Three-dimensional Architecture Of Nephrons in The Normal And Cystic Kidney
Mar 17, 2022
for more information:ali.ma@wecistanche.com
Thomas Blanc1,2,7, Nicolas Goudin3,7, Mohamad Zaidan1,4,7, Meriem Garfa Traore3, Frank Bienaime1,5, Lisa Turinsky1, Serge Garbay1, Cle´ment Nguyen1, Martine Burtin1, Ge´rard Friedlander1,6, Fabiola Terzi1,8and Marco Pontoglio1,8
1Institut National de la Santé et de la Recherche Médicale U1151, Centre National de la Recherche Scientifique UMR8253, Université de Paris, Institut Necker Enfants Malades, Département « Croissance et Signalisation », Paris, France; 2Service de Chirurgie Viscérale et Urologie Pédiatrique, AP-HP, Hôpital Necker Enfants Malades, Paris, France; 3Structure Fédérative de Recherche Necker, US24-UMS3633, Paris, France; 4Service de Néphrologie-Transplantation, AP-HP, Hôpital Bicêtre, Le Kremlin-Bicêtre, France; 5Service d’Explorations Fonctionnelles, AP-HP, Hôpital Necker Enfants Malades, Paris, France; and 6Service d’Explorations Fonctionnelles, AP-HP, Hôpital Européen Georges Pompidou, Paris, France
KEYWORDS: cystic kidney disease; kidney; nephron; nephronophthisis; tissue clearing
Translational Statement
Clearing methods provide a unique tool for understanding how morphological changes lead to pathological consequences. Kidneys are critical organs that maintain body homeostasis through water, metabolite, and electrolyte handling, which are crucially dependent on the complex 3-dimensional structure of nephrons. When this structural organization is altered, renal pathophysiology ensues. In the present study, we have developed a powerful method based on optical clearing, multiphoton microscopy, and digital tracing to study the kidney at the single-nephron level under physiological and pathological conditions. In particular, we provide the first 3-dimensional reconstruction of a polycystic kidney at the scale of single nephrons. This method can be applied to several pathological contexts, allowing us to better understand the complex process of kidney deterioration and, consequently, to develop more targeted therapeutic strategies
The kidney maintains body homeostasis via its complex 3-dimensional (3D) nephron structure. When this structural organization is altered, renal pathophysiology ensues. Chronic kidney diseases are characterized by the development of renal lesions. Intriguingly, pathologists have reported a wide heterogeneity in the distribution of lesions during chronic kidney diseases.1,2 Whether this reflects the fact that specific nephron segments can be differently damaged, or alternatively, that some nephrons have different susceptibilities to be injured as a whole, is unknown. To address this issue, it is mandatory to reconstruct the 3D (Three-dimensional) shape of nephrons in pathological contexts.
To our knowledge, nephrons have only been fully reconstructed using serial 2D sections.3–5 Although standard histological sectioning provides high resolution, 3D reconstructions are laborious and hard to obtain. This is primarily due to mechanical distortions, which is an inescapable effect of the slicing procedure. Furthermore, 3D( reconstruction from 2D images does not allow direct imaging of 3D structures in whole-mounted tissues.
Multiphoton microscopy has improved our ability to detect morphological changes in thick sections. However, a significant limitation is the shallow accessible depth due to light scattering. By minimizing refractive index differences, clearing agents have dramatically improved the ability to image depths.6,7 Despite the fact that the first clearing agent was introduced a century ago,8 it is only recently that several clearing protocols have been developed, mostly for the brain.9–17 Recent studies have demonstrated their potential for imaging other solid organs, such as the liver, pancreas, or kidney.18–26
Polycystic kidney disease, a genetically heterogeneous disorder involving mutations in several ciliary genes, is the most common hereditary kidney disease.27,28 Polycystic kidney disease is characterized by the development of cysts that lead to the complete destruction of the kidney.28 Microdissection studies suggested that cysts can develop either as an “out pushing” of a nephron segment, as in autosomal dominant polycystic kidney disease; as ecstatic expansion of collecting ducts (CDs), as in autosomal recessive polycystic kidney disease; or exclusively in medullary tubules, as in nephronophthisis (NPHP).27,29 However, how cysts develop in 3D (Three-dimensional) and organize with each other and normal neighboring tubules is still unknown.
Here, we combined optical clearing with multiphoton microscopy to provide new insights into how nephrons are shaped and organized in physiological conditions and how they are modified during disease, such as cystogenesis. Our results demonstrate that there are 3 types of nephrons and that vessels may influence nephron spatial organization. Interestingly, we observed that nephrons tend to lie in specific planes. Unexpectedly, when we applied this technique to jck mice, we observed that cysts developed only in specific nephron segments with fusiform cysts intermingled with normal tubules.
Click to Cistanche side effects for kidney disease
RESULTS
Experimental setup
To study the shape of entire nephrons in relation to their environment, we adopted a technique based on optical clearing and multiphoton microscopy (Supplementary Figure S1A). By comparing different optical clearing protocols, we determined that for the kidney, benzyl alcohol/ benzyl benzoate (BABB) was the most appropriate in terms of efficacity (Supplementary Figure S1B and C and Supplementary Table S1).
3D (Three-dimensional) reconstruction and digital tracing require high-quality images with high resolution and a high signal-to-background ratio. We observed that the optical signal provided by native fluorescence was not uniform over kidney sections. In particular, images from the medulla were poorly contrasted with a low signal-to-background ratio (Supplementary Figure S2). To improve the signal-to-background ratio, we stained sections with periodic acid–Schiff (Supplementary Figure S1D), because we empirically noted that this staining gave rise to high contrast in thin sections during 2-photon fluorescence. However, the signal was not uniform throughout thick periodic acid–Schiff–stained kidney sections (Supplementary Figure S1E). Therefore, we stained kidney sections with fluorochrome-coupled lectins: peanut agglutinin and wheat germ agglutinin, which stain glomeruli and tubules, and Griffonia simplicity I folia agglutinin, which marks blood vessels.30 Remarkably, with the combination of these lectins, the signal-to-background ratio was dramatically increased in whole kidney sections, with a significant improvement in both the XY plane and the z-axis (Supplementary Figure S1E and Supplementary Movie S1).
3D nephron segment visualization
To visualize the shape of whole nephrons, we traced their paths, beginning at the urinary pole of the glomerulus and ending at the CD junction (Supplementary Movie S2). In spite of the fact that the lectins used globally bind all nephron segments, we noticed that when carefully inspected, these lectins were characterized by a specific pattern in the way they stained basal or luminal membranes in the different nephron segments. In other words, we took advantage of the stereotypical pattern of lectin staining to identify the different nephron segments. In this way, we could easily identify 6 entities: proximal tubule (PT), thin limb (TL) of the Henle loop (HL), thick ascending limb (TAL) of HL, distal convoluted tubule (DCT), connecting tubule (CNT), and CD. The transition between PT and TL was characterized by the sudden loss of the typical PT brush border signal and by the decrease in lumen width, which was almost virtually absent in TL (Figure 1a). Conversely, the transition between TL and TAL was characterized by a significant increase in external tubular diameter and relatively constant lumen width (Figure 1b). In all nephrons, the transition of TAL to DCT was systematically found at the vascular pole of the glomerulus. This transition was characterized by a sudden enlargement of external tubular diameter that was predominantly due to a significant increase in lumen diameter (Figure 1c). The transition between DCT and CNT was characterized by a reduction in both tubular diameter and lumen (Figure 1d). Finally, the connection between CNT and cortical CD was characterized by a sudden increase in both tubular diameter and lumen (Figure 1e). The quantification of these morphological transitions was statistically significant (Figure 1f), and its pertinence was verified with staining with specific tubular markers (Supplementary Figures S3–S5 and Supplementary Movie S3–S5).
The shape and size of PTs differ according to their depth
The staining of thick kidney sections allowed us to trace the coordinates of the spatial evolution of entire PT segments. Interestingly, we discovered that their shape was extremely variable and determined by the position of their glomeruli in the cortex. Globally, we identified 3 patterns. The first corresponded to the most superficial nephrons (SNs) (Figure 2a; Supplementary Movie S6), which occupied the most external part of the cortex (the most external 30% close to the renal capsule). Their convoluted parts formed compact structures with only 6 to 7 convolutions around their own glomerulus, occupying very small and tightly packed spaces. The convolution of these SNs ended in long and straight pars recta that descended into the medulla. At the opposite end of this spectrum, we observed a second pattern of nephrons located in the deepest part of the cortex (40% more internal depth), next to the medulla (juxtamedullary nephrons [JNs]) (Figure 2a; Supplementary Figure S6 and Supplementary Movie S6). Their PTs were characterized by an initial short loop that systematically descended toward the papilla and then U turned to ascend toward the renal capsule. In contrast to SNs, JN proximal convoluted tubules were characterized by large coils that evolved around their own glomerulus and occupied large, loosely packed domains in the juxtamedullary cortex. The convoluted part of the PT of JNs had 10 to 15 convolutions, forming large domains. Finally, the third pattern included nephrons located in the middle portion of the cortex (middle nephrons [MNs]) (Figure 2a; Supplementary Movie S6). Interestingly, these tubules had a shape and spatial orientation that represented a transition between SNs and JNs. In fact, they contained 8 to 9 convolutions with coils that were larger than SNs, but smaller than JNs. In addition, MNs had longer pars recta than JNs. Interestingly, a global comparison of the shapes of the PTs showed that SNs and MNs had a homogeneous pattern whereas JNs were extremely heterogeneous (Supplementary Figure S6). In addition, the spatial reconstruction of several nephrons revealed that PTs never intermingle (Supplementary Movie S6), indicating that each nephron occupies its own individual space in the cortex.
Morphometric analyses confirmed the large differences in the size and shape of SNs and JNs, with an intermediate aspect for the MNs. The PT of JNs was more than 2-fold longer than the PT of SNs (Figure 2b); however, straightness was greater in SNs than in JNs (Figure 2c), consistent with the observation that JN tubules were much more tortuous (Figure 2a; Supplementary Figure S6).
Figure 1 | Morphological criteria used to identify nephron segments. (a–e) Morphology of the different segments of the nephron in control mice. Nephrons were traced from the urinary pole of the glomerulus to the collecting duct (proximal tubule [PT]: turquoise; thin limb [TL] of the Henle loop: light gray; thick ascending limb [TAL] of the Henle loop: dark gray; distal convoluted tubule [DCT]: pink; connecting tubule [CNT]: yellow; cortical collecting duct [CCD]: orange). The arrows indicate the diameter of the different nephron segments. (Continued)
Figure 1 | (Continued) (f) Quantification of the outer tubular diameter of the different nephron segments. Data are expressed as mean±SEM. Analysis of variance followed by the Tukey-Kramer test: **P < 0.01, ***P < 0.001.
Figure 2 | Three-dimensional (3D) proximal tubule reconstruction reveals 3 different shapes and organizations. (a) Representative 3D reconstruction images of proximal tubules through the entire renal cortex, divided into 3 levels (white lines) according to the shape of proximal tubules (PTs): superficial (blue), middle (green), and juxtamedullary (red) cortex, corresponding to the outer 30%, middle 30%, and inner 40% of the cortex, respectively. Note the different shapes of the 3 types of nephrons and in particular the length and tortuosity of the juxtamedullary tubules. Bar=150 mm. (b,c) Quantification of the (b) length and (c) straightness of the 3 types of proximal tubules. (d) Quantification of the volume of glomeruli from each cortex region. Data are expressed as mean SEM. Analysis of variance followed by the Tukey-Kramer test: juxtamedullary PT versus superficial PT: ###P < 0.001; juxtamedullary PT versus middle PT: ***P < 0.001; superficial PT versus middle PT: $$$P < 0.001.
Figure 3 | Three-dimensional (3D) nephron reconstruction reveals 3 different types of nephrons. (a) Representative 3D images of a whole nephron (left panels) from glomeruli to collecting duct in the superficial cortex (blue), middle cortex (green), and juxtamedullary region (red) in control mice. Nephron segmentation (right panels) for the 3 different types of nephrons. Bar=150 mm. (b) Quantification of the inner distance between the 2 limbs of the Henle loop for the 3 types of nephrons. (c) Quantification of the nephron length for the (continued)
Figure 3 | (continued) 3 types of nephrons. (d) Quantification of the length of the different segments of the nephron according to the type of nephron. Data are expressed as mean±SEM. Analysis of variance followed by the Tukey-Kramer test: juxtamedullary nephron (JN) versus superficial nephron (SN): ###P < 0.001; JN versus middle nephron (MN): *P < 0.05, **P < 0.01, ***P < 0.001. CNT, connecting tubule; DCT, distal convoluted tubule; PT, proximal tubule; TAL, thick ascending limb of the Henle loop; TL, thin limb of the Henle loop.
Glomerular size varies with depth
We then analyzed the depth and diameter of glomeruli randomly selected from each of the 3 zones. Morphometric analyses revealed that glomeruli were located on average 783 and 355 mm from the renal capsule in JNs and SNs, respectively. As previously reported, we observed that the size of glomeruli varied according to their depth. In particular, the glomerular volume of JNs was 3 times greater than that of SNs and MNs (Figure 2d).
3D nephron reconstruction
We then traced the spatial evolution of entire nephrons from PT to CNT. The global view of their 3D (Three-dimensional) reconstruction confirmed that the nephron shape differed among SNs, MNs, and JNs (Figure 3a; Supplementary Movie S7). This difference was mainly due to the structure of PT. In addition, we observed that HL had a different 3D spatial organization. In particular, TL and TAL were more distantly separated in JNs than in SNs and MNs (Figure 3b). Morphometric analyses revealed that JNs were globally 2-fold longer than SNs (Figure 3c). The increased length was proportionally distributed in all segments, except for TAL and DCT, which tended to have a similar absolute length in all nephron types (Figure 3d). This suggested that nephron elongation is a surprisingly homogenously patterned process.
The arcuate vessels influence juxtamedullary PT convolution
To gain a more comprehensive perspective of the kidney structure, we took advantage of Griffonia simplicifolia agglutinin staining, which allowed us to trace the arcuate vessels and their branches into the cortical radiate vessels (Figure 4a; Supplementary Movie S8). Interestingly, the concomitant 3D (Three-dimensional) reconstruction of nephrons and vessels showed that the arcuate vessels tend to be in a surprising parallel spatial organization with nearby tubules. In particular, we observed that the path followed by specific JNs (constrained JNs) tends to follow the path of the nearby vessel (Figure 4b; Supplementary Movie S9). Remarkably, this bias (observed only in a subset of JN nephrons) accounted for the large heterogeneity in the shapes of PT (Supplementary Figure S6). Morphometric analyses revealed that the “constrained” PTs were significantly longer and more tortuous (reduced straightness) than the “nonconstrained” ones (Figure 4c and d). It is worth noting that the convoluted parts of both SNs and MNs were positioned above the arcuate vessels and were not constrained.
Cistanche-chronic kidney disease
Tubules tend to lie in planes
Interestingly, an inspection of the 3D (Three-dimensional) shapes of nephrons indicated that they tend to lie in planes (Figure 5a; Supplementary Movie S10). To quantify this parameter and to assess the possible statistical bias of this conformation, we measured the general tendency of tubules to bend out of the plane defined by their loops. Our measurements showed that compared to a simulated model of random walking generated with the same set of consecutive measured angles and length between consecutive points (Figure 5b), the tubules tended to evolve with a smaller deviation from a plane (Figure 5c). This indicated that tubules tended to restrict their path along a plane. The observed differences were highly statistically significant (P < 0.001).
3D nephron reconstruction reveals a specific pattern for cyst development
To determine whether our protocol allows for tracing nephrons in disorganized pathological tissues, we characterized the shapes and spatial distribution of cysts in jack mice, a widely used model of NPHP and cystic disease.31–33 3D images showed that in spite of the appearance of prominent cysts, the general 3D (Three-dimensional) course of nephrons did not differ significantly from that of control mice (Supplementary Movie S11). Similarly, morphometric analyses confirmed that the global length of nephrons, glomerular volume, and length of each nephron segment were comparable with those of control mice (Supplementary Figure S7). Of note, because of cyst development, we were unable to differentiate TL from TAL in HL. All the nephrons developed fusiform cysts (Supplementary Movies S11 and S12). One of the most striking features was the observation that fusiform cystic dilations occurred in multiple locations along the same nephron segment (Supplementary Movies S12 and S13). Intriguingly, we observed that cysts never developed in PTs and HL descending limbs (Figure 6a; Supplementary Movie S12). In contrast, they were predominantly detected in HL ascending limbs, DCTs, CNTs, and in the upper part of CDs, in continuity with CNT (Figure 6a; Supplementary Movie S12). In particular, we observed that DCT, as well as CNT segments, had a peculiarly increased probability to develop cysts in their respective more distal parts (Figure 6b). Quantitative morphometric analyses of 3D reconstruction images revealed that the total cyst volume per nephron was particularly high in CNT (Figure 7a). In addition, we observed that cyst enlargements in HL and DCT were greater in JNs than in SNs and MNs (Figure 7a; Supplementary Movie S12). Moreover, cyst incidence was significantly higher in JNs than in the other types of nephrons (Figure 7b). We also observed that the average glomerular inter distance (calculated on the nearest 5 glomeruli) tended to be particularly increased in cystic mice, especially in the more superficial cortex (Supplementary Figure S8).
Figure 4 | Vessels determine the nephron shape. (a) Three-dimensional reconstruction of the vessel architecture from the arcuate vessels to the cortical radiate vessels. Bar ¼ 200 mm. (b) Representative images of “nonconstrained” (upper panel) and “constrained” (lower panel) juxtamedullary proximal tubules. Bar ¼ 100 mm. (c,d) Quantification of the (c) length and (d) straightness of “nonconstrained” and “constrained” juxtamedullary proximal tubules. Data are expressed as mean±SEM. Mann-Whitney test: “constrained” versus “nonconstrained”: *P < 0.05, ***P < 0.001.
3D (Three-dimensional) reconstruction of numerous cysts revealed that they were extremely variable in shape and volume (Figure 6c; Supplementary Movies S11 and S13). In the HL ascending limb, cysts had a fusiform shape with a small diameter. In DCT, cysts had a rather spherical and larger shape and were interspersed among nondilated tubules. Interestingly, the transition between DCT and CNT was systematically characterized by a normal nondilated structure. In contrast, in CNT, we systematically observed a contiguous dilated cystic structure directly in contact with the CD. Interestingly, in CD, this structure progressively shrunk to a normal tubule size. More distally, no further cysts could be detected in the contiguous portion of the CD. Another consistent feature was the peculiar spatial organization of cysts in the HL ascending limb. In fact, although cysts were deeper and closer to the loop in SNs, they were more superficial and closer to DCT in JNs (Figure 6c). Again, cysts occupied an intermediate position in MNs (Figure 6c).
Figure 5 | Nephron tubules tend to evolve as a flat structure. (a) A typical path of a nephron proximal tubule. This path illustrates its tendency to lie in a plane. The planarity of the evolution of the path can be better seen when properly rotated and aligned with the observer point of view (see the right-hand representation of the same tubular segment, which, on the left-hand side, is represented on a different angle). (b) Schematic representation of a tubular path composed of 4 segments (between 5 points defining a given tubular path shown as an example). The degree to which paths tend to go “out of their plane” can be represented by the angle indicated here as “beta.” (Continued)
Figure 5 | (Continued) This angle is measured between the segment p3-p4 with respect to the plane (defined by the immediately preceding points p3, p2, and p1) and represented by a dark gray rectangle in the scheme. Calculation of beta angles was then recursively calculated all along with the set of points in a tubular path. To evaluate the extent of the bias of these beta angles, we simulated a random walking (Monte Carlo simulation [MCs]) by using the same set of consecutive alpha angles of each tubular path (measured between all consecutive segments). This random walking generated a path with an identical set of alpha angles and randomized beta angles. (c) Violin plots of the global distribution of beta angles and the result of MCs in the proximal tubule (PT), thin limb (TL) of the Henle loop, thick ascending limb (TAL) of the Henle loop, distal convoluted tubule (DCT), and connective tubule (CNT). MCs, P < 0.001.
DISCUSSION
Traditional histological methods are limited in their ability to detect pathological changes that affect the global shape of nephrons. Here, we present a powerful approach based on tissue clearing that has the potential to address crucial questions on kidney architecture with unprecedented spatial detail under normal and pathological conditions. Our results confirmed the existence of 3 types of nephrons, which differ in their location, shape, and size, consistent with their functional specificities. Moreover, we demonstrated that nephrons tend to lie in planes and to adapt to vessel spatial organization. Intriguingly, when we applied this technique to a model of cystic kidney disease, we observed that cysts develop in all nephrons, but only in specific segments. Interestingly, we showed that cyst shape varies according to the nephron segment and that along the same nephron, cysts intercalate with normal nondilated tubules. Altogether, these results provide the first 3D (Three-dimensional) characterization of the spatial arrangement of nephrons and vessels and importantly provide the basis for understanding a pathological process, such as cystogenesis.
Kidney optical clearing is challenging because of high levels of autofluorescence and cell density. By comparing different clearing protocols, we identified in BABB a powerful technique for implementing the visualization and reconstruction of structures located in the deepest part of the kidney, that is, the medulla. Moreover, BABB is rapid and scalable and, once cleared, samples can be stored for months before image acquisition. Another main advantage of this technique is its low cost. Clearing agents may result in shrinking or enlargement of structures.9–17 However, because these modifications of kidney size are isotropic, the relative measurements are not expected to be affected or bias their interpretation. One of the most important limitations is the annotation of structures that is highly time-consuming. Nevertheless, there is a growing number of techniques based on deep learning that should rapidly overcome this limitation.
Consistent with the results obtained by classical techniques, our study showed that nephrons differ in their shape and length according to their position. In particular, we observed that JNs have more developed convoluted PTs and larger glomeruli and are 2 times longer than SNs. The increased length is the result of a harmonious process because lengthening affects proportionally all nephron segments. We also observed that JNs have larger HL. Altogether, these data clearly show that the microanatomy of SNs and JNs is significantly different and that the MNs display intermediate features. Interestingly, physiological studies have shown that nephrons are also functionally different.2 The morphogenetic events that underlie these differences are not yet known. It is thus tempting to speculate that the particular structure of JNs may account for the specificity of their functions.
Interestingly, by providing for the first time the 3D (Three-dimensional) reconstruction of nephrons and their surrounding vessels, we discovered a spatial constraint between the arcuate vessels and a subgroup of nephrons. Thus, it is conceivable to think that vessels may dictate the way nephrons lengthen during development.34,35 The inspection of the 3D shapes of nephrons combined with computational simulations also revealed that nephrons never intermingle and that each nephron tends to lie in a plane. The functional significance of this observation remains to be elucidated.
Cyst development during polycystic kidney disease is still an intriguing process. NPHP, a pathological condition characterized by cyst development, is the most common genetic disease–causing end-stage renal disease in children and adolescents. Twenty NPHP genes have been identified so far.36 Among them, NEK8 encodes a member of the never in mitosis A-related kinase family, which plays a role in cilia function and cell cycle progression.37 Nek8 was originally characterized as the gene mutated in jck mice.32 Notably, a mutation in the same protein domain has been shown to lead to NPHP9 in humans.38 2D studies suggested that cyst development dramatically differs among the different forms of polycystic kidney disease.27,29 In particular, in NPHP, cysts seem to be derived exclusively from CD and DCT.31 Although informative, these immunohistochemical studies cannot argue against the possibility that the loss of specific tubular markers39 artificially accounts for this observation. Thus, our 3D (Three-dimensional) study provides the first clear evidence that cyst development is a process that may involve only specific nephron segments. Interestingly, although NIMA (Never In mitosis gene A)–Related Kinase 8 (NEK8) is expressed in the cytoplasm of all nephron segments, its expression in cilia is restricted to DCT and CD. Because only these segments are prone to develop cysts,31 we can postulate that the disruption of NEK8 ciliary function is the crucial event for cyst formation. Intriguingly, we have also observed that fusiform cysts are in continuity with normal nondilated tubules. The fact that a recessive germline mutation leads to a pathological phenotype in only a subset of cells suggests that a second event might trigger cyst development in these cells, as suggested for autosomal dominant polycystic kidney disease.40,41
Figure 6 | Three-dimensional (3D) nephron reconstruction reveals that cysts develop only in specific nephron segments. (a) Representative 3D images with cyst volume rendering of a whole nephron (left panels) from glomeruli to collecting duct in the superficial cortex (blue), middle cortex (green), and juxtamedullary region (red) in jack mice. Nephron segmentation (right panels) for the 3 different types of nephrons. Bar=150 mm. (b) Probability to develop cysts in the different nephron segments in jack mice. (Continued)
Figure 6 | (Continued) The horizontal axis is a normalized length of nephrons corresponding to 50 bins (proximal tubule [PT], turquoise; Henle loop [HL], white; distal convoluted tubule [DCT], pink; connecting tubule [CNT], yellow). (c) Representative 3D cysts reconstruction images with cyst volume rendering of the ascending limb of HL (left panels), DCTs (middle panels), and CNTs and cortical collecting ducts (CCDs) (right panels) in superficial (upper panels), middle (middle panels), and juxtamedullary (lower panels) nephrons in jack mice. Bar= 50 mm.
Figure 7 | Characterization of cyst distribution and volume throughout the nephron. (a) Quantification of the total cyst volume (sum of the cysts per nephron) for each nephron segment (Henle loop [HL], white; distal convoluted tubule [DCT], pink; connecting tubule [CNT], yellow) and (b) length of tubule occupied by the cysts for each type of nephron (superficial: blue; middle: green; juxtamedullary: red) in jack mice. Data are expressed as mean±SEM. Analysis of variance followed by the Tukey-Kramer test: juxtamedullary nephron [JN] versus superficial nephron [SN]: # P < 0.05; JN versus middle nephron [MN]: *P < 0.05, **P < 0.01.
We also observed that cyst enlargement is more pre-dominant in JNs than in SNs, at least considering the cysts that locate between PT and CNT. Consistently, it has been reported that in the context of glomerulosclerosis, glomerular lesions are found more frequently in JNs than in SNs.1,21,42 Whether overwork due to the increased singlenephron glomerular filtration rates and transport/enzymatic activities accounts for the increased susceptibility of JNs to deteriorate is an interesting hypothesis that deserves further investigation.
In summary, we describe a new technique to image kidneys in 3D (Three-dimensional) with sufficient molecular specificity and resolution to directly record the spatial and quantitative distribution of specifically labeled internal structures (nephrons, vessels, and cysts). Given the convenience, speed, and ability for direct quantification, we anticipate that this technique will become a powerful tool for understanding kidney pathophysiology.
METHODS
Animals
All animal procedures were approved by the “Services Vétérinaires de la Préfecture de Police de Paris” and by the ethical committee of Université Paris Descartes. Two-month-old FVB/N mice (n = 20) were used to set up the experimental condition for kidney clearing. The study was then performed in 2-week-old jack male mice (n = 4) and control littermates (n = 3).
Preparation of renal tissues
Before killing, mice were perfused, via intracardiac catheterization, with 25 ml of heparinized saline (1000 IU/l) followed by 75 ml of 4% paraformaldehyde in phosphate-buffered saline (PBS). Experiments were conducted under xylazine (Rompun 2%, Bayer, Leverkusen, Germany, 6 mg/g of body weight) and ketamine (Clorketam 1000, Vetoquinol SA, Lure, France, 120 mg/g of body weight) anesthesia.
For clearing studies, kidneys were fixed in 4% paraformaldehyde for 4 hours and embedded in 4% agarose and 1.5-mm-thick sections surrounding the renal hilum were cut and stored in PBS at 4℃. Kidney sections were first stained and then cleared.
For immunohistochemical studies on thin sections, kidneys were fixed in 4% paraformaldehyde overnight and paraffin-embedded and4-mm sections were cut.
Staining
Periodic acid–Schiff staining. The 1.5-mm kidney sections were incubated in pure or 1:100 PBS-diluted periodic acid–Schiff for5 minutes at room temperature before clearing.
Immunohistochemistry on thin paraffin-embedded sections. Four-micrometer sections of paraffin-embedded kidneys were heated for antigen retrieval and incubated overnight at 44℃ with flfluorochrome-coupled lectins to trace the tubules (rhodamine peanut agglutinin [RL-1072-5] and rhodamine wheat germ agglutinin [RL-1022-5], Vector Laboratories, Burlingame, CA, diluted at1:200) and segment-specific primary antibody to depict distinct tubular segments (biotinylated Lotus tetragonolobus lectin [B-1325], Vector Laboratories, diluted at 1:100; mouse anti-Calbindin D28K[D-4], Santa Cruz, Heidelberg, Germany, diluted at 1:200; goat anti-AQP2 [C-17], Santa Cruz, diluted at 1:200). The next day, the sections were incubated with the secondary antibody for 1 hour at room temperature (Alexa Fluor 488 conjugate [S32354], Invitrogen; anti-goat Alexa Fluor 488 [A-11055], Invitrogen; anti-mouse AlexaFluor 488 [A-21202], Invitrogen, Carlsbad, CA; all diluted at 1:500)before being colored with 40,6-diamidino-2-phenylindole. All images were acquired with a Nikon Eclipse E800 microscope (Champignysur Marne, France) and prepared using Fiji software (version 1.50).Lectin staining on thick sections. The 1.5-mm-thick kidney sections were incubated at 44℃ for 1 month in Texas Red or flfluorescein isothiocyanate coupled lectins: peanut agglutinin (RL-1072-5) and wheat germ agglutinin (RL-1022-5), diluted at 1:100 in 0.1%PBS azide and 0.1% Triton-X. Sections were then washed every day for 2 weeks with PBS before clearing.
Cistanche-kidney function
Optical clearing
For Scale clearing, 1.5-mm kidney sections were incubated in ScaleA2 (4 M urea, 0.1% Triton X-100, 10% glycerol) or ScaleB4 (8 Murea, 0.1% Triton X-100) for 2 weeks, up to 1 year, at 4℃.
For BABB clearing, 1.5-mm kidney sections were dehydrated in sequential rinses of ethanol at room temperature. The samples were then incubated in BABB (Sigma-Aldrich, Saint Louis, MO) in a 1:2ratio for at least 2 days at 4℃.
Two-photon microscopy image acquisition
Tissues were imaged with aqueous gel on an inverted multiphoton microscope (LaVision BioTec) equipped with a Mai Tai HP Titanium-Sapphire Laser (Spectra-Physics, Santa Clara, CA) (Supplementary Figure S1A). The excitation wavelength was 815 nm at 8% of the power. We used a 20Xwater immersion objective (XLUMPLFL20XW, Olympus [Tokyo, Japan]; numerical aperture, 0.95; working distance, 2.0 mm). For fluorescence acquisition, we used a non-descanned detector at 80% with a bandpass filter of 593/40 nm.
The first step consisted of 2D mosaic image acquisition at the middle z stack by using suboptimal acquisition parameters (400 mm X400 mm and 1011X1011 pixels, with 10% overlap and a line averaging at 2) to guide the selection of the most relevant central fields (Supplementary Figure S9A). Once the XY grid and z fields of interest were defined, parameters were adjusted to acquire high-quality images. Such an approach reduced the region of interest to 5X12 fields per z stack. Then, we acquired 850 z stacks/optical slices (1-mm step size) surrounding the renal hilum in the middle part of the kidney section.
Image stitching, processing, and tracing Each stack was tiled according to the method developed by Preibisch et al., 43 which allows the stitching of a large collection of images by using the “Grid/Collection stitching” plugin of Fiji software (HTTP:// Fiji.sc/Fiji). Because we observed an important shift between 2 adjacent images after stitching, we wrote a custom-matching program in Python. This program compensates for the image shift, allowing us to correctly align the images (Supplementary Figure S9B). Stitched-corrected images were analyzed using Imaris version 8.4.2 (Bitplane, Zurich, Switzerland). Tubules and vessels were traced using the filament tracer package manual mode of Imaris and joined by an Imaris Xtension that we developed with MATLAB (https://www.dropbox.com/s/sm9u5em3orjrpmc/Standalone_Join_ Filament_Tool.zip?dl=0) (Supplementary Figure S9C). Glomeruli and cysts were 3D reconstructed using the manual mode of the surface package of Imaris. The 3D (Three-dimensional) glomeruli spatial organization was displayed using the manual mode of the spots package of Imaris.
Morphometry
Tubular segment length, nephron length, and straightness were determined using the filament tracer package of Imaris. According to Imaris software, straightness was quantified as the ratio between the distance between 2 points and the length of the nephron segment path. Thirty-one PTs and 12 whole nephrons were reconstructed and measured in 3 wild-type mice, whereas 37 PTs and 17 whole nephrons were reconstructed and measured in 4 jack mice. Cyst and glomerular volumes were determined using the “surface package” of Imaris. Thirty-one and 34 glomeruli were measured in wild-type and jack mice, respectively. Seventy-four cysts were measured in total.
The “out of the plane” angles of tubular turns (indicated as “beta”) were measured between the plane (defined by 3 consecutive points) and the direction with the next consecutive fourth point in space (Figure 5b). To evaluate the statistical significance of the observed bias, we simulated the distribution of beta angles with a random rotation of beta angles (Monte Carlo simulation) in structures that were composed of the same segments with the same set of alpha angles.
For the calculation of the probability to develop a cystic lesion along with nephron segments, we represented each nephron by a set of points in space. Each point was annotated with regard to the presence of cyst versus normal structure and the type of segment. The probability at each standardized length (set at 50 bins) was calculated as the ratio between the numbers of points with cystic annotation and the total number of points in the bin.
For the calculation of the average glomerular inter distance, we considered for each glomerulus the average glomerular interdistance calculated on the nearest 5 glomeruli.
Data analyses and statistics
Data are expressed as mean±SEM. Differences between the experimental groups were evaluated using analysis of variance followed, when significant (P < 0.05), by the Tukey-Kramer test. When only 2 groups were compared, the Mann-Whitney test was used. Statistical analyses were performed using Graph Prism software (SanDiego, CA, version 9.0.0).
DISCLOSURE
All the authors declared no competing interests.
ACKNOWLEDGMENTS
We thank the Laboratoire Experimentation Animale et Transgenese (LEAT), Histology and Imaging Platforms of Structure Federative de Recherche Necker for technical assistance. We thank Pierre Isnard, Marie-Claire Gubler, and Nicolas Kuperwasser for critical reading of the manuscript. This work was supported by Institut National de la Santé et de la Recherche Médicale, Université Paris Descartes, Assistance Publique – Hôpitaux de Paris, Agence Nationale de la Recherche, Whoami Laboratoire d’Excellence Who am I, Roche Pharma Research and Early Development (Basel, Switzerland), and Institut Roche de Recherche et de Médecine Translationnelle (Paris, France).
AUTHOR CONTRIBUTIONS
TB, NG, and MZ designed and performed the experiments and analyzed the data. FB, LT, MGT, and SG also performed some experiments and analyzed the data. MB and CN performed the mouse experiments. TB and MZ also contributed to writing the manuscript. GF revised the manuscript. FT and MP provided the conceptual framework and designed the study, supervised the project, and wrote the paper.
Cistanche-Three-dimensional architecture of nephrons kidney
SUPPLEMENTARY MATERIAL
Supplementary File (PDF)
Figure S1. Experimental setup, acquisition procedure, and image processing.
Figure S2. Limits of autofluorescence signal in XY and z resolutions.
Figure S3. A two-dimensional image of the transition between PT and TL on 4-mm paraffin-embedded kidney sections.
Figure S4. A two-dimensional image of the transition between TAL and DCT on 4-mm paraffin-embedded kidney sections.
Figure S5. A two-dimensional image of the transition between DCT and CNT on 4-mm paraffin-embedded kidney sections.
Figure S6. Three-dimensional reconstruction reveals 3 different shapes of proximal tubules.
Figure S7. Three-dimensional reconstruction reveals that the length and segmentation of nephrons are not affected by cysts.
Figure S8. The density of glomeruli in control and cystic kidneys.
Figure S9. Acquisition procedure and posttreatment image processing. Table S1. Comparison of clearing methods. Supplementary File (Movies)
Movie S1. Lectin staining markedly improves the resolution and signal-to-background ratio.
Movie S2. Tracing of the path of a tubule.
Movie S3. Three-dimensional reconstruction of a cleared kidney showing the junction between the proximal tubule and the thin limb of the Henle loop.
Movie S4. Three-dimensional reconstruction of a cleared kidney showing the junction between the thick ascending limb of the Henle loop and the distal convoluted tubule.
Movie S5. Three-dimensional reconstruction of a cleared kidney showing the junction between the distal convoluted tubule and the connecting tubule.
Movie S6. The shape and size of proximal tubules differ according to their depth.
Movie S7. Three-dimensional nephron reconstruction in control mice.
Movie S8. Tracing of the path of a vessel from an arcuate vessel to a cortical radiate vessel.
Movie S9. The arcuate vessels model the shape of the juxtamedullary proximal tubules.
Movie S10. Nephrons tend to lie in a plane.
Movie S11. Three-dimensional nephron reconstruction shows a specific pattern for cyst development.
Movie S12. Three-dimensional nephron segmentation reveals that cysts develop in specific nephron segments.
Movie S13. Spatial arrangement and interrelationship of nephrons and vessels in a polycystic kidney.
REFERENCES
1. Rich AR. A hitherto undescribed vulnerability of the juxtamedullary glomeruli in lipoid nephrosis. Bull Johns Hopkins Hosp. 1957;100:173–186.
2. Bankir L, Bouby N, Trinh-Trang-Tan MM. Heterogeneity of nephron anatomy. Kidney Int Suppl. 1987;20:S25–S39.
3. Christensen EI, Grann B, Kristoffersen IB, et al. Three-dimensional reconstruction of the rat nephron. Am J Physiol Physiol. 2014;306:F664–F671.
4. Zhai XY, Birn H, Jensen KB, et al. Digital three-dimensional reconstruction and ultrastructure of the mouse proximal tubule. J Am Soc Nephrol. 2003;14:611–619.
5. Zhai X-Y, Thomsen JS, Birn H, et al. Three-dimensional reconstruction of the mouse nephron. J Am Soc Nephrol. 2006;17:77–88.
6. Puelles VG, Moeller MJ, Bertram JF. We can see clearly now: optical clearing and kidney morphometrics. Curr Opin Nephrol Hypertens. 2017;26:179–186.
7. Seo J, Choe M, Kim S-Y. Clearing and labeling techniques for large-scale biological tissues. Mol Cells. 2016;39:439–446.
8. Spalteholz W. Über das Durchsichtigmachen von menschlichen und tierischen Präparaten und seine theoretischen Bedingungen, nebst Anhang: Über Knochenfärbung. Leipzig, Germany: S. Hierzel; 1914.
9. Hama H, Kurokawa H, Kawano H, et al. Scale: a chemical approach for fluorescence imaging and reconstruction of the transparent mouse brain. Nat Neurosci. 2011;14:1481–1488.
10. Ertürk A, Becker K, Jährling N, et al. Three-dimensional imaging of solvent-cleared organs using 3DISCO. Nat Protoc. 2012;7:1983–1995.
11. Kuwajima T, Sitko AA, Bhansali P, et al. ClearT: a detergent- and solvent-free clearing method for neuronal and non-neuronal tissue. Development. 2013;140:1364–1368.
12. Ke MT, Imai T. Optical clearing of fixed brain samples using SeeDB. Curr Protoc Neurosci. 2014;66:2.22.1–2.22.19.
13. Chung K, Deisseroth K. CLARITY for mapping the nervous system. Nat Methods. 2013;10:508–513.
14. Tomer R, Ye L, Hsueh B, Deisseroth K. Advanced CLARITY for rapid and high-resolution imaging of intact tissues. Nat Protoc. 2014;9:1682–1697.
15. Yang B, Treweek JB, Kulkarni RP, et al. Single-cell phenotyping within transparent intact tissue through whole-body clearing. Cell. 2014;158: 945–958.
16. Susaki EA, Tainaka K, Perrin D, et al. Whole-brain imaging with single-cell resolution using chemical cocktails and computational analysis. Cell. 2014;157:726–739.
17. Alanentalo T, Asayesh A, Morrison H, et al. Tomographic molecular imaging and 3D quantification within adult mouse organs. Nat Methods. 2007;4:31–33.
18. Parra SG, Chia TH, Zinter JP, et al. Multiphoton microscopy of cleared mouse organs. J Biomed Opt. 2010;15:036017.
19. Renier N, Wu Z, Simon DJ, et al. iDISCO: a simple, rapid method to immunolabel large tissue samples for volume imaging. Cell. 2014;159: 896–910.
20. Lee H, Park JH, Seo I, et al. Improved application of the electrophoretic tissue clearing technology, CLARITY, to intact solid organs including brain, pancreas, liver, kidney, lung, and intestine. BMC Dev Biol. 2014;14:1–7.
21. Iversen BM, Amann K, Kvam FI, et al. Increased glomerular capillary pressure and size mediate glomerulosclerosis in SHR juxtamedullary cortex. Am J Physiol Physiol. 1998;274:F365–F373.
22. Angelotti ML, Antonelli G, Conte C, Romagnani P. Imaging the kidney: from light to super-resolution microscopy. Nephrol Dial Transplant. 2021;36:19–28.
23. Puelles VG, van der Wolde JW, Schulze KE, et al. Validation of a three-dimensional method for counting and sizing podocytes in whole glomeruli. J Am Soc Nephrol. 2016;27:3093–3104.
24. Pannabecker TL, Dantzler WH, Layton HE, Layton AT. Role of three-dimensional architecture in the urine concentrating mechanism of the rat renal inner medulla. Am J Physiol Physiol. 2008;295:F1271–F1285.
25. Saritas T, Puelles VG, Su X-T, et al. Optical clearing in the kidney reveals potassium-mediated tubule remodeling. Cell Rep. 2018;25:2668–2675.e3.
26. Schuh CD, Polesel M, Platonova E, et al. Combined structural and functional imaging of the kidney reveals major axial differences in proximal tubule endocytosis. J Am Soc Nephrol. 2018;29:2696–2712.
27. Braun DA, Hildebrandt F. Ciliopathies. Cold Spring Harb Perspect Biol. 2017;9:a028191.
28. Bergmann C, Guay-Woodford LM, Harris PC, et al. Polycystic kidney disease. Nat Rev Dis Prim. 2018;4:50.
29. Wilson PD. Polycystic kidney disease. N Engl J Med. 2004;350:151–164.
30. Laitinen L, Virtanen I, Saxén L. Changes in the glycosylation pattern during embryonic development of mouse kidney as revealed with lectin conjugates. J Histochem Cytochem. 1987;35:55–65.
31. Smith LA, Bukanov NO, Husson H, et al. Development of polycystic kidney disease in juvenile cystic kidney mice: insights into pathogenesis, ciliary abnormalities, and common features with human disease. J Am Soc Nephrol. 2006;17:2821–2831.
32. Liu S, Lu W, Obara T, et al. A defect in a novel Nek-family kinase causes cystic kidney disease in the mouse and in zebrafish. Development. 2002;129:5839–5846.
33. Franke M, Baeßler B, Bechtel J, et al. Magnetic resonance T2 mapping and diffusion-weighted imaging for early detection of cystogenesis and response to therapy in a mouse model of polycystic kidney disease.Kidney Int. 2017;92:1544–1554.
34. Barry DM, McMillan EA, Kunar B, et al. Molecular determinants of nephron vascular specialization in the kidney. Nat Commun. 2019;10: 5705.
35. Perretta-Tejedor N, Jafree DJ, Long DA. Endothelial-epithelial communication in polycystic kidney disease: role of vascular endothelial growth factor signaling. Cell Signal. 2020;72:109624.
36. Srivastava S, Molinari E, Raman S, Sayer JA. Many genes-one diseases? Genetics of nephronophthisis (NPHP) and NPHP-associated disorders. Front Pediatr. 2017;5:287.
37. Fry AM, O’Regan L, Sabir SR, Bayliss R. Cell cycle regulation by the NEK family of protein kinases. J Cell Sci. 2012;125:4423–4433.
38. Otto EA, Trapp ML, Schultheiss UT, et al. NEK8 mutations affect ciliary and centrosomal localization and may cause nephronophthisis. J Am Soc Nephrol. 2008;19:587–592.
39. Wilson PD. Apico-basal polarity in polycystic kidney disease epithelia. Biochim Biophys Acta. 2011;1812:1239–1248.
40. Happé H, Leonhard WN, van der Wal A, et al. Toxic tubular injury in kidneys from Pkd1-deletion mice accelerates cystogenesis accompanied by dysregulated planar cell polarity and canonical Wnt signaling pathways. Hum Mol Genet. 2009;18: 2532–2542.
41. Piontek K, Menezes LF, Garcia-Gonzalez MA, et al. A critical developmental switch defines the kinetics of kidney cyst formation after loss of Pkd1. Nat Med. 2007;13:1490–1495.
42. Ikoma M, Yoshioka T, Ichikawa I, Fogo A. Mechanism of the unique susceptibility of deep cortical glomeruli of maturing kidneys to severe focal glomerular sclerosis. Pediatr Res. 1990;28:270–276.
43. Preibisch S, Saalfeld S, Tomancak P. Globally optimal stitching of tiled 3D (Three-dimensional) microscopic image acquisitions. Bioinformatics. 2009;25: 1463–1465.
