original article
http://www.kidney-international.org & 2009 International Society of Nephrology
An active renal crystal clearance mechanism in rat and man Benjamin A. Vervaet1, Anja Verhulst1, Simonne E. Dauwe1, Marc E. De Broe1 and Patrick C. D’Haese1 1
Laboratory of Pathophysiology, Department of Medicine, University of Antwerp, Antwerp, Belgium
The kidney has several defense mechanisms to avert nephrocalcinosis by preventing intratubular crystal formation and adherence. Little is known about the fate of luminally adhered crystals. In order to study post–crystal adhesion defense mechanisms we quantified the number and morphology of crystal-containing tubules in rats at various time points following ethylene glycol administration as well as in renal biopsies of patients diagnosed with nephrocalcinosis of different etiology. In rats, nephrocalcinosis was completely cleared by epithelial overgrowth of adherent crystals, which were then translocated to the interstitium and subsequently disintegrated. These processes correlated with a low to moderate infiltration of inflammatory cells. Patients with nephrocalcinosis due either to acute phosphate nephropathy, primary hyperoxaluria, preterm birth, or transplantation also showed epithelial crystal overgrowth independent of the underlying disorder or the nature of the crystals. Our study found a quantitative association between changes in tubular and crystalline morphology and crystal clearance, demonstrating the presence of an important and active nephrocalcinosis-clearing mechanism in both rat and man. Kidney International (2009) 75, 41–51; doi:10.1038/ki.2008.450; published online 10 September 2008 KEYWORDS: nephrocalcinosis; reversibility; epithelial overgrowth; crystal–cell interaction; crystal adhesion; crystal clearance
Correspondence: Patrick C. D’Haese, Laboratory of Pathophysiology, Department of Medicine, University of Antwerp, Universiteitsplein 1, 2610 Wilrijk, Antwerp, Belgium. E-mail:
[email protected] Received 21 December 2007; revised 14 May 2008; accepted 4 June 2008; published online 10 September 2008 Kidney International (2009) 75, 41–51
One of the major functions of the kidney is to remove solute and waste material from the body. During this process, the concentrations of calcium (Ca) and its anionic ligands phosphate (P) and oxalate (Ox) can increase up to the level of supersaturation, increasing the risk of crystal formation. Subsequent adherence of intratubular calcium phosphate (CaP) and/or calcium oxalate (CaOx) crystals to undifferentiated/regenerating tubular epithelial cells defines intratubular nephrocalcinosis.1–3 These crystal-bearing epithelia might compromise proper tubular function as undifferentiated cells cannot exert the multiple functional processes of mature, fully polarized epithelial cells. In addition, intratubular-adhered crystals can engender inflammation and may hinder tubular fluid dynamics.4 However, the kidney has developed several defense mechanisms acting at different levels. At the physiological level, a high concentration of calcium itself is able to reduce antidiuretic hormone-stimulated water permeability of the collecting duct through the calcium sensing receptor, leading to an increased urinary volume and a reduced risk of supersaturation.5,6 At the physicochemical level, the urinary supersaturation capacity can be increased and crystal formation, growth, and aggregation can be delayed or prevented by micro- and macromolecular urinary constituents such as citrate, magnesium, and proteins.7–10 At the level of crystal–cell interactions, normal differentiated tubular epithelia have no affinity for crystals and crystal retention can be prevented by coating crystals with urinary macromolecules.1,2,7,11,12 Altogether, the combined actions of these mechanisms make the kidney intrinsically resistant to intratubular nephrocalcinosis. However, what would happen if these renal defense mechanisms are inadequate or fail and crystals adhere to the tubular epithelium? Will this necessarily result in permanent tubular loss and decreased renal function or does the kidney have an additional survival mechanism? On the one hand, it is obvious from human pathology and experimental studies that extensive nephrocalcinosis, as observed in acute phosphate nephropathy and primary hyperoxaluria, may lead to renal damage and functional deterioration.13 On the other hand, the effect of less-extensive intratubular crystal adhesion on renal function and tubular morphology is less straightforward. In kidney transplant patients, Pinheiro et al.14 reported a 12-year allograft survival rate of 75% in the absence of nephrocalcinosis, whereas in the 41
original article
BA Vervaet et al.: Renal crystal clearance
presence of nephrocalcinosis allograft survival decreased to 48%. Although these data suggest an association between nephrocalcinosis and a decreased allograft survival, it should be noted that the survival of almost half of the allografts with nephrocalcinosis is not influenced by crystal deposition. Also, in several prospective and retrospective studies, in which expreterm neonates with nephrocalcinosis were compared with birth-weight and postnatal (or gestational) age-matched controls without nephrocalcinosis, no clear evidence for an association between neonatal nephrocalcinosis and renal dysfunction in the long term was found.15–18 Importantly, these and other preterm follow-up studies described resolution of nephrocalcinosis (in up to 75% of patients) with age and discontinuation or adaptation of neonate diuretic therapy.15–17,19,20 Even so, in patients with obesity-related jejunoileal bypass-induced enteric hyperoxaluria, where kidneys progress towards renal insufficiency secondary to oxalate nephropathy, restoration of gastrointestinal continuity has a stabilizing or partial reversible effect in the decreasing renal function.21–24 On the basis of the above observations, we hypothesize that progression of nephrocalcinosis not only depends on the extent, rate, and duration of intratubular crystal adhesion, but also is additionally countered by a post-adhesion crystal clearing mechanism, which might avoid, slow down or even reverse the progression towards renal insufficiency. In other words, kidneys in which crystal adhesion exceeds a certain threshold might deteriorate, whereas below this threshold the kidney may (continuously) clear the adhered crystal deposits and secure renal function. To find evidence of such a defense mechanism, the recovery phase following an established ethylene glycol (EG)-induced nephrocalcinosis was investigated with respect to crystal clearance and concurrent
In the control animals, the median of urinary oxalate excretion was 0.77 mg/24 h (range: 0.02–1.06) (Figure 1a). Supplementation of the drinking water with 0.75% EG induced hyperoxaluria, with an oxalate excretion of 2.57 mg/ 24 h (range: 0.29–14.36) during EG administration. Hyperoxaluria persisted, at a lower level, until 2 days after stopping EG administration (1.31 mg/24 h; range: 1.11–4.14). At day 5 after arrest of EG administration oxalate excretion had dropped to control values. During the study, serum creatinine (0.40 mg/100 ml; range: 0.10–2.90) levels of EGadministered animals did not differ significantly from controls (0.35 mg/100 ml; range: 0.20–0.50, P40.05). Urine production and water consumption moderately increased during EG administration and evolved to normal levels when EG was withdrawn (data not shown). Crystalluria
Control animals did not show CaOx crystals in their urinary pellets, however coffin and trapezoidal-shaped triphosphate crystals were present occasionally. The calcium content in pellets of centrifuged urine of controls was low (0.16 mg/24 h;
Crystalluria 2.4
12 10 8 6 a 4 2 0
a
2.0 1.6 1.2 0.8 0.4 a 0
Ctr #
EG
+2
+5
+10 +25
Time (days)
Ctr #
EG
+2
Renal crystal content Number of von Kossa-positive tubules
Urinary pellet calcium (mg/24 h)
a
14 Urinary oxalate (mg/24 h)
RESULTS Urine and serum biochemistry
+5
+10 +25
Time (days)
2000
a
1800 1600
a
1400 1200 1000 800
a
600 400 200
d
0 Ctr
8
Oxaluria 16
changes in tubulointerstitial phenotype. As a clinical correlate, human biopsy specimens of four different clinical settings known to be associated with various degrees of nephrocalcinosis were morphologically evaluated; that is two disorders with extensive crystal deposition—acute phosphate nephropathy (CaP) and primary hyperoxaluria (CaOx)—and two disorders with less prominent nephrocalcinosis—preterm infants (CaOx) and kidney transplant patients (CaOx and CaP).
+2
+5
+10
+25
Time (days)
Figure 1 | Urinary biochemistry and nephrocalcinosis. (a) Urinary oxalate excretion showing the transient effect of EG administration with hyperoxaluria persisting until day 2 after arrest of EG administration. (b) Urinary pellet calcium shows clear crystalluria during EG administration only. (c) Renal crystal content during recovery after arrest of EG administration. Quantification of the amount of crystalcontaining tubules/sites on Von Kossa-stained sections shows clearance of nephrocalcinosis. Data are presented as individual values (diamonds) and median (horizontal bars). Black diamonds represent the animals where phenotypical evaluation (Figure 2) was performed. Asterisk represents combined individual data of animals killed at days 1, 2, 3, and 4 during EG administration. # represents values of control animals of each group killed during and after the EG administration period. N represents values of control animals of each group killed after arrest of EG administration. In the graphs, letters a–d represent Po0.05 versus control (Ctr), þ 2, þ 5, and þ 10, respectively, by Mann–Whitney U-test. 42
Kidney International (2009) 75, 41–51
original article
BA Vervaet et al.: Renal crystal clearance
range: 0.06–0.57) (Figure 1b). However, during EG administration the urinary pellets of EG-administered animals clearly contained CaOx crystals (mainly bipyramidal calcium oxalate dihydrate) and the median of centrifugeable calcium excretion increased significantly to 1.18 mg/24 h (range: 0.12–2.28), indicating crystalluria. Within 48 h after arrest of EG administration CaOx crystals were no longer observed by microscopy and the amount of centrifugeable calcium significantly decreased to below control levels. Renal crystal content
Control rat kidneys showed no von Kossa-positive crystals in the tubules (Figure 1c). At 2 days after withdrawal of EG from the drinking water 364 crystal-containing tubules/sites (range: 2–1326) per section were present, not significantly increasing to 536 (range: 1–1782) by day 5. By day 10 after arrest of EG administration the amount of crystal-containing tubules/sites decreased to 175 (range: 1–490) and further decreased to 1 (range: 0–4) by day 25 (Figure 1c). For reason of clarity quantification per renal region (cortex, outer stripe of outer medulla, inner stripe of outer medulla, and inner medulla) was not presented separately as it did not add additional information to the outcome that nephrocalcinosis was reversible. Morphology of crystal-containing tubules
In accordance with a previous study,1 the low-dose and short-term administration of EG, as used in this study, did not result in overt renal tubular cell necrosis, nor in massive interstitial cellular infiltration. Evaluation of the morphology of the crystal–tubular epithelial interaction and the crystal associated interstitial reaction pattern in cortex, outer stripe of outer medulla, inner stripe of outer medulla, and inner medulla revealed five different phenotypical groups: (1) normal tubules containing intraluminal crystals either lying free in the lumen or adjacent to the epithelium, (2) tubules with crystals adhered to flattened regenerating epithelial cells, either or not positive for proliferating cell nuclear antigen (PCNA), (3) tubules with crystals overgrown by the tubular epithelium, (4) interstitial crystals, and (5) granulomas (Figure 2). The results of the separate renal regions were combined for reason of clarity, resulting in the evaluation of 564, 533, 420, and 10 crystal-containing tubules/interstitial sites (out of three or five animals) at days 2 (n ¼ 3), 5 (n ¼ 3), 10 (n ¼ 3), and 25 (n ¼ 5) after arrest of EG administration, respectively. The mean (± s.d.) relative frequencies of the five phenotypes of the three or five animals at the different time points after arrest of EG administration are presented in Figure 2. Normal crystal-containing tubules and tubules with adherent crystals were present from day 2 after arrest of EG administration on with a relative frequency of 12.5±2.0 and 54.5±5.2%, respectively, and the prevalence of both these phenotypes gradually decreased to zero by day 25 of EG recovery (Figure 2a and b). The phenotypical group of overgrown crystals, defined as crystals covered either by Kidney International (2009) 75, 41–51
dividing flattened cells (Figure 2d) or by more differentiated, polarized epithelia (that is cuboidal cells), presenting basement membranes adjacent to the crystals (Figure 2e and f), were already seen shortly after arrest of EG administration (day 2) and represented 28.3±4.2% of all evaluated crystalcontaining tubules/sites at that time, increasing to 49.9±6.4% at day 5 and slightly decreasing to 41.4±10.2% by day 10. At day 25 these overgrown crystals could no longer be detected. The frequency of sites with interstitial crystals, including both intact and disrupted/scattered crystal deposits (Figure 2h, i and j), almost tripled from 4.8±1.1% at day 2 to 13.4±6.8% at day 5 and showed a further 3.5-fold increase to 46.3±14.2% at day 10. The few remaining crystal deposits present at day 25 after arrest of EG administration either were present at interstitial sites (85.0±33.5%) or clearly resided in granulomatous like structures (15.0±33.5%) (Figure 2k). During the 4-day EG administration period only an occasional overgrowth structure, but no interstitial crystals neither granulomas were observed (data not shown). Overall, no clear intracellular crystals were seen. Also, by visualization of brush border (remnants) on epithelial cells distinct from the crystal adhesion site in crystal-containing tubules, periodic acid Shiff (PAS) staining revealed that EG-induced intratubular nephrocalcinosis particularly affects proximal tubules of cortex and the corticomedullary junction. In addition, a number of distal tubules were affected and some of them also showed the phenotypic changes observed in the proximal tubules. Morphology of crystals
From a morphological point of view adherent or overgrown crystals in general showed rather large, uniform intact crystal lattices (Figure 3a and b) whereas crystal deposits in an interstitial environment either presented smaller crystallites at their periphery, or sometimes even were completely scattered and interspersed by cellular material (Figure 3c and d). The surface of crystals found in the scarce granulomas had a microcrystalline appearance, clearly differing from that of the adhered and overgrown crystals. Inflammation
No massive interstitial cellular infiltration was observed. However, evaluation of OX-1 (CD45, common leukocyte antigen) stained sections revealed a time and tubular phenotype dependent association of leukocytes with crystalcontaining tubules and crystals (Figure 4a). Although no quantitative data could be obtained for tubules with nonadhered and adhered crystals on day 10 after arrest of EG administration due to the low abundance of these particular phenotypes at that time point, the association of OX-1 þ cells with crystal-containing tubules and crystals increased with recovery time (day 2oday 5oday 10) and with progression of the renal crystal clearing process (adheredoovergrownointerstitial; Figure 4a). In contrast, ED-1 þ cells (CD68, monocyte/macrophage) were found only moderately co-localizing with crystal-containing tubules/sites 43
original article
BA Vervaet et al.: Renal crystal clearance
(and crystals) and their association additionally decreased with time (Figure 4b). No neutrophils were found to be associated with crystal-containing tubules/sites or crystals.
General morphological evaluation of the renal tissue revealed no edema and only an exceptional enlargement of the interstitial compartment.
Crystal-containing tubules with normal epithelium 100%
Percentage of scored tubules
90% 80% 70% 60% 50% 40% 30% 20%
a
10% 0% Time (days)
a +2
+5
+10
+25
Tubules with crystals adhering to flattened regenerating cells 100%
Percentage of scored tubules
90% 80% 70% 60% 50%
a
40% 30% 20%
a,b
10% 0% Time (days)
a +2
+5
+10
+25
Tubules with crystals overgrown by the tubular epithelium 100%
Percentage of scored tubules
90% 80% 70% 60%
a
a,b
+5
+10
50% 40% 30% 20% 10%
0% Time (days)
b,c +2
+25
Figure 2 | Continued. 44
Kidney International (2009) 75, 41–51
original article
BA Vervaet et al.: Renal crystal clearance
Tubules with interstitial crystals 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Time (days) +2
Percentage of scored tubules
a,b
a,b
a
+5
+10 +25
Granuloma Percentage of scored tubules
100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Time (days) +2
564
+5
+10 +25
Summary of graphs ‘a’ to ‘k’
Symbol legend
533
Crystal (location)
420
10
Percentage of scored tubules
100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Time (days) +2
a,b,c
Number of crystal-containing tubules scored out of three animals per time point. Tubules with nonadherent crystals Tubules with crystals adhering to flattened regenerating cells Tubules with crystals overgrown by the tubular epithelium Tubules with interstitial crystals Granuloma
+5
Flattened PCNA-positive tubular epithelial cells Overgrowing PCNA-positive tubular epithelial cells Brush border (Re-) differentiated tubular epithelial cell Basement membrane
+10 +25
Figure 2 | Frequency of phenotype. Images of serial sections stained with von Kossa (left) and PAS/PCNA (right), respectively. The amount of (a) normal tubules with crystals and (b) tubules with crystals adherent to regenerating cells decreases during recovery after arrest of EG administration. Concomitantly the amount of tubules with (c) crystals overgrown by the tubular epithelium and (g) the number of sites with interstitial crystals increases. (k) By day 25 only a few crystal-containing granulomas are found. (d) Crystals overgrown by young flattened epithelial cells, which most likely differentiate into (e) the polarized epithelia covering the crystals and presenting basement membrane deposition (e: PAS/PCNA, insert; f: Jones stain) and restoration of brush border (e: PAS/PCNA, insert). (g) Interstitial crystals either not (h) or already (i, j) being disintegrated in the subepithelial environment. (l) Cumulative summary of graphs ‘a’ to ‘k’. Concomitant with the decrease in renal crystals (Figure 1) the frequency of overgrown and interstitial crystals increases, reaching up to almost 90% by day 10. Of the few crystals left on day 25 only a minority are found in granulomas. Data are presented as mean (±s.d.) relative frequencies of the five phenotypes of three animals at time points þ 2, þ 5, and þ 10, and five animals at time point þ 25. In the graphs, letters a–c represent Po0.05 versus time point þ 2, þ 5, and þ 10, respectively, by w2-test. Kidney International (2009) 75, 41–51
45
original article
BA Vervaet et al.: Renal crystal clearance
Tubule with adherent crystals
Interstitial crystals
25 microns
25 microns
Tubule with overgrown crystals 25 microns
25 microns
Granuloma 20 microns
2 microns
1 micron
Figure 3 | Morphology of crystals. Images of serial sections either stained with the Von Kossa method (outer left images in a, b, c, d, e; crystals are black) or prepared for SEM (right hand images). (a) Adhered crystals and (b) overgrown crystals show rather large intact crystal lattices. (c, d) Interstitial crystals either show smaller crystallites at their periphery (c, white arrows) or are completely scattered and interspersed with cellular material (d), suggesting crystal disintegration. (e) The surface of crystals in granulomatous structures has a microcrystalline appearance, as shown in the consecutive magnifications.
Clinical study
As no serial sections were available from the different human pathologies, a precise morphological score, based on serial von Kossa and PAS/PCNA-stained sections, could not be performed. Nevertheless, the single von Kossa/hematoxylin and eosin-stained sections allowed to identify crystals overgrown by a new tubular epithelium. In this way, within the clinical settings of acute phosphate nephropathy, preterm neonates, and transplant patients, the average frequency of clear overgrowth structures (relative to the total number of crystal-containing tubules/sites per biopsy) was assessed for each disorder (Figure 5). Acute phosphate nephropathy, preterm infants and transplant patients showed clear evidence of crystal overgrowth in 24.4±24.3, 11.8±8.9, and 28.4±34.0% of crystal-containing tubules/sites, respectively. In primary hyperoxaluria quantification could not be performed as the extent of nephrocalcinosis and renal damage did not allow proper evaluation of tubular morphology. Nevertheless, several clear overgrowth structures could be identified in these biopsy specimens (Figure 5). DISCUSSION
Intratubular nephrocalcinosis can be defined as the microscopic observation of CaOx or CaP crystals adhered to the luminal membrane of undifferentiated/regenerating tubular epithelial cells.1–3 Intratubular nephrocalcinosis is considered 46
to be the result of failing or overwhelmed defense mechanisms acting at the pre-crystal adhesion level. In the present study, however, the observed changes in frequencies of the different crystal-associated tubular phenotypes during recovery after arrest of EG administration together with the morphological changes of the crystals, and the concomitant decrease/disappearance in the number of renal crystal deposits, allows the quantitative description of a post-crystal adhesion scenario, also to be found in humans. After adhesion of the crystals to the regenerating tubular epithelium or, occasionally, even to denuded basement membranes in cases of severe tubular injury resulting in cell loss, the epithelial cells neighboring the adhesion site stretch and proliferate over the adhered crystals. This reaction is independent of the nature of the tubular segment, as both proximal and distal tubules show overgrowth, and of the nature of the deposits, as CaOx (primary hyperoxaluria), CaP (acute phosphate nephropathy), cystine (cystine stone formers25), and 2,8-dihydroxyadenine crystals (as seen in kidneys of rats with adenine-induced chronic renal failure used to study vascular calcification;26 data not shown) are found to be overgrown by the tubular epithelium. This newly formed monolayer then differentiates into a mature epithelium—with basement membranes adjacent to the crystals and restoration of brush border microvilli on the opposite luminal side (in case of proximal tubules)—thereby excludKidney International (2009) 75, 41–51
original article
BA Vervaet et al.: Renal crystal clearance
n In te rs tit i
al
O ve rg ro w
re d he Ad
100%
No OX-1+ cells in proximity of crystal containing tubule/site
90% 80%
OX-1+ cell(s) in proximity of crystal containing tubule/site
70% 60%
OX-1+ cell(s) in proximity of crystals
50% 40% 30% 20% 10% 0%
2
5 10
data of four tubules data of eight tubules no data due to low abundance of designated phenotypes
No data
No data
Percentage of crystal containing tubules
ep No ith rm el al iu m
OX-1 (leukocytes)
2
5 10
2 5 10
2
5 10
Days after arrest of EG administration
al tit i rs te In
O ve rg ro w n
Ad
he
re
d
ep No ith rm el al iu m
ED-1 (monocytes/macrophages)
100%
No Ed-1+ cells in proximity of crystal containing tubule/site
80%
Ed-1+ cell(s) in proximity of crystal containing tubule/site
70% 60%
Ed-1+ cell(s) in proximity of crystals
50% 40%
data of four tubules data of three tubules data of 14 tubules no data due to low abundance of designated phenotypes
20% 10%
No data
30%
No data
Percentage of scored tubules
90%
0%
2 5 10
2
5 10
2
5 10
2 5 10
Days after arrest of EG administration
Figure 4 | Inflammation. Relative percentage of the different crystal containing tubular phenotypes associated with (a) OX-1 þ (CD45, common leukocyte antigen) and (b) ED-1 þ (CD68, monocyte/macrophage marker) cells at the different time points after arrest of EG administration. Whereas the association of OX-1 þ cells with crystal-containing tubules/sites and crystals increases with recovery time, the association with ED-1 þ cells is moderate and transient during renal crystal clearance.
ing the crystalline deposit from the tubular lumen and restoring epithelial integrity of the tubule. During this overgrowth process the crystals do not seem to undergo morphological changes, but are merely directed towards the interstitium together with the epithelial cells of the original adhesion site. Following this translocation into the interstitium crystal disintegration becomes morphologically apparent. Crystals are found broken up into smaller crystallites with cellular material interspersing and surrounding the crystalline remnants, shielding them from the surrounding tissue. The biochemical mechanism by which the overgrown crystals disintegrate and eventually disappear is currently unclear. It is evident, however, that the adhered crystal deposits, ab initio, are too large to undergo endocytosis and subsequent lysosomal degradation. Therefore, following their overgrowth, an initial extracellular crystal disintegrating process has to take place. Biologically, this could be attained (and controlled) by local acidification by polarized proton secretion and/or release of lysosomal content (protons and proteases), resulting in a low-pH-induced chemical dissolution of the mineral phase and proteolysis of the organic Kidney International (2009) 75, 41–51
crystal matrix.27 The identity of the cells and pathways involved therein has not been provided yet. However, monocytes, macrophages, and multinucleated giant cells have been identified in the vicinity of interstitial crystals in EG-administered rats and glyoxylate-injected mice.28–30 In addition, de Water et al.31 showed in vitro phagocytosis and dissolution of CaOx crystals by macrophages, whereas Papadimitriou and Wee32 reported a selective release of lysosomal enzymes from cell populations containing multinucleated giant cells. In the present rat study no major inflammatory reaction was observed. Nevertheless, leukocytes (OX-1) were found associated with crystal-containing tubules/sites and crystals and their appearance increased with time and with the progressive changes of the tubular phenotype during the extrusion of adhered crystals from the tubular lumen to the interstitium. Monocytes and macrophages (ED-1) were only moderately and transiently found associated with crystal-containing tubules/sites, with only a minority of them actually located adjacent to crystal deposits. Although monocytes/macrophages are functionally capable of clearing crystals, their limited association suggests either a transient or minor involvement in renal crystal clearance. 47
original article
BA Vervaet et al.: Renal crystal clearance
Acute phosphate nephropathy (APN)
Crystal
Transplant patients
Primary hyperoxaluria
Preterm infants
Overgrowing tubular epithelial cells APN
Transplants
Preterm inf.
Prim. hyperoxal.
Number of patients
16
10
8
2
Number of biopsies
16
20
8
3
Number of crystal containing biopsies
16
12
8
3
Number of crystal containing biopsies showing overgrown crystals
16
6
7
3
24.4 ± 24.3
28.4 ± 34.0
11.8 ± 8.9
No data
Mean (± s.d.) percentage of crystals showing overgrowth
Figure 5 | Frequency and examples of clear overgrowth structures in human kidney biopsies with nephrocalcinosis of different etiology. Owing to the extent of nephrocalcinosis and renal damage in the biopsies of primary hyperoxaluria no quantitative morphological data could be obtained.
Also, no morphologically apparent neutrophils could be identified in the vicinity of crystal deposits. Whether, in general, leukocyte infiltration is merely due to a local crystalinduced inflammation or whether they are actively recruited to participate in crystal clearance cannot be deduced from the current data, nor can the involvement of lymphocytes, other interstitial cells or the epithelial cells covering the crystals be excluded. The importance of the proposed crystal handling mechanism is corroborated by the following observations. First, the striking decrease in the number of renal crystals in the rat study parallels the relative increase of definite morphological changes of crystal-containing tubules; that is crystal overgrowth and apparent translocation jointly reach up to 87% of crystal deposits by day 10 after arrest of EG administration. Second, alternative clearing mechanisms quantitatively seem to be less important. For example, the clearance of nephrocalcinosis might be explained by detachment of crystals from the regenerating/undifferentiated tubular epithelium, once it has regained its normal noncrystal-binding phenotype. Such a crystal clearance mechanism, however, is unlikely to be of great importance as the number of crystal-containing tubules between days 2 and 5 after arrest of EG administration does not change significantly, whereas the relative frequency of regenerating crystal-bearing epithelia decreases and that of overgrown and interstitial crystals increases. In addition, hyperoxaluria is no longer present at day 5, making it improbable that the loss of crystals by crystal detachment is compensated by the 48
adhesion of newly formed passing crystals. Third, only a minority of crystal deposits (o0.5% of the original amount at day 2 after arrest of EG administration) is found to end up in granulomas, suggesting that the vast majority of crystals must already have been cleared through a faster alternative mechanism. Moreover, as observed with scanning electron microscopy, the different ( ¼ microcrystalline) morphology of granuloma-embedded crystals suggests that handling of crystals by granulomas might be of a different nature. Fourth, as the low-dose and short-term administration of EG, in accordance with a previous study,1 did not result in acute renal failure, nor in overt epithelial necrosis nor massive interstitial cellular infiltration, the vast majority of the large amount of interstitial crystals is unlikely to be the result of dystrophic calcification. In humans, tubular epithelial overgrowth was observed, to various extents, in all four pathologies under study (acute phosphate nephropathy, primary hyperoxaluria, transplants patients, and preterm infants), corroborating the idea that the human kidney is able to actively counter progression of nephrocalcinosis and suggesting that this type of renal crystal handling is a common reaction of the tubular epithelium towards adhered crystals, unrelated to the underlying disease and composition of the crystals. This does not exclude the possibility, however, that the underlying disorder or a particular treatment could influence the kidney’s potency to clear crystals. This phenomenon might be occurring in transplant patients, whose allografts continuously are being stressed by immunosuppressants and for whom it is known Kidney International (2009) 75, 41–51
BA Vervaet et al.: Renal crystal clearance
that the prevalence of nephrocalcinosis increases with time after transplantation.33,34 Therefore, it is possible that the individual outcome of the balance between crystal adhesion and crystal clearance not only depends on the extent, duration and rate of crystal adhesion, but also on a ‘renal (epithelial) health threshold’ above which initiation and accomplishment of the crystal clearing process is ensured. In addition, two considerations have to be made. Whereas in rats the recovery phase was investigated at definite time points in the absence of a continuous hyperoxaluric/crystalluric challenge, human samples were studied in a crosssectional manner without information on the extent, rate or duration of crystal formation/deposition at the moment of biopsy. Second, it is important to carefully consider whether interstitial crystal deposits, as observed in intestinal bypass surgery23 for example, are deposited de novo or are the result of translocation after epithelial overgrowth. In view of this it is important to mention that Evan et al. recently described intraluminal and interstitial crystal deposits (Randall’s plaques) in stone forming patients due to primary hyperparathyroidism.35 Whether this and our observations can be linked mechanistically or merely are independent parallel processes remains to be determined. In conclusion, our quantitative temporal morphological evaluation of the fate of intratubular adhered crystals indicates the presence of an active crystal clearing mechanism in rat and man, involving epithelial crystal overgrowth, and interstitial disintegration. This mechanism seems to act independently of the underlying disorder or the nature of the crystals. Although epithelial overgrowth of crystals has been observed earlier,29,36 the present paper, for the first time, describes a clear quantified association in time between changes in tubular and crystal morphology on the one hand and disappearance of crystals on the other. In addition, evidence is presented corroborating the quantitative importance of this renal defense mechanism in rat and man. MATERIALS AND METHODS Experimental design A total of 56 male Wistar rats (250–300 g; Iffa Credo, Brussels, Belgium) were divided into eight groups (n ¼ 7 each). Within each group one (control) animal received normal drinking water, whereas six received drinking water supplemented with EG (0.75% vol/vol) during 4 days. For determination of oxaluria, crystalluria, and serum creatinine during EG administration four groups were killed, respectively, at days 1, 2, 3, and 4 during the 4-day EG administration period. After arrest of the 4-day EG administration period the remaining four groups received normal drinking water and were killed, respectively, at days 2, 5, 10, and 25 of recovery for determination of oxaluria, crystalluria, serum creatinine, and evaluation of nephrocalcinosis and renal morphology. All animals had free access to standard chow and drinking solution. At 24 h before killing, animals were housed individually in metabolic cages to collect 24-h urine samples and to monitor fluid intake. After sedation (Nembutal) of the rats, a blood sample was taken and kidneys were excised and decapsulated. Sagittal slices of renal tissue were immediately fixed in either methacarn (60% methanol, 30% Kidney International (2009) 75, 41–51
original article
chloroform, 10% acetic acid) or Dubosq-Brasil fixative (47% ethanol, 11.7% H2O, 23.5% formaldehyde, 17.6% acetic acid, and 4 mM picric acid) for 4 h, rinsed with 70% ethanol and embedded in low –melting point paraffin (52 1C; Kendall, Mansfield, MA, USA). Urine and serum samples were frozen at 20 1C until biochemical analysis. The experiments were approved by the Ethical Committee for Animal Experiments of the University of Antwerp. Urine and serum biochemistry For the determination of the urinary oxalate 5 ml urine samples were acidified with 100 ml hydrochloric acid (1 M) after which oxalate was determined by a quantitative enzymatic colorimetric assay (Sigma Diagnostics, Deisenhofen, Germany). The concentrations of creatinine in serum were analyzed on a routine autoanalyzer system (Vitros 750 XRC). Crystalluria For the determination of the amount of urinary crystals 3 ml of urine was centrifuged (10 min, 2500 r.p.m.). The urinary pellet was resuspended in 500 ml tris saline buffer and acidified with 100 ml of 12% HCl, thereby dissolving precipitated crystals. The calcium content in this solution was determined by flame atomic absorption spectrometry (PerkinElmer, Norwalk, CT, USA) and is considered a measure of crystalluria. In addition, crystalluria was evaluated by visualization of urinary sediments of a 1 ml centrifuged urine sample (10 min, 2500 r.p.m.) by optical light microscopy ( 100). Renal crystal content Renal calcium deposits were visualized by von Kossa staining. Deparaffinized Dubosq-Brasil-fixed 4-mm tissue sections were incubated in 5% silver nitrate for 45 min. Slides were rinsed in water incubated in 1% pyrogallic acid for 3 min, rinsed with water, fixed in 5% sodium thiosulfate for 1 min, and counterstained with hematoxylin and eosin. In each sagittal kidney section (n ¼ 1 per rat), the renal crystal content was quantified by counting the number of tubules/interstitial sites presenting black-stained crystals in cortex, outer stripe of outer medulla, inner stripe of outer medulla, and inner medulla. Morphology of crystal-containing tubules Two serial methacarn-fixed tissue sections (4 mm) of representative animals (with a sufficient amount of renal crystals) of each group (n ¼ 3 for days þ 2, þ 5, and þ 10; n ¼ 5 for day þ 25) were stained with von Kossa and PAS/PCNA, respectively. Briefly, in PAS/ PCNA staining, tissue sections were blocked with normal horse serum and incubated with primary label (mouse anti-human PCNA; Dako, Carpinteria, CA, USA). Sections were subsequently incubated with secondary label, biotinylated horse anti-mouse antibody (Vector, Burlingame, CA, USA). Finally, avidin–biotin peroxidase complex (Vector) and diaminobenzidine were used to detect PCNA. No staining was observed when primary labels were omitted. Subsequently sections were PAS-stained for morphological evaluation. Nuclei were counterstained with methyl green. The morphology of 100 (cortex), 50 (outer stripe of outer medulla), 25 (inner stripe of outer medulla), and 25 (inner medulla) crystal-containing tubules/sites per animal identified on the von Kossa-stained section, was evaluated on the serial PAS/PCNAstained section. In this way the morphology of crystal-containing tubules could be evaluated and quantified specifically, whereas the inevitable loss of crystals during PAS/PCNA staining was surpassed. 49
original article
BA Vervaet et al.: Renal crystal clearance
Results are represented as the mean (±s.d.) frequency of a certain phenotype relative to the total number of crystal-containing tubules/ sites per time point.
administration was assessed by the w2-test in combination with Bonferroni correction when more than two groups were compared. Statistics were performed with SPSS 14.0 for Windows software.
Morphology of crystals Two serial methacarn-fixed tissue sections (4 mm) of animals (n ¼ 1 each) killed at day 3 during and at days 10 and 25 after the 4-day EG administration period were stained with von Kossa and prepared for scanning electron microscopy, respectively. Several crystal-containing tubules/sites with different morphology were selected on the von Kossa-stained sections and crystal morphology was evaluated on the serial section by SEM.
DISCLOSURE
Inflammation Two series of three serial methacarn-fixed sections (4 mm) of animals (n ¼ 1 each) killed at days 2, 5, and 10 after arrest of EG administration were stained with von Kossa, PAS/PCNA, and either OX-1 (CD45, leukocyte common antigen) or ED-1 (CD68, monocyte/macrophage marker), respectively. Briefly, for OX-1 or ED-1 staining, tissue sections were blocked with normal horse serum and incubated with primary label (mouse anti-rat leukocyte common antigen (OX-1) or mouse anti-rat ED-1 (Pharmingen, San Diego, CA, USA)). Sections were subsequently incubated with secondary label, biotinylated horse anti-mouse antibody (Vector). Finally, avidin–biotin peroxidase complex (Vector) and diaminobenzidine or aminoethylcarbazole were used to detect OX-1 or ED-1, respectively. No staining was observed when primary labels were omitted. Nuclei were counterstained with methyl green. A total of 20 crystal-containing tubules/sites of a particular phenotype, as identified on the von Kossa- and PAS/PCNA-stained sections, were scored according to their association with interstitial OX-1 and ED-1-positive cells. Three types of association were included: (1) no OX-1 þ and/or ED-1 þ cells adjacent to crystal-containing tubule/site, (2) OX-1 þ and/or ED-1 þ cells adjacent to crystalcontaining tubule, and (3) OX-1 þ and/or ED-1 þ cells adjacent to the crystals. Results are presented as the frequency of a particular type of association relative to the total number of scored crystal-containing tubules/sites of a particular phenotype per time point.
All the authors declared no competing interests. ACKNOWLEDGMENTS
We express our gratitude to Dr Markowitz (Columbia College, New York), Dr Van Woerden (Erasmus University, Rotterdam), Dr Mengel, and Dr Gwinner (Hannover Medical School, Germany) for the generous gift of the human kidney biopsy samples of patients with acute phosphate nephropathy, primary hyperoxaluria, transplanted kidneys, and preterm infants. We thank Walter Dorinne´ and Steven Verberckmoes for the assistance with the SEM imaging and Ludwig Lamberts for general technical assistance, as well as the biochemistry laboratory of the Antwerp University Hospital. REFERENCES 1.
2.
3.
4.
5.
6.
7.
8.
9.
Clinical study Tissue sections of formalin fixated renal biopsies from patients with acute phosphate nephropathy (n ¼ 16 patients, 1 biopsy per patient, 1 section available per biopsy), primary hyperoxaluria (n ¼ 2, 3 biopsies, 1 section available per biopsy), transplants patients (n ¼ 10, 2 biopsies each, 1 section available per biopsy), and preterm infants (n ¼ 8, 1 biopsy each, 1 section available per biopsy) were stained with the von Kossa method in order to visualize calcium deposits and quantify/evaluate the number and morphology of crystalcontaining tubules/sites.
10. 11.
12.
13.
14.
Statistics Data are either presented as mean±standard deviation, as individual values and as median ( þ range). For serum creatinine, oxaluria, crystalluria, and the amount of crystal-containing tubules comparisons between the study groups were assessed using a Kruskal–Wallis test, followed by a Mann–Whitney U-test in combination with Bonferroni correction when more than two groups were compared. A value of Po0.05 was considered significant. Comparison of the frequencies of the different phenotypes between the different time points after arrest of EG 50
15.
16.
17. 18.
Asselman M, Verhulst A, De Broe ME et al. Calcium oxalate crystal adherence to hyaluronan-, osteopontin-, and CD44-expressing injured/ regenerating tubular epithelial cells in rat kidneys. J Am Soc Nephrol 2003; 14: 3155–3166. Verhulst A, Asselman M, Persy VP et al. Crystal retention capacity of cells in the human nephron: involvement of CD44 and its ligands hyaluronic acid and osteopontin in the transition of a crystal binding- into a nonadherent epithelium. J Am Soc Nephrol 2003; 14: 107–115. Verhulst A, Asselman M, De Naeyer S et al. Preconditioning of the distal tubular epithelium of the human kidney precedes nephrocalcinosis. Kidney Int 2005; 68: 1643–1647. Umekawa T, Chegini N, Khan SR. Oxalate ions and calcium oxalate crystals stimulate MCP-1 expression by renal epithelial cells. Kidney Int 2002; 61: 105–112. Hebert SC. Extracellular calcium-sensing receptor: implications for calcium and magnesium handling in the kidney. Kidney Int 1996; 50: 2129–2139. Hebert SC, Brown EM, Harris HW. Role of the Ca(2+)-sensing receptor in divalent mineral ion homeostasis. J Exp Biol 1997; 200(Part 2): 295–302. Kumar V, de la Vega LP, Farell G et al. Urinary macromolecular inhibition of crystal adhesion to renal epithelial cells is impaired in male stone formers. Kidney Int 2005; 68: 1784–1792. Mo L, Huang HY, Zhu XH et al. Tamm-Horsfall protein is a critical renal defense factor protecting against calcium oxalate crystal formation. Kidney Int 2004; 66: 1159–1166. Worcester EM. Inhibitors of stone formation. Semin Nephrol 1996; 16: 474–486. De Yoreo JJ, Qiu SR, Hoyer JR. Molecular modulation of calcium oxalate crystallization. Am J Physiol Renal Physiol 2006; 291: F1123–F1131. Verkoelen CF, van der Boom BG, Houtsmuller AB et al. Increased calcium oxalate monohydrate crystal binding to injured renal tubular epithelial cells in culture. Am J Physiol 1998; 274(5 Part 2): F958–F965. Kumar V, Farell G, Lieske JC. Whole urinary proteins coat calcium oxalate monohydrate crystals to greatly decrease their adhesion to renal cells. J Urol 2003; 170: 221–225. Markowitz GS, Stokes MB, Radhakrishnan J et al. Acute phosphate nephropathy following oral sodium phosphate bowel purgative: an underrecognized cause of chronic renal failure. J Am Soc Nephrol 2005; 16: 3389–3396. Pinheiro HS, Camara NO, Osaki KS et al. Early presence of calcium oxalate deposition in kidney graft biopsies is associated with poor long-term graft survival. Am J Transplant 2005; 5: 323–329. Saarela T, Lanning P, Koivisto M. Prematurity-associated nephrocalcinosis and kidney function in early childhood. Pediatr Nephrol 1999; 13: 886–890. Kist-Van Holthe JE, van Zwieten PH, Schell-Feith EA et al. Is nephrocalcinosis in preterm neonates harmful for long-term blood pressure and renal function? Pediatrics 2007; 119: 468–475. Porter E, McKie A, Beattie TJ et al. Neonatal nephrocalcinosis: long term follow up. Arch Dis Child Fetal Neonatal Ed 2006; 91: F333–F336. Abitbol CL, Bauer CR, Montane B et al. Long-term follow-up of extremely low birth weight infants with neonatal renal failure. Pediatr Nephrol 2003; 18: 887–893. Kidney International (2009) 75, 41–51
original article
BA Vervaet et al.: Renal crystal clearance
19.
20. 21.
22.
23. 24.
25.
26.
27.
Saarela T, Vaarala A, Lanning P et al. Incidence, ultrasonic patterns and resolution of nephrocalcinosis in very low birthweight infants. Acta Paediatr 1999; 88: 655–660. Hoppe B, Duran I, Martin A et al. Nephrocalcinosis in preterm infants: a single center experience. Pediatr Nephrol 2002; 17: 264–268. Dhar NB, Grundfest S, Jones JS et al. Jejunoileal bypass reversal: effect on renal function, metabolic parameters and stone formation. J Urol 2005; 174: 1844–1846. Hassan I, Juncos LA, Milliner DS et al. Chronic renal failure secondary to oxalate nephropathy: a preventable complication after jejunoileal bypass. Mayo Clin Proc 2001; 76: 758–760. Mole DR, Tomson CR, Mortensen N et al. Renal complications of jejunoileal bypass for obesity. QJM 2001; 94: 69–77. Ehlers SM, Posalaky Z, Strate RG et al. Acute reversible renal failure following jejunoileal bypass for morbid obesity: a clinical and pathological (EM) study of a case. Surgery 1977; 82: 629–634. Evan AP, Coe FL, Lingeman JE et al. Renal crystal deposits and histopathology in patients with cystine stones. Kidney Int 2006; 69: 2227–2235. Persy V, Postnov A, Neven E et al. High-resolution X-ray microtomography is a sensitive method to detect vascular calcification in living rats with chronic renal failure. Arterioscler Thromb Vasc Biol 2006; 26: 2110–2116. Fleming DE, Van Riessen A, Chauvet MC et al. Intracrystalline proteins and urolithiasis: a synchrotron X-ray diffraction study of calcium oxalate monohydrate. J Bone Miner Res 2003; 18: 1282–1291.
Kidney International (2009) 75, 41–51
28.
29.
30.
31.
32.
33. 34.
35.
36.
de Water R, Noordermeer C, Houtsmuller AB et al. Role of macrophages in nephrolithiasis in rats: an analysis of the renal interstitium. Am J Kidney Dis 2000; 36: 615–625. de Water R, Noordermeer C, van der Kwast TH et al. Calcium oxalate nephrolithiasis: effect of renal crystal deposition on the cellular composition of the renal interstitium. Am J Kidney Dis 1999; 33: 761–771. Okada A, Nomura S, Higashibata Y et al. Successful formation of calcium oxalate crystal deposition in mouse kidney by intraabdominal glyoxylate injection. Urol Res 2007; 35: 89–99. de Water R, Leenen PJ, Noordermeer C et al. Cytokine production induced by binding and processing of calcium oxalate crystals in cultured macrophages. Am J Kidney Dis 2001; 38: 331–338. Papadimitriou JM, Wee SH. Selective release of lysosomal enzymes from cell populations containing multinucleate giant cells. J Pathol 1976; 120: 193–199. Nankivell BJ, Borrows RJ, Fung CL et al. The natural history of chronic allograft nephropathy. N Engl J Med 2003; 349: 2326–2333. Gwinner W, Suppa S, Mengel M et al. Early calcification of renal allografts detected by protocol biopsies: causes and clinical implications. Am J Transplant 2005; 5: 1934–1941. Evan AE, Lingeman JE, Coe FL et al. Histopathology and surgical anatomy of patients with primary hyperparathyroidism and calcium phosphate stones. Kidney Int 2008; 74: 223–229. Lieske JC, Spargo BH, Toback FG. Endocytosis of calcium oxalate crystals and proliferation of renal tubular epithelial cells in a patient with type 1 primary hyperoxaluria. J Urol 1992; 148: 1517–1519.
51