by Mann–Whitney U test.

^cStatistical analyses performed by Fisher exact test.

^dStatistical analyses were performed by Kruskal–Wallis test.

Observation of RPs and the Surrounding Tissue

During the endoscopic intrarenal operation, RPs were observed as plain white calcification regions that were covered with the papillary epithelium when viewed through a nephroureteroscope. Some RPs coexisted with ductal plugs in the same renal calyx papilla. Hematoxylin-eosin staining showed destruction of the papillary epithelium layer and interstitial cellular disorder surrounding RPs. The RPs were positive for von Kossa staining but negative for Pizzolato staining, indicating that RPs contained CaP but did not have a CaOx component (Figure 1).

Figure 1.

Endoscopic and microscopic distribution of RPs. Representative photographs show renal papillary tissues from both normal and RP mucosa. The endoscopic image shows renal papilla mucosa in the upper calyx during retrograde intrarenal surgery. The normal papilla shows fleshy smooth mucosa without bleeding or calcification. Some RPs are showing as a white patchy lesion (arrow heads) as well as a ductal plug (arrow) within the same papilla. Micro tissues were stained with hematoxylin-eosin, von Kossa (for detection of CaP crystals), and Pizzolato (for detection of CaOx crystals) staining. *Location of RP. Magnification, ×400.

Energy dispersive x-ray (EDX) microanalysis revealed that the spectra of both calcium and phosphorus matched those for the RP region; other regions did not show spectra for both (Figure 2A). Transmission electron microscopy (TEM) showed that there were numerous collagen fibers in both interstitial cellular spaces around RPs and outside interstitial spaces around the basement membranes of renal tubular cells (Figure 2B). Immunohistochemical TEM showed much more diffuse and higher expression of OPN, considered to represent the matrix of CaOx and CaP stones, in RPs compared with both renal tubular cells and interstitial cells without RPs (Figure 2C).

Figure 2.

Ultrastructural observations obtained using EDX microanalysis and TEM. (A) EDX microanalysis of RPs. Upper images are microphotographs of non-, calcium (Ca)-, and phosphorus (P)-staining tissues. Lower images show spectra of carbon (C), oxygen (O), sodium (Na), Osmium (Os), Ca (arrow head), and P (arrow) for each tissue. L1, lesion 1 (the nonplaque area); L2, lesion 2 (another nonplaque area); RP, RP area. (B) Ultrastructural details of collagen fibers surrounding RP and normal renal tubular cells from non-RP lesion detected in papillary tissue by TEM. N, nucleus; arrow, basement membrane. Scale bar, 2 μm. (C) Immunoelectron microscopy analysis of OPN by TEM. OPN-positive area is indicated by black dots (arrow) on RPs and a renal tubular cell. Arrow, basement membrane. Scale bar, 1 μm.

Gene Profiling of Papillary Tissue of CaOx SFs and Controls

Microarray analysis was performed to compare the gene expression profiles of papillary tissue from nonstone patients (C group) and non-RP (N group) and RP (P group) tissue from CaOx SFs. Cluster analysis demonstrated that gene expression profiles of CaOx SFs (including both N and P groups) markedly differed from those of nonstone patients (C group) (Figure 3A). The scatter diagram showed significantly different gene expression between the C and N/P groups and smaller differences in gene expression between the N and P groups (Figure 3B).

Figure 3.

Comparison of gene expression in renal papillary tissues among RP and normal tissue from CaOx SFs and normal tissue from control patients by using microarray analysis. (A) Cluster analysis of the expression of all 50,599 genes on a human array chip. C group, normal papillary tissues from control patients; N group, normal papillary tissues from CaOx SFs; P group, RP papillary tissues from CaOx SFs. (B) Scatter plots of gene expression difference between C and N (left), C and P (center), and N and P groups (right).

Figure 4A shows the microarray comparison results for genes that showed a twofold increase or decrease in gene expression among P and N groups in CaOx SFs, with the C group as control patients. Compared with the C groups, the P and N groups had 6019 and 8274 genes with increased expression, respectively, and 70%–97% (5860 genes) of them were mutually shared with both P and N groups. In contrast, compared with the C groups, P and N groups had 451 and 577 genes with decreased expression, respectively, and 71%–91% (412 genes) of them were common with P and N groups. Additionally, 21 and ten genes in the P group showed increased and decreased expression compared with the N group, respectively (Figure 4).

Figure 4.

Scheme of the microarray analysis demonstrated diversity of gene expression difference among each group. (A) Comparison of each group and gene expression changes. Each extracted value represents the number of statistically different genes with >2.0- or <0.5-fold difference in expression between groups. P, RP papillary tissue group from CaOx SFs; N, normal papillary tissue group from CaOx SFs; C, normal papillary tissue group from control patients. (B) Venn diagram of a select number of genes that showed >2.0-fold difference in expression in both normal and RP papillary tissue of CaOx SFs compared with normal papillary tissue of control patients (left red figure), and the number of genes that showed <0.5-fold difference in expression in both normal and RP papillary tissue of CaOx SFs compared with normal papillary tissue of control patients (right blue figure).

Ingenuity canonical pathway analysis indicated significant activation of the following pathways in the N and P groups compared with the C group: cAMP-mediated signaling, coagulation system, gα signaling, extrinsic prothrombin activation pathway, and calcium signaling (Table 2).

Table 2.

High/Lowest activation z-score canonical pathways in both RP and normal papillary tissue, where gene expression differs from that in the control mucosa by at least twofold

Ingenuity Canonical Pathways	P Value	Ratio	z-Score	Molecules in Pathway
acAMP-mediated signaling	0.001	0.17	3.307	AKAP12,HTR5A,PTGDR,MC3R,PTGER3,PDE3A,PDE4A,TAAR1,CHRM3,HRH3,OPRL1,PPP3R2,MC5R,AKAP14,PPP3R1,CNR2,PDE4D,PDE11A,DRD3,MC2R,HTR6,RGS2,CALML5,GRM8,OPRM1,SSTR3,GNAI1,DRD5,RAPGEF3,GRM4,MC4R,XCR1,P2RY14,AKAP4,CNGB1,LPAR1,PDE5A,ADCY10
aCoagulation system	0.002	0.29	2.530	F11,BDKRB2,PLG,F9,PROS1,PROC,VWF,F7,TFPI,FGG
aGαs signaling	0.01	0.18	2.500	HTR6,RGS2,ADD2,HTR5A,MC3R,PTGDR,TAAR1,GNG13,DRD5,RAPGEF3,MC4R,CHRM3,GNB1,MC5R,GNB4,CNGB1,MRAS,ADCY10,GNG12,MC2R
aExtrinsic prothrombin activation pathway	0.01	0.31	2.236	PROS1,PROC,F7,TFPI,FGG
CREB signaling in neurons	0.40	0.11	2.138	POLR2F,CALML5,GRM8,GRID1,GRIA1,GRIK3,GNAI1,GNG13,GRM4,PRKCG,GNB1,GRM5,GNB4,PLCG2,MRAS,PLCB1,ADCY10,PLCL1,GNG12
aGαi signaling	0.02	0.17	1.807	OPRM1,PTGER3,GRM8,SSTR3,GNAI1,GNG13,GRM4,HRH3,OPRL1,XCR1,GNB1,GNB4,P2RY14,LPAR1,CNR2,CAV1,MRAS,ADCY10,DRD3,GNG12
Dopamine receptor signaling	0.28	0.13	1.633	MAOB,PPP1R14D,PPP2R5B,NCS1,DRD5,PPP1CB,SLC18A1,DRD3,ADCY10,CALY
Glutamate receptor signaling	0.06	0.18	1.342	GRM5,GNB1,CALML5,SLC17A7,GRM8,GRIA1,GRID1,GRIK3,DLG4,GRM4
Phospholipase C signaling	0.27	0.12	1.279	MYL10,PPP1CB,GNG13,PPP3R2,GNB1,GNB4,PPP1R12B,PPP3R1,MRAS,PLCB1,ARHGEF2,GNG12,CALML5,ARHGEF4,MEF2A,PLA2G1B,RAPGEF3,ARHGEF17,CD79A,MYL1,PRKCG,MYL9,RHOQ,PLCG2,MEF2D,ADCY10,ARHGEF10,FNBP1
Chemokine signaling	0.19	0.14	1.265	ROCK2,CCR3,CALML5,PPP1R12B,PLCG2,MRAS,GNAI1,PLCB1,PPP1CB,NOX1
Synaptic long term depression	0.29	0.12	1.213	GRM8,GRID1,GRIA1,PPP2R5B,PLA2G1B,GNAI1,PPP1R17,GRM4,PRKCG,GRM5,PRKG1,PLCG2,PLB1,GUCY1A2,MRAS,PLCB1,PLCL1
aCalcium signaling	0.02	0.15	1.155	CHRNA1,CALML5,TRPC1,GRIA1,MYH14,TRDN,MEF2A,TNNI3,CHRNB4,SLC8A2,TPM1,TPM2,ATP2A2,MYL1,TRPC7,ATP2B2,ATP2A1,MYL9,PPP3R2,CHRNG,ACTA2,ATP2B3,MEF2D,PPP3R1,TNNI1,SLC8A1,ATP2B4
PPARα/RXRα activation	0.48	0.11	1.147	CYP2C18,ACVR1,ADIPOQ,CYP2C9,BMPR2,ABCA1,CYP2C19,GK,PLCG2,INS,MRAS,TGFB2,TGFB3,SMAD4,PLCB1,PLCL1,ADCY10,ACVR1C,MAP4K4
Gα12/13 signaling	0.05	0.15	−1.414	VAV2,F2RL2,MYL10,CDH4,MEF2A,MYL1,CDH11,ROCK2,MYL9,CDH9,CDH12,LPAR1,VAV3,MEF2D,MRAS,CDH17,CDH8,CDH13

CREB, cAMP responsive element binding protein; PPARα/RXRα, peroxisome proliferator activated receptor alpha/retinoid X receptor alpha.

^aP<0.05 indicates significant difference.

Per disease and function analyses, upregulated genes common to both P and N groups were categorized based on cell/neuron hyperpolarization, fertilization, ion/carbohydrate/monosaccharide transport, duct cell differentiation, androstenedione modification, and endocrine cell depolarization. Downregulated genes common between P and N groups were categorized to obesity, cell attachment, tensile strength of the skin, leukemia, endocrine gland hypoplasia, anion homeostasis, phosphatidylserine distribution, glycemic control, and metabolic bone disease (Table 3).

Table 3.

High/Lowest activation z-score categories of disease and functions in both RP and normal papillary tissue gene expressions differed from control mucosa by twofold

Categories	Diseases or Functions Annotation	P Value	Activation z-Score	Predicted Activation State	Numbers of Molecules
Most increased
Cellular function and maintenance	Hyperpolarization	0.001	3.307	Increased	14
Cellular function and maintenance	Hyperpolarization of cells	0.001	2.752	Increased	8
Cellular function and maintenance	Hyperpolarization of neurons	0.002	2.591	Increased	7
Embryonic development, organismal development,  reproductive system development and function	Fertilization	0.01	2.425	Increased	28
Molecular transport	Transport of ion	0.002	2.343	Increased	86
Carbohydrate metabolism, molecular transport	Transport of carbohydrate	0.01	2.319	Increased	33
Carbohydrate metabolism, molecular transport	Transport of monosaccharide	0.01	2.243	Increased	29
Cellular development, tissue development	Differentiation of duct cells	0.001	2.236	Increased	5
Endocrine system development and function, lipid  metabolism, small molecule biochemistry	Modification of androstenedione	0.01	2.169	Increased	5
Cell morphology, cellular function and maintenance,  endocrine system development and function	Depolarization of endocrine cells	0.003	2.000	Increased	4
most decreased
Nutritional disease	Obesity	0.004	−2.730	Decreased	80
Cell-to-cell signaling and interaction	Attachment of cells	0.003	−2.566	Decreased	26
Hair and skin development and function	Tensile strength of skin	0.002	−2.195	Decreased	7
Cancer, hematologic disease, immunologic disease,  organismal injury and abnormalities	Leukemia	0.001	−2.183	Decreased	233
Developmental disorder, endocrine system  disorders	Hypoplasia of endocrine gland	0.006	−2.157	Decreased	8
Cellular function and maintenance, small molecule  biochemistry	Homeostasis of anion	0.001	−2.078	Decreased	15
Carbohydrate metabolism, lipid metabolism,  molecular transport, small molecule biochemistry	Distribution of phosphatidylserine	0.01	−2.000	Decreased	4
Organismal functions	Glycemic control	0.001	−2.000	Decreased	6
Connective tissue disorders, metabolic disease,  skeletal and muscular disorders	Metabolic bone disease	0.001	−2.000	Decreased	18

The top 100 up- or downregulated genes and top networks in the N and P groups compared with those in the C group are shown in Supplemental Tables 1–3.

Gene Profiling of RP Tissue in CaOx SFs

The top eight genes that were upregulated by >2-fold or downregulated by <0.5-fold in the P group compared with the N group of CaOx SFs are listed in Table 4.

Table 4.

Top upregulated and downregulated genes comparing RP with normal papillary tissue

Agilent ID	Gene Symbol	Entrez Gene Name	Fold Change	P Value	Location	Type
Genes with increased  expression in plaque  mucosa
A_23_P169437	LCN2	Lipocalin 2	6.167	0.02	Extracellular space	Transporter
A_33_P3243887	IL11	IL 11	3.949	0.01	Extracellular space	Cytokine
A_24_P190472	SLPI	Secretory leukocyte peptidase inhibitor	3.407	0.03	Cytoplasm	Other
A_24_P208825	MUC4	Mucin 4, cell surface associated	2.826	0.04	Plasma membrane	Other
A_24_P64167	PTGS1	PG-endoperoxide synthase 1	2.824	0.02	Cytoplasm	Enzyme
A_33_P3369371	GPX3	Glutathione peroxidase 3	2.717	0.03	Extracellular space	Enzyme
A_23_P164047	MMD	Monocyte to macrophage differentiation-associated	2.390	0.01	Plasma membrane	Kinase
A_23_P253350	C8orf4	Chromosome 8 open reading frame 4	2.022	0.04	Other	Other
Genes with decreased  expression in plaque  mucosa
A_23_P150555	SCGB1D2	Secretoglobin, family 1D, member 2	−2.322	0.001	Extracellular space	Other
A_33_P3252003	KCNJ1	Potassium channel, inwardly rectifying subfamily J, member 1	−2.317	0.01	Plasma membrane	Ion channel
A_21_P0007591	NAV2	neuron navigator 2	−2.255	0.02	Nucleus	Other
A_23_P84666	GDPD1	Glycerophosphodiester phosphodiesterase domain containing 1	−2.213	0.02	Other	Enzyme
A_24_P136029	SLC12A1	Solute carrier family 12 (sodium/potassium/chloride transporter), member 1	−2.112	0.03	Plasma membrane	Transporter
A_23_P29057	KCNJ6	Potassium channel, inwardly rectifying subfamily J, member 6	−2.110	0.001	Plasma membrane	ion channel
A_33_P3230166	NALCN	Sodium leak channel, nonselective	−2.068	0.001	Plasma membrane	ion channel
A_33_P3253723	AQP1	Aquaporin 1 (Colton blood group)	−2.005	0.001	Plasma membrane	transporter

Network analyses of genes whose expression differed by >2-fold or <0.5-fold in the P group compared with the N group are shown in Supplemental Table 4. The top-scored network demonstrated upregulation of lipocalin (LCN) 2, IL 11, secretory leukocyte peptidase inhibitor (SLPI), mucin 4, PG-endoperoxide synthase (PTGS) 1, monocyte to macrophage differentiation (MMD), and chromosome 8 open reading frame 4, and downregulation of potassium channel inwardly rectifying subfamily J member 1, solute carrier family 12 member 1 (SLC12A1), and sodium leak channel nonselective (NALCN) were related to extracellular proinflammatory cytokine and intracellular signal pathways (Figure 5).

Figure 5.

Network function of inflammatory response, cellular movement, cell-to-cell signaling, and interaction between RP and normal papillary tissues in CaOx SFs (defined by either >2.0- or <0.5-fold difference in expression; P<0.01). Red shapes indicate upregulated mRNAs, while gray shapes indicate downregulated mRNAs. Different shapes and prediction outlines are indicated in the legend box.

IL1β and TNF were determined to be upstream cytokines upregulated among genes in the P group compared with the N group (Supplemental Table 5).

Toxicity analysis showed that LCN2, IL11, glutathione peroxidase (GPX) 3, and aquaporin (AQP) 1 were responsible for ARF, renal ischemia-reperfusion injury, cardiac hypertrophy, and oxidative stress (Table 5).

Table 5.

Toxicity analysis of genes which expressed over twofold differences between RP and normal papillary tissue

Ingenuity Toxicity Lists	P Value	Ratio	Genes in Lists
ARF panel (rat)	0.001	0.03	LCN2, AQP1
Persistent renal ischemia-reperfusion injury (mouse)	0.02	0.03	LCN2
Cardiac hypertrophy	0.04	0.01	GPX3, IL11
Oxidative stress	0.04	0.01	GPX3

Validation of Up/Downregulated Molecules in RP Papillary Tissues of CaOx SFs

To validate the mRNA and protein expression results, we used quantitative RT-PCR (qPCR) and immunohistochemical staining. The mRNA expression of LCN2, IL11, SLPI, PTGS1, GPX3, and MMD in the P group was significantly higher and that of secretoglobin family 1D member 2 (SCGB1D2), SLC12A1, and NALCN was significantly lower than that in the N group (Figure 6, A and B).

Figure 6.

mRNA validation by qPCR shows significant higher expression of LCN2, IL11, SLPI, PTGS1, GPX3, and MMD; whereas lower expression of SCGB1D2, SLC12A1, and NALCN is shown in P group compared with N group. mRNA validation of genes showing either (A) >2.0- or (B) <0.5-fold difference in expression in RP papillary tissue compared with that in normal papillary tissue of CaOx SFs. The expression of each gene investigated was determined using qPCR performed using TaqMan assays. The data are presented as means±SEMs. Control values are the average of the data for the C group. *P<0.05 for comparisons between the N and P groups; ^#P<0.05 compared with the C groups. C8orf4, chromosome 8 open reading frame 4; GDPD1, glycerophosphodiester phosphodiesterase domain containing 1; KCNJ1, potassium channel inwardly rectifying subfamily J member 1; MUC4, mucin 4; NAV2, neuron navigator 2.

Based on the qPCR results, we examined the protein expression of differentially regulated genes by immunohistochemistry. Widespread LCN2, IL11, GPX3, and SCGB1D2 protein expression was noted in urothelial cells, tubular epithelial cells, and the interstitial spaces of renal papillae. PTGS1, MMD, SLC12A1, and NALCN proteins were mainly expressed in the cells of the epithelium, tubules, and interstitial spaces. LCN2, IL11, GPX3, and MMD expression was relatively strong, whereas SCB1D2, SLC12A1, and NALCN expression was weaker in the P group than in the other groups (Figure 7, A and B).

Figure 7.

Protein validation by immunohistochemistry shows strong expression of LCN2, IL11, PTGS1, GPX3, and MMD; whereas weak expression of SCGB1D2, SLC12A1, and NALCN is shown in P group compared with N group. Immunohistochemical distribution of genes showing either (A) >2.0- or (B) <0.5-fold difference in expression in RP papillary tissue compared with that in normal papillary tissue of CaOx SFs. Genes were selected according to mRNA validation results obtained using qPCR. The locations of RPs are indicated as asterisks in each representative microphotograph. Magnification, ×400.

Validation of Proinflammation and Apoptosis-Related Molecules Between RPs and Normal Papillary Tissues

According to network, upstream regulator, and toxicity analyses, we further compared proinflammatory gene expression and apoptosis among C, N, and P groups. Expression of IL1B in the P group and that of nitric oxide synthase 2 and TNF in the N and P groups was markedly higher than those in the C group. The number of cells stained positively for CD68, CD138, neutrophil elastase, and terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL), which represented Mϕs, plasma cells, and neutrophils, respectively, increased in the P group compared with the other groups (Figure 8).

Figure 8.

Validation of proinflammatory and apoptotic assay in RP and normal papillary tissues from CaOx SFs and normal papillary tissues from control patients. (A) mRNA expression. The data are presented as the means±SEMs. Control values are the average values of the data for the C group. ^#P<0.05 compared with the C groups. NOS2, nitric oxide synthase 2. (B) Immunohistochemical staining. Cells stained positive for CD68, CD138, and neutrophil elastase (each arrow) are shown in each representative microphotograph. Magnification, ×400.

Discussion

Several studies have reported both basic and clinical significance of RPs for idiopathic CaOx SFs.^18–21 Recent reports regarding RPs have tended to use either computed- and microscopic-technical or clinical characterization with 24-hour urine samples. Using microcomputed tomography (micro-CT), Williams et al. reviewed the concept of stone growth in RPs, in which CaOx crystals start to adhere to RPs that originated from papillary interstitium apatite through the loss of the papillary epithelium and were then excreted in the renal collecting system.¹⁸ Nuclear magnetic resonance spectroscopy revealed that the apatite in RPs was composed of CaP and that RPs included variable proportions of protein, glycosaminoglycan, lipid, and carbonate.¹⁹ Another study using x-ray microanalysis and electron microscopy indicated that RPs had high zinc levels and were associated with calcifying collagen fibers as well as crystals present in membrane-bound vesicles.^20,21

The clinical association between RPs and risk of CaOx stone formation is yet controversial. The urinary volume, urinary calcium and citrate excretion, and serum osteocalcin levels seem to be related to RPs^22–24; however, reports indicate that ductal tubular plugging, and not RPs, is more relevant for kidney stone risk as predicted by 24-hour urine and CT imaging.^25,26 Although the rate of urinary metabolic abnormality is much lower in Japan than in the United States, the urolithiasis prevalence in both countries is approximately 10%.²⁷ This finding shows that not only metabolic but also molecular-based investigations are essential to characterize kidney stone pathogenesis. Since both morphologic and molecular analyses are required to analyze the relevance of RPs for CaOx SFs in order to determine the pathogenesis of RPs, we performed genomic and immunohistochemical analyses using papillary tip samples from patients with kidney stones in this study.

Here, we examined RPs that were of sufficient size to be visualized by endoscopy. Positive results were obtained for the large RPs in von Kossa but not Pizzolato staining, indicating that they contained CaP and not CaOx. Large amounts of collagen fibers were found in the interstitial space surrounding the basement membrane, both with and without RPs. The RPs showed diffuse OPN expression. Our results are consistent with previous reports suggesting that RPs are composed of CaP and that their origin is correlated with collagen fibers and OPN expression.^21,28 Increased collagen fiber and OPN expression plays a crucial role in RP growth; they also contribute to dramatic changes in the expression of other molecules, such as those involved in inflammation and immunity, oxidative stress, and sodium/potassium transporter and channels.

Comparison with the C group of non-SFs showed that the N and P groups of CaOx SFs had many common up- or downregulated genes. Canonical pathway analysis showed the activation of signaling pathways including cAMP, coagulation, Gα, and calcium in both RPs and normal papillary tissues of CaOx SFs compared with normal papillary tissues of control patients. According to disease and function analyses, cellular hyperpolarization, reproductive development, differentiation, and molecular transport also increased, but nutrition levels, cell-to-cell attachment, and organismal development decreased in the papillae of CaOx SFs compared with those of controls. Additionally, other networks related to the cardiovascular system, immune response, and inflammatory disease were raised as the top associated networks in differences between papillae of CaOx SFs and controls. The diversity of the results was derived from the heterogeneity of sample tissues, which consisted of a large amount of renal tubular and interstitial cells and a small amount of urothelial and immune cells as shown in Figure 8B and Supplemental Figure 1. However, the predicted association between CaOx SFs and non-SFs as described above supports the united RP formation theory^16,29–31 involving disorders of cellular structure, signaling, differentiation, mineral density, tissue inflammation, and vascular formation.

Microarray, network, and validation analyses showed that the RP papillary tissue of CaOx SFs had higher LCN2, IL11, PTGS1, GPX3, and MMD expression and lower SLC12A1 and NALCN expression. Network analysis demonstrated that these genes were directly related to each other via extracellular signal-regulated kinase 1/2,^32,33 Akt/phosphatidylinositol 3-kinase, protein kinase a (Pka), and proinflammatory cytokines. LCN2, which is also known as neutrophil gelatinase-associated LCN, is expressed in tubular cells, Mϕ, and neutrophils and is related to cellular apoptosis and inflammation.^34,35 IL11, a member of the IL6 family, is associated with oxidative stress and compensatory proliferation.³⁶ Some studies reported that both neutrophil gelatinase-associated LCN and IL11 are important biomarkers for AKI.^37–39

PTGS1, also known as cyclooxygenase-1, acts as a vasoconstrictor in the kidney and contributes to development of arterial hypertension.⁴⁰ Stoller et al. have supportively hypothesized the theory that kidney stones and RPs could be caused by renal vascular injury.^29,30 GPX3 is found in the renal tubular cells in basement membranes, which indicates the existence of oxidative stress.⁴¹ Toxicity analysis showed that ARF, ischemia-reperfusion injury, and oxidative stress were associated with gene expression of LCN2, GPX3, IL11, and AQP1, such that not only renal tubular and urothelial cell damage but also some kinds of vascular injury resulted in RP formation. Although the Akt/PI3K pathway is related to suppression of renal cell apoptosis and facilitation of vascularization, MMD positively regulates Akt/PI3K activation in Mϕ.^42,43 Since MMD expression occurs via stimulation of proinflammatory Mϕ,⁴³ MMD upregulation resulted in activation of proinflammatory cytokines and oxidative stress.

The downregulated genes SLC12A1 and NALCN encode a membrane transporter and channel, respectively. SLC12A1 is a sodium-potassium-chloride transporter found in the ascending limb of the loop of Henle and is responsible for Batter syndrome type 1. SLC12A1 deficiency results in renal hypokalemia, alkalosis, hypercalciuria, and nephrocalcinosis, and we previously found that CaOx SFs had single-nucleotide polymorphisms of SLC12A1.^44,45 NALCN is a nonselective sodium leak channel that is related to osmoregulation mainly in neuron cells.⁴⁶ Although the contribution of NALCN to renal papillary tissue has not yet been identified, deficiency of this gene might cause renal cell damage and alteration in intratubular mineral saturation with SLC12A1.

Finally, in addition to the detected inflammatory network and validation of each candidate gene responsible using the RP site, the upstream regulator analysis showed activation of IL1B and TNF; therefore, we validated the major proinflammatory cascade and cell expression. As a recent study mentioned,⁴⁷ these inflammation and cellular disorders might contribute to RP formation (Figure 9).

Figure 9.

Schema of hypothetical theory of RP formation, based on this study’s results. (A) Anatomy of renal parenchyma. Renal papilla including loop of Henle, interstitial space, and urothelial cells was focused on in the following schemas. (B) Normal status of renal papillary tissue between the loop of Henle and renal calyx. Aligned urothelial cells, interstitial cells, and tubular cells can be seen. (C) Accumulation of immune cells (macrophages and plasma cells), collagen fibers, and OPN; cellular apoptosis occurs due to inflammation and oxidative stress due to activated molecules, such as LCN2, IL11, PTGS1, GPX3, and MMD. The disorder of SLC12A1 and NALCN causes apoptosis of interstitial cells. (D) Aggregation and precipitation of apatite crystals with collagen fibers and OPN result in the replacement from apoptotic interstitial and urothelial cells to RP. (E) A CaOx stone eventually develops and is attached to the RP.

There are some limitations to this study. First, the lack of clinical data like those for 24-hour urine samples limits the clinical significance of the gene expression profiling of patients. Second, the heterogeneity and mechanical damage of biopsied papillary tissue may have resulted in inconsistency among analyses, with indication of an inflammatory response and limited infiltration of immune cells as shown in immunohistochemistry. Third, comparison of papillary tissues with or without RPs from the same kidney decreased the detection of other responsible candidate genes not statistically but clinically.

In conclusion, we found compelling evidence that genes related to renal injury, vasoconstriction, oxidative stress, Mϕ, and sodium/potassium transporters and channels contribute to RP development in CaOx SFs via proinflammatory activation through the mitogen-activated protein kinase and Akt/PI3K pathways. To our knowledge, this is the first study reporting the gene expression profile of RP papillary tissue, which would contribute to developing molecular targeted therapies for idiopathic CaOx stones.

Concise Methods

Patients

This study was approved by the Nagoya City University ethics board (No. 929). All participants provided informed consent. We obtained biopsies of renal papillary tissues from 23 idiopathic CaOx stone patients who had undergone percutaneous intrarenal surgery (PCNL) or retrograde intrarenal surgery (RIRS) at our institutions during November 2013 to April 2015. The age range of the stone-forming patients and controls was 20 to 80 years. Patients with active urinary tract infection, metabolic and autoimmune disease, carcinoma, and severe hydronephrosis (grade 3 or 4 according to The Society for Fetal Urology guidelines) were excluded.

Samples were individually collected from two different regions in each patient: renal papillary tissue with RP (P group) and normal papillary tissue without RP (N group). We also collected normal renal papillary tissue as a control (C group) from seven patients who underwent ureteroscopy or nephrectomy either for screening for urothelial tumor and hemorrhage or for adhesion of adrenal tumor without urolithiasis. For the nephrectomy, we cut the papillary mucosa tissue with a scalpel.

All the enrolled patients underwent physical examination, blood and urine tests, and ultrasonography and computed tomography imaging before treatment.

Surgical Procedure

For the PCNL, we accessed the renal collecting system from the lower pole in order to fragmentize and remove the pelvic stone with an 18 Fr mini-PCNL tract (Karl Storz, Tuttlingen, Germany) and lithotripter (Swiss LithoClast; Boston Scientific, Marlborough, MA). For the RIRS, ureteral access sheaths (12/14Fr Flexor; Cook Medical Inc., Bloomington, IN) were inserted for all patients. We used a flexible ureteroscope ([Flex-X2; Karl Storz] or [URF-V; Olympus, Tokyo, Japan]) and a holmium laser lithotripsy system (VersaPulse; Boston Scientific). After the removal of fragments, we obtained renal papillary tissue samples from the upper calyx using either BIGopsy (Cook Medical Inc.) or Piranha (Boston Scientific) biopsy forceps. Each group of samples was preserved in both 4% paraformaldehyde (PFA) for immunohistochemical analysis and RNAlater (Qiagen, Hilden, Germany) for microarray and qPCR analyses. Laser coagulation was performed if there was uncontrollable bleeding in the biopsied region by irrigation. A ureteral catheter was inserted at the end of surgery for all patients who underwent PCNL or RIRS.

Microscopic Analysis of RP Papillae

Slices of 4% PFA-preserved renal sections from the RP group were examined by hematoxylin-eosin, Pizzolato, and von Kossa staining, as described previously.^48,49

EDX analysis was performed to measure the components of inorganic calcification of RP. The paraffin-embedded sections were dewaxed and washed with phosphoric acid buffer. The sections were refixed first with 2.5% glutaraldehyde, and then with 2% osmium tetroxide. Dehydration was performed using a 50%–100% ethanol series. The samples were embedded in epoxy resin, coated with platinum, and then photographed with a scanning electron microscope (S-4800; Hitachi, Tokyo, Japan). The elemental spectra of the RPs in the specimens were determined by performing energy-dispersive x-ray analysis using a Horiba EMAX-5770 system (Horiba, Kyoto, Japan).

The microstructure of the RP and surrounding tissue was examined using TEM. The 4% PFA sections were perfusion-fixed in 0.1 mol/L phosphoric acid buffer (20 ml) and 2.5% glutaraldehyde (20 ml), extracted, washed with phosphoric acid buffer, and fixed with 2% osmium tetroxide for 2 hours. The tissues were dehydrated using a graded ethanol series (50%–100%), embedded in epoxy resin, and polymerized at 60°C for 48 hours. Super slices (99 nm) were double stained with uranium and lead and examined under a JEM-1011 TEM microscope (JEOL, Tokyo, Japan). For immunohistochemical staining for TEM, the tissues were incubated overnight with polyclonal anti-human OPN (O-17) rabbit IgG (IBL Co. Ltd., Gunma, Japan) at 4°C in the same blocking solution. The secondary antibody was goat anti-rabbit IgG gold colloidal particles (10 nm; BBI Solutions, Cardiff, UK). The specimens were stained with 2% uranyl acetate for 5 minutes and modified Sato’s lead solution for 1 minute.^50,51

Microarray Analysis

Total RNA was extracted from the tissues in RNAlater using an RNeasy Micro Kit (Qiagen). cDNA amplified using the Ovation Pico System (Nugen, San Carlos, CA) was subjected to transcriptome analysis using Agilent SurePrint G3 microarrays. Microarray data were analyzed using the GeneSpring 13.1 program (Agilent Technologies, Santa Clara, CA). Greater than twofold changes in gene expression between groups were deemed to be significantly different (P<0.01). All microarray data were deposited in Gene Expression Omnibus (Acc. No: GSE 73680).

Data were analyzed through the use of IPA (Qiagen Redwood City Inc., Redwood City, CA). Functional analysis was used to identify the biologic functions and/or diseases that were most significant for the data set. The right-tailed Fisher exact test was used to calculate the P value determining the probability that each biologic function and/or disease assigned to that data set was assigned due to chance alone. A network is a graphical representation of the molecular relationships between molecules that are supported by at least one reference from the literature, from a textbook, or from canonical information stored in the Ingenuity Knowledge Base.

qPCR

For the qPCR, we used the amplified cDNA that was used for microarray analysis. To assess the gene expression results obtained by microarray analysis, validation experiments were performed using TaqMan Gene Expression Assays (Life Technologies, Grand Island, NY) for each complete mRNA sequence. The primers used are listed in Supplemental Table 6. qPCRs were performed using a TaqMan Fast Universal PCR Master Mix (4352042; Applied Biosystems) with a 7500 Fast RT-PCR System (Applied Biosystems). Each gene’s expression was normalized to that of the internal control, glyceraldehyde-3-phosphate dehydrogenase.

Immunohistochemical Staining

Immunohistochemistry for LCN2, IL11, SLPI, PTGS1, GPX3, MMD, SCGB1D2, SLC12A1, NALCN, CD68, CD138, and neutrophil elastase was performed on 4-μm-thick sliced biopsy sections. The antibodies listed in Supplemental Table 7 were used as primary antibodies. The reactions were analyzed using a Histofine Simple Stain Kit for mouse, rat, or rabbit IgG (Nichirei Biosciences Inc., Tokyo, Japan) according to the manufacturer’s instructions.

Evaluation of Apoptosis

TUNEL assays were performed to detect apoptotic cells by using an in situ cell death detection kit (Roche Applied Science, Indianapolis, IN).

Statistical Analyses

All data have been expressed as mean±SD. The statistical analyses were performed using two-way ANOVA for comparisons among three or more groups, or the Mann–Whitney U test for comparisons between two groups. Categorical data were compared using Fisher exact test. All the statistical analyses were performed using Statistical Analysis System, version 9.1 (SAS Institute Inc., Cary, NC). Values of P<0.05 were considered statistically significant.

Disclosures

None.