Urinary nephrospheres indicate recovery from acute kidney injury in renal allograft recipients – a pilot study

Patients admitted to the Department of Medicine III, Division of Nephrology and Dialysis of the Medical University of Vienna with AKI were enrolled into this study. AKI was defined according to KDIGO guidelines of the International Society of Nephrology [8]. The Ethics Committee (EC) of the Medical University of Vienna approved the conduct of this study (EC number 1043/2016). All patients gave their informed consent to participate in this study. Clinical data were obtained from detailed chart reviews. Urine collection was performed every 24 h at 7:30 AM. The urine samples were taken from an indwelling urinary catheter and frozen for storage at − 80 °C.

Paraffin-embedded urinary sediment cells were cut in 4 μm sections, deparaffinized and treated as described earlier. In brief, following tissue antigen unmasking in citrate buffer, slides were washed in phosphate buffered saline (PBS) and incubated with the primary antibody. Following two washing steps with PBS the secondary antibody was applied and incubated for one hour at room temperature. After a second washing step, the horseradish peroxidase (HRP) substrate/chromogen mixture was applied under microscopic observation and left until a visible color-reaction had developed. Counterstaining was performed using Mayer’s hematoxylin.

Cytopreparations were fixed in acetone for 10 min. The primary antibody (either mouse monoclonal, mAb, or rabbit monospecific polyclonal) was applied and slides were incubated overnight. The next day, the secondary antibody (goat-anti-mouse, Alexa546/goat-anti-rabbit, Alexa488) was applied and after a final PBS washing step, 40 μl of DAPI solution was added onto the cell preparation. Slides were mounted in Vectashield® mounting medium for fluorescence (Vector® Laboratories, Inc. Burlingame, CA) under a coverslip. Antibodies used were AQP1 (made in rabbit) from Merck MilliporeTM, CASR (mAb) from ThermoFisher® Scientific and podocalyxin-like protein 1 (PODXL) (mAb) from the Clinical Institute of Pathology of the Medical University of Vienna.

Cellular agglomerates were isolated from urine. After centrifugation of urine the pallet was suspended in culture medium. This was layered on a Ficoll-Paque™ Plus gradient and centrifuged for 20 min in a HettichTM Rotanta 46 RSC centrifuge. The interface was then removed, washed in culture medium and seeded into a 12-well plate. Every third day the culture medium was replaced by RPMI-1640 containing 10% FCS and renal tubular cell culture hybrid mix (1:1) [9]. The seventh day the cell layer was passaged using trypsin to liberate the adherently growing cells from the tissue culture surface. Individual cell clones were picked under microscopic observation during trypsin treatment when just liberating from the surface. In a second isolation step, nephrospheres were handpicked with a micropipette under invert microscopic observation. The isolates were suspended in culture medium for incubation and further propagation, in order to initiate tissue culture. Alternatively, an aliquot of the resultant cell suspension was applied to the funnel of the cytocentrifuge (Shandon, UK) and centrifuged onto microscopic slides.

One hundred mg of normal kidney tissue obtained from tumor nephrectomy were minced in 1000 μl TriFastTM using the peqlab® mincing device (peqlab® Biotechnologie GmbH, Erlangen, Germany). Furthermore, the urine sediment pallet was resuspended in 50 μl urine which was then lysed in 1000 μl TriFastTM. RNA extraction was performed as described in the company’s manual.

cDNA was amplified using AQP1, CASR, and NPHS2 specific probes from TaqMan (Thermo Fisher Scientific®) and 2x TaqMan Universal Master Mix in a StepOnePlus qPCR machine (Applied Biosystems®). GAPDH was used as a housekeeping gene to normalize the individual expression levels of AQP1 and CASR regarding the cycle threshold (Ct) [10]. The individual sample gene expression level was normalized to human kidneys obtained from tumor nephrectomy.

Unfortunately, we were not able to obtain daily follow-up urine samples from all patients, which represents a limitation of this study. Furthermore, qPCR was performed for the entire urinary sediment specimen and not for nephrospheres only. In addition, three urinary sediment samples of the transplanted group were not investigated with qPCR.

To determine whether urinary sediments of AKI patients with allografted kidneys differed from those patients who were no kidney transplant recipients, samples were collected from 45 kidney allograft recipients. Urine collection was undertaken over time, from initial AKI diagnosis until kidney function was restored, or renal replacement therapy was established. Following centrifugation, sediments were recovered and examined microscopically by staining as well as a variety of immunofluorescence methods. Apoptosis was examined using light microscopy. As a control, sediments of 19 non-allograft AKI control patients were also analyzed.

The results of hematoxylin/eosin staining revealed multilayered, three-dimensional cell agglomerates, comprising between ten to more than 100 cells, in 17 of the allograft recipients (Table 1). These cells contained vacuoles and infrequently underwent mitosis (Fig. 1). These spherical structures, herewith termed nephrospheres, were observed both during the period of serum creatinine decrease characteristic of AKI as well as at the time of renal function regain. Appearance of nephrospheres in urine was independent of the duration after kidney transplantation. These nephrospheres could be observed seven days after onset of AKI and over a period of seven days, depending on the individual patient, during improvement of kidney function (Fig. 2). Based on light microscopic examination, it was possible to identify the cells comprising the nephrospheres as tubular cells, albeit their exact source along the nephron remained to be determined. While tubular cells were found in the urine of both transplanted and non-transplanted AKI patients, nephrospheres were exclusively observed in the transplanted group. All patients who excreted nephrospheres in urine recovered from AKI (n = 17), while only 39.29% (n = 11) of kidney transplant patients who did not exhibit nephrospheres in urine (n = 28) recovered. Thereby presence of nephrospheres in urine showed a sensitivity of 60.71% and a specificity of 100% for recovery from AKI with a positive predictive value (PPV) of 100% and a negative predictive value (NPV) of 60.71%.

In contrast, sediments from control patients contained mostly apoptotic tubular cells displaying features such as nuclear condensation, fragmentation, membrane lysis, and expulsion of the nucleus. Some of the tubular cells were found in groups of four to five cells, but these groups differed markedly from the agglomerates seen in transplant recipients (Fig. 3).

Therefore, it would seem that nephrospheres might distinguish between healing transplanted AKI patients and otherwise not transplanted ones, and it would be useful to be able to expose the structures’ biological source along the nephrons.

To determine the precise origin of the tubule cells incorporated into nephrospheres, immunofluorescence staining was performed using commercially available marker proteins AQP1, indicative for the proximal tubule, and CASR specific for the distal tubule. In addition, staining was also undertaken for PODXL, a protein specific for podocytes forming the kidneys’ filtration apparatus. Furthermore, immunohistology staining was also performed for the endometrial stroma marker CD10, which is a relevant protein in renal embryogenesis, and GATA3 and PAX8, which are kidney-specific transcription factors [11]. Standard immunofluorescence staining controls were applied, including the use of tubular urinary sediment cells as positive controls for AQP1 and CASR.

The results showed that CASR (red) expression was present on almost all nephrospheres but varied in its intensity (Fig. 4). Furthermore, levels of AQP1 (green) expression varied between the spheres and among the cells therein (Fig. 4). Staining for the podocyte marker PODXL was positive in only one patient. As indicated in Table 2, CASR+ nephrospheres additionally displaying an AQP1- phenotype were more frequent than AQP1+ agglomerates, indicating that slightly more cells contained within nephrospheres emerged from the distal tubule. The nephrosphere-phenotype was consistent for each patient. Paraffin-embedded nephrospheres were positive for the highly reliable immunohistochemical marker CD10 as well as for GATA3 and PAX8 (Additional file 1: Figure S1) [7].

Hence, since the majority of nephrospheres were decorated with CASR and less so by AQP1, it would be reasonable to assume that they originated from the distal tubule or that distal tubule cells covered proximal tubule cells situated within the nephrosphere. Moreover, the presence of the markers CD10, GATA3, and PAX8, hinted to an embryogenic potential for these kidney-borne cells, and it would be interesting to determine whether these structures had any proliferative potential and could be propagated in vitro for further studies.

As recent data has shown, in vitro engineered nephrospheres possess the ability to expand and repopulate kidney scaffolds and, therefore, we thought it would be interesting to examine whether patient-derived nephrospheres could grow under in vitro conditions. Hence, isolated nephrospheres were seeded in culture using renal tubular cell culture mix and observed by invert microscopy.

When first put into culture, these freshly isolated cell agglomerates either displayed rod-like or tubular structures (Fig. 3). After two to three days in culture nephrosphere cells started to adhere to the surface of the culture plate. Four to five days in culture thereafter, individual new clones started to form (Fig. 5). In the following days, these clones were observed to be further expanding at varying growth rates, displaying morphological diversity. These freshly formed clones were eventually isolated and transferred into fresh media within new wells, where about 3% kept expanding, and could be passaged for a third and fourth time, without showing senescence or a growth stop even after 27 days in culture.

As a result the morphology of these adherent cells varied from those covering large surface areas to rounded cell bodies extending long pseudopodic structures, also above the basal layer of other cells (Fig. 6 upper panel). Phase contrast observation of bulk cultures of nephrospheres showed proliferative foci (Fig. 6d) among cells covering large areas of which some displayed mitotic features (Fig. 6e) and extending ruffles (Fig. 6f), indicative of cells forming contact with neighboring ones. After 27 days in culture, cells started to show signs of senescence with developing vacuoles and loss of cell adhesion.

Hence, nephrospheres appeared to possess at least some regenerative powers, and it seemed logical to investigate whether they had the ability to form kidney tubules.

To determine if newly formed nephrospheres were still expressing marker proteins specific for kidney tubular cells, freshly formed nephrospheres were in situ stained for AQP1, CASR, DAPI and PODXL after 20 days in bulk culture.

All newly grown agglomerates demonstrated AQP1 (green) staining in various intensities, going from very high (Fig. 6a) to moderate intensities (Fig. 6b). AQP1+ cells could be observed covering the plate and were directly growing along AQP1- cells, with which they were forming cell-to-cell contact (Fig. 6a, b). Moreover AQP1+ cells could be depicted developing long dendritiform, podocytic extrusions (Fig. 6c). These cells started to grow proliferatively (Fig. 6d), exhibiting cell division indicated by mitotic cells, which were found in telophase (Fig. 6e) and cells with podocytic extrusions (Fig. 6f).

Thus it appears that nephrospheres grown in bulk culture over 20 days still expressed marker proteins specific to the proximal kidney tubule and thereby still belonged to functional renal cells. The cells ability to grow, form podocytic protrusions and undergo mitosis indicated that cells from nephrospheres still possess potential to grow and probably differentiate.

In addition it was considered important to express and compare the overall representation of kidney tubular proteins in transplanted to non-transplanted AKI patients.

In order to establish quantitative data about presence of tubular cells and regenerative glomerular cells among urinary sediment cells, qPCR for AQP1, CASR, and NPHS2 were performed from urinary sediment of non-transplant patients (n = 19) and transplanted patients (n = 42). Normal human kidney tissue was used as reference (Additional file 1: Figure S2). Individual patient results of AQP1, CASR, and NPHS2 are presented as relative expression level in comparison to normal human renal tissue. In order to calculate AQP1, CASR, and NPHS2 expression in sediment cells GAPDH expression was chosen as reference.

In summary, the extent of CASR mRNA expression was much lower when compared to the extent of AQP1 expression. This indicated that mRNA of distal tubular cells were less frequently found in the urinary sediment of transplanted and non-transplanted AKI patients than those originating from the proximal tubule. No topographical differences of tubular cell origin in urinary sediments between transplanted and non-transplanted AKI patients could be demonstrated via qPCR.