- Email: [email protected]
Flow cytometric analysis of urine lymphocytes isolated from patients with renal transplants—purification of urine lymphocytes
Journal of Immunological Methods 213 Ž1998. 145–155
Flow cytometric analysis of urine lymphocytes isolated from patients with renal transplants—purification of urine lymphocytes Jozef ´ Stachowski a,b,c,) , Claudia Barth a, Maria Lewandowska-Stachowiak b, Petra Lammerding a , Dariusz Runowski b, Conrad A. Baldamus a a Department of Nephrology, UniÕersity of Cologne, Joseph-Stelzmann Str. 9, 50924 Koeln, Germany Department of Clinical Chemistry, UniÕersity of Cologne, Joseph-Stelzmann Str. 9, 50924 Koeln, Germany Department of Nephrology, Clinic of Pediatric Diseases, UniÕersity School of Medical Sciences, Szpitalna Str. 27 r 33, 60372 Poznan, ´ Poland b
c
Received 24 March 1997; revised 6 October 1997; accepted 1 December 1997
Abstract Early diagnosis of rejection is a pivotal problem in renal transplantation. Recent advances in urinary cell analysis using flow cytometry are still burdened with difficulties concerning urine lymphocyte ŽUL. isolation. The analysis of lymphocytes washed out with the urine from the kidney transplant offers a tool to monitor noninvasively the intragraft immune response. However, the demand for optimal isolation of UL with high viability and good separation of other cell types has not, as yet, been met. The present study was undertaken to evaluate the optimal conditions for harvesting UL in order to perform adequate UL analysis by flow cytometry. We found that UL viability is mainly dependent on the time of urine harvesting. Low UL viability was caused by high urine osmolality due to high concentrations of urea and glucose. In contrast, high protein concentrations protected UL viability. Hence, the following algorithm of adequate UL isolation for flow cytometric analysis was established: Ž1. Collection of morning urine directly onto foetal calf serum ŽFCS: 30% vrv.; Ž2. UL isolation within 2 h; Ž3. Erythrocyte lysis with subsequent two-step density gradient isolation of UL from residual erythrocytes, granulocytes ŽFicoll-Isopaque, 1.077 grcm3 . and from uroepithelial cells Ž30% methylglucamine 3,5-diacetomido-2,4,6-triiodobenzoicum, 1.085 grcm3 .; Ž4. Flow cytometric analysis of UL using the ‘live gate’ setting in the area of blood lymphocyte cluster. Adequate UL isolation and special settings of the flow cytometer may provide a useful tool for early diagnosis and the noninvasive monitoring of renal transplant rejection. q 1998 Elsevier Science B.V. All rights reserved. Keywords: Urine lymphocytes; Flow cytometry; Kidney transplant rejection
Abbreviations: APM: aminopeptidase M; ATN: acute tubular necrosis; DTC: distal tubular cells; FCS: foetal calf serum; FIT: fluorescein-isothiocyanate; FSC: forward scatter; g-GT: g-glutamyltransferase; MFI: mean fluorescence intensity; PBS: phosphate buffered saline; PE: phycoerythrin; PI: propidium iodide; PS: phospholipid phosphatidylserine; PTC: proximal tubular cells; RD: receptor density; SSC: side scatter; THG: Tamm–Horsfall glycoprotein; UL: urine lymphocytes; Uropolinum: methylglucamine 3,5-diacetomido-2,4,6-triiodo benzoicum ) Corresponding author. Tel.: q49-221-460-1951; fax.: q49-221-460-1945; e-mail: [email protected]. 0022-1759r98r$19.00 q 1998 Elsevier Science B.V. All rights reserved. PII S 0 0 2 2 - 1 7 5 9 Ž 9 8 . 0 0 0 1 3 - 1
146
J. Stachowski et al.r Journal of Immunological Methods 213 (1998) 145–155
1. Introduction
2. Materials and methods
Early diagnosis of rejection episodes is still a pivotal problem in renal transplantation. Since mononuclear cells washed out with the urine seem to reflect the intragraft immune response, the analysis of activation markers on the cell surface or within the cell offers the possibility of noninvasive monitoring. Flow cytometry is the technique of choice for the analysis of cell surface antigen expression ŽHammer et al., 1993; Lao et al., 1993; Lee et al., 1992; Marcussen et al., 1992; Radio et al., 1993.. Recent developments in detecting target antigens within the cytoplasm and nucleus have made flow cytometry applicable not only to surface but also to intracellular molecules such as cytokines and nuclear antigens ŽJung et al., 1993; Lan et al., 1996; Teague and El-Naggar, 1994.. However, flow cytometry requires a complicated cell preparation process in which cell purification, staining and fixation are crucial. Optimal staining time depends on the cell type under analysis Že.g., blood vs. urine lymphocytes. and on the characteristics of the selected antigen or receptor. In order to perform adequate flow cytometric analysis the following requirements have to be met: high viability and good separation of the lymphocytes from dead cells, debris and other cell types Že.g., uroepithelial cells.. The cytological profile in the urine of renal allograft recipients has so far been determined by light microscopy using direct smear and monoclonal antibody staining ŽDooper et al., 1991; Kyo et al., 1992a,b; Kyo and Mihatsch, 1991; Segasothy et al., 1989; Simpson et al., 1989; Xiao et al., 1992.. Only a few publications have applied flow cytometry to the analysis of urine sediment ŽGorski et al., 1992; ´ Jorge et al., 1991; Lao et al., 1993; Marcussen et al., 1992; Roberti et al., 1995.. None of these studies controlled for lymphocyte viability and purity following urine harvesting. Moreover, optimal urine collecting conditions including osmolality, urea, glucose and protein concentration, which may influence adequate flow cytometric analysis of UL, were not defined. Therefore, the data available to date are not sufficient for reproducible results using flow cytometry as a routine diagnostic tool. This study optimises and defines the conditions of adequate UL isolation for flow cytometric analysis.
2.1. HarÕesting of urine samples Fifty-six renal allograft recipients were followed posttransplant. Morning urine samples Žcatheter or spontaneous midstream. were collected on alternate days for the first 3 weeks, twice weekly for the next 2 weeks, and weekly for the subsequent time course. All collected samples were examined immediately and after 1, 2, 3, 4 and 5 h in order to estimate the influence of time on the viability and cell surface receptor expression of UL. 2.2. Viability of urine lymphocytes dependent on time and on urinary concentration of protein, glucose and urea In order to determine the optimal time point for UL analysis by flow cytometry after urine harvesting, cell viability was examined. The viability of cells recovered from the interphase of density gradient centrifugation was assessed by trypan blue dye exclusion. Dead UL were distinguished from apoptotic UL directly after urine harvesting and before UL isolation Žsee Section 2.4.. The influence of the time interval between urine harvesting and flow cytometric analysis, as well as the influence of urine osmolality and different concentrations of protein, glucose and urea in the urine influencing UL viability, was analysed simultaneously. The pH value of the urine sample was adjusted to 7.0. Correlation coefficients for viability were calculated over time with respect to urine osmolality, urine urea, protein and glucose concentrations using SigmaStat for Windows, MATHCAD software or an in-house program. 2.3. Isolation of urine lymphocytes— separation from uroepithelial cells Collected urine specimens supplemented with FCS Ž30% vrv. were centrifuged for 10 min at 200 = g. The supernatant was decanted, the cells washed with PBS ŽDulbecco without Caq2 , Mgq2 . and resuspended in PBS containing 30% FCS ŽDifco, Detroit, MI, USA.. Subsequent isolation of the lymphocyte population and separation from granulocytes, as well
J. Stachowski et al.r Journal of Immunological Methods 213 (1998) 145–155
as uroepithelial cells Žproximal and distal tubular cells, collecting duct, urothelium., were performed using two-step density gradient centrifugation Ž300 = g, 30 min. on Ficoll-Isopaque ŽLymphoflot-Biotest Diagnostics, Germany; density 1.077 grcm3 and ca. 810 mosMrkg. and on 30% sodium methylglucamine 3,5-diacetomido-2,4,6-triiodobenzoicum ŽUropolinum, Polfa, Poland; density 1.085 grcm3 and ca. 880 mosMrkg.. In samples with massive haematuria, lysis of erythrocytes was performed using 0.84% NH 4 Cl before density gradient centrifugation Ž Ficoll or Uropolinum.. 2.4. Flow cytometry Experiments were performed on the FACScan flow cytometer ŽBecton-Dickinson, Heidelberg, Germany. with FACScan research and Consort 30 software and subsequent additional analysis using DATA-Mate ŽBecton-Dickinson. software. An argon ion laser operating at 488 nm was used to illuminate the cells. The filters used to measure fluorescein ŽFITC. and phycoerythrin ŽPE. fluorescence were 530-nm band-pass and 630-nm long-pass filters, respectively. For data acquisition we used a linear amplification mode for forward-scatter ŽFSC. and side-scatter ŽSSC. parameters and logarithmic amplification for fluorescence 1 ŽFL1-FITC. and fluorescence 2 ŽFL2 s PE.. Fluorescence histograms of UL of at least 5000 counts were generated using a ‘live gate’ set-up on the blood lymphocytes isolated from the same individual using FSC vs. SSC scattergram for the background Žcomparison.. The percentage of positive cells was measured from a cut-off set using an isotype matched nonspecific control antibody, while the mean channel fluorescence ŽMFI s mean fluorescence intensity presented in arbitrary units. was measured over the entire lymphocyte subpopulation. MFI was proportional to the number of receptor binding sites on the cell surface and was determined as the mean receptor density per cell ŽRD.. UL were stained in double-colour fluorescence with the monoclonal antibodies anti-CD45RA labelled with FITC and anti-CD4 conjugated with PE ŽBecton-Dickinson.. Data analysis was carried out using dot plot statistics of either Consort 30, FACScan or Data Mate software ŽBecton-Dickinson..
147
In order to characterise the events in those areas excluded from the ‘live gate’ setting, the whole urine sediment Ž n s 52. was stained with Annexin-V-FITC as a marker for apoptosis and propidium iodide ŽPI. as a marker for cell death ŽPharmingen. ŽVermes et al., 1995.. Annexin-V is a 35–36-kDa Ca2q-dependent phospholipid-binding protein that has a high affinity for the membrane phospholipid phosphatidylserine ŽPS.. In apoptotic cells, PS is translocated from the inner to the external cellular membrane; hence, Annexin-V-FITC bound to PS identifies cells undergoing apoptosis, especially in the earlier stages of apoptosis. PI is a standard flow cytometric viability probe distinguishing viable from nonviable cells ŽVermes et al., 1995.. Tubular epithelial and urothelial cells were detected by monoclonal antibodies URO8 and URO9, respectively, and were estimated in parallel by indirect fluorescence staining ŽSegasothy et al., 1989.. Improved analysis of tubular epithelial cells was performed using mAbs against aminopeptidase M ŽAPM, mAb CD13. and against g-glutamyltransferase Žg-GT, mAb 138H11. as markers of proximal tubular cells ŽPTC. ŽFischer et al., 1990.. The distal tubular cells ŽDTC. were identified using a mAb recognising Tamm–Horsfall glycoprotein ŽTHG; mAb 5A2., a specific antigen for the thick ascending limb and the early distal convoluted tubule ŽBaer et al., 1997.. Flow cytometric localisation of these cell types was performed using PAINT-A-GATE software. These parameters of events excluded from the ‘live gate’ were assessed on a separate set of 52 urine samples from renal transplant patients. Hence, these parameters are not included in this study.
3. Results 3.1. Viability of urine lymphocytes— dependence on the time interÕal after urine harÕesting, urine osmolality and concentration of urea, glucose, and protein in the urine Urine osmolality is mainly determined by the concentrations of urea, glucose and protein. These concentrations may vary widely even in physiological conditions ŽTables 1–3, Fig. 1.. UL viability
J. Stachowski et al.r Journal of Immunological Methods 213 (1998) 145–155
148
Table 1 Viability a of UL: influence of time and osmolality
Table 3 Viability of UL: influence of time and urine glucose concentration
Osmolality ŽmosM.
Time course after urine-sample harvesting 1h 2h 3h 4h 5h
Glucose Žmgrdl.
300–400 400–500 500–600 600–700 700–800 800–900 900–1000
100 100 96 45 50 20 11
- 20 20–50 50–100 100–200 200–400 400–600 600–800 800–1000 1000–2000 2000–4000
90 87 80 72 33 18 10
75 50 38 24 11 5 4
42 38 25 16 12 7 4
33 25 20 12 8 7 5
a
Viability is expressed as a percentage of all cells. Mean values for 1322 analysed urine samples are presented.
Time course after urine-sample harvesting 1h
2h
3h
4h
5h
100 98 94 83 80 71 45 32 20 9
95 90 92 87 80 82 70 33 18 5
60 50 37 24 12 11 6 5 4 3
20 22 21 17 14 15 13 11 6 3
22 22 20 22 11 10 7 7 4 2
a
Viability is expressed as a percentage of all cells. Mean values for 1322 analysed urine samples are presented.
is diminished by urea- and glucose-associated high osmolality after urine harvesting ŽTables 2 and 3.. The urine samples with high glucose concentrations were collected from diabetic patients, including a few diabetics with a poor blood-glucose adjustment, where the urine glucose reached values between 1000–4000 mgrdl ŽTable 3.. However, at the same time these patients showed a urine pH decrease, which may reveal additional negative influence on the UL viability in vitro Ždata not shown.. On the other hand, a high protein concentration in the urine had a highly protective effect on UL viability ŽFig. 1.. The separate influence of each parameter on UL viability does not reflect their mutual influence exerted at the same time under physiologic conditions. Hence, the urine-derived data were pooled and analysed by the six-parameter data matrix test taking into consideration the influence of protein, glucose and
Table 2 Viability of UL: influence of time and urea concentration Urea Žmgrdl.
Time course after urine-sample harvesting 1h
2h
3h
4h
5h
- 500 500–1000 1000–1500 1500–2000 2500–3000 3500–4000
98 96 96 91 60 40
85 90 70 60 56 32
68 50 25 20 18 8
40 30 20 10 8 4
30 20 10 5 2 0
a
Viability is expressed as a percentage of all cells. Mean values for 1322 analysed urine samples are presented.
urea concentrations, osmolality and time after urine collection on UL viability in vitro. This multiparameter regression analysis confirmed our findings of the protective effect of protein on UL viability. Computer simulation showed that the negative effect of a high urea concentration on UL viability Žup to 3590 mgrdl. was diminished by a urine protein concentration of 8100 mgrdl, which extended the viability to 96% for 2 h or to 76% for 3 h after urine harvesting. Therefore, the high correlation coefficient Žrange: 0.90–0.92. between protein concentration ŽG 8100 mgrdl. and UL viability was still present even when the concentration of urea or glucose was higher than 3600 mgrdl or 1000 mgrdl, respectively. Thus, in order to benefit from the protective effect of protein, urine samples were adjusted with FCS Ž10–30 grdl. as exogenous protein ŽFig. 1.. This procedure ensured that in the time course of 2–3 h following urine harvesting, UL viability ranged between 88% to 100% ŽFCS,30 grdl., independent of an unpredictable urine composition. 3.2. Algorithm of urine lymphocyte isolation The experiments concerning the influence of osmolality and different concentrations of protein, urea and glucose in the urine suggest the following schedule of UL collection in order to reach the optimal conditions of UL isolation. Urine samples Ž50–150 ml. have to be collected in the morning and should
J. Stachowski et al.r Journal of Immunological Methods 213 (1998) 145–155
149
Fig. 1. Viability of UL influenced by urine protein concentration and the time after urine harvesting. Urine samples Ž n s 1322. were aliquoted according to the protein concentration and incubated at room temperature for 5 h. At 1-h intervals UL were isolated by density gradient centrifugation and their viability was assessed. Some urine samples Ž n s 120. were adjusted to 10 grdl, 20 grdl or 30 grdl with exogenous FCS immediately after urine harvesting. The bar graph presents the mean values.
be directly supplemented with FCS Ž30% vrv.. The isolation procedure should be performed within a maximum of 2 h. A UL count of 5.5 = 10 3 –10.5 = 10 4rml can be obtained depending on the clinical conditions. UL were found to be mainly present in patients with acute or chonic rejection; the count was extremely low in patients with stable transplant function. The FCS-supplemented urine was centrifuged for 10 min at 200 = g and the sediment was resuspended in PBS q 30% FCS. This diluted urine sediment was centrifuged on Ficoll with subsequent centrifugation on 30% methylglucamine 3,5-diacetomido-2,4,6-triiodobenzoicum ŽUropolinum .. The density gradient of Uropolinum was determined
in a multiple series of experiments designed to optimise the procedure for maximal recovery of UL and minimal contamination by erythrocytes, epithelial cells and granulocytes ŽTable 4.. After this centrifugation procedure the isolated cell population contained ) 92% lymphocytes and 4%–6% monocytic cells as determined by peroxidase and May– Gruenwald–Giemsa staining. The epithelial cells were almost totally separated from the lymphocytes as demonstrated by phase-contrast microscopy ŽFig. 2; Table 4.. When only a small number of epithelial cells was present in urine samples Ž- 30rml., centrifugation on a Ficoll density gradient yielded sufficient cells for submission to flow cytometry analysis.
Table 4 Efficiency of separating urine sediment epithelial cells from lymphocytes using density gradient centrifugation on sodium methylglucamine 3,5-diacetomido-2,4,6-triiodobenzoicumrUropolinumr a Uropolinum
Interphase epithelial cells: lymphocytes Sediment epithelial cells: lymphocytes a
40% Ž1.096grcm3 , 1060mosMrkg.
30% Ž1.085grcm3 , 880mosMrkg.
25% Ž1.076grcm3 , 800mosMrkg.
20% Ž1.065grcm3 , 700mosMrkg.
15% Ž1.054grcm3 , 630mosMrkg.
10% Ž1.045grcm3 , 490mosMrkg.
1r2
1r4
1r3
1r3
0r1
0r0
2r0
3r0
3r1
3r1
4r3
1r1
Data are expressed as a ratio of uroepithelial cells to UL present in 20 representative high-power fields Ž40 = ..
150 J. Stachowski et al.r Journal of Immunological Methods 213 (1998) 145–155 Fig. 2. Phase-contrast microscopy analysis of sequential steps of the two-step density gradient centrifugation. ŽA. Interphase from the first-step density gradient centrifugation on the Ficoll–Isopaque. UL contaminated with uroepithelial cells; ŽB. Interphase of the second-step density gradient centrifugation on Uropolinum. Pure UL; ŽC. Sediment after Uropolinum centrifugation. Urothelial cells.
J. Stachowski et al.r Journal of Immunological Methods 213 (1998) 145–155
Fig. 3. Flow cytometry instrument setting in UL analysis. Simultaneous analysis of peripheral blood lymphocytes and UL was performed to set an appropriate lymphocyte gate, which was used as a ‘live gate’ in the UL studies. ŽA. Lymphocyte ‘live gate’ setting in peripheral blood lymphocytes. ŽB. Urine sediment before UL isolation Žapprox. 2 h after harvesting.; Ž1. cell debrisqerythrocytes, Ž2. UL, Ž3. apoptotic cellsqdead cells, Ž4. tubular epithelium ŽPTC.qapoptotic cellsqdead cells, Ž5. leucocytesqtubular epithelium ŽDTC., Ž6. urothelial cells. ŽC. UL after two-step centrifugation. ŽD. UL after two-step centrifugation—counting only in the ‘live gate’.
151
152 J. Stachowski et al.r Journal of Immunological Methods 213 (1998) 145–155 Fig. 4. Flow cytometric analysis of UL markers CD45RA and CD4 in relation to the time interval between urine harvesting and UL isolation. Dot blots of double-colour fluorescence of CD45RArFITC-FL1r and CD4rPE-FL2r. Percentage and mean fluorescence intensity ŽMFI. of CD45RA negative and CD45RA positive clusters are presented for one representative experiment performed with UL isolated from a patient with acute rejection of a kidney transplant.
J. Stachowski et al.r Journal of Immunological Methods 213 (1998) 145–155
153
The separation of UL from epithelial cells was necessary since some epithelial cells Že.g., urothelial cells s 25–80 m m. are much larger than lymphocytes and monocytes Ž6–8 m m and 12–20 m m, respectively.. In some transplant patients the number of uroepithelial cells was extremely high compared to the lymphocyte counts in urine. Therefore, in these cases the relative number of UL was very low after urine harvesting. We attempted to exclude uroepithelial cells by dual scatter gating. However, the cytometric measuring time was extremely prolonged when the urine lymphocytes were not first isolated by the proposed two-step centrifugation. It proved difficult to show the largest uroepithelial cells Žsize approx. 80 m m. as a separate cluster well demarcated from the right axis display in a dual scatter on a linear scale ŽFig. 3B.. A logarithmic scale, additionally used in dual scatter, did not improve the separation of this urothelial cluster from the lymphocyte cluster, dead cells and debris Ždata not shown..
this percentage could reach 10–25% and 12–38%, respectively, if the FCS was not adjusted Ždata not presented.. The analysis of UL surface markers was crucially influenced by the time interval between collection and preparation of UL. Adequate flow cytometric analysis was best up to 1 h following urine harvesting ŽFig. 4.. The receptor density of the surface molecules Že.g., CD45RA and CD4. vanished depending on the time interval of urine collection and subsequent UL isolation. Two separate clusters of CD45RAyCD4q and CD45RAqCD4q cells could be observed only in the first hour after urine harvesting. However, the collection of urine on PBS q 30% FCS extended the useful harvesting time up to 2–3 h. The same tendency was observed for other surface molecules such as CD3 and HLA-DRrIar Ždata not shown..
3.3. Phenotyping of urine lymphocytes using flow cytometry
Urine cytology is a noninvasive method. Large amounts of cell material can be harvested, but only small numbers of UL can be concentrated. Cell disintegration, depending on urine osmolality, makes the evaluation of cell surface markers difficult. Shrinkage and overstaining are possible. The present study is unique in that it is the first to define the optimal conditions of urine collection and flow cytometer settings in order to analyse the UL markers in renal transplant rejection. We have found that UL viability is mainly dependent on the time interval after urine harvesting and isolation. Directly after urine harvesting only a few UL revealed signs of death Ž5–7%. or apoptosis Ž7–9%.. This portion increased proportionally with time after voiding Ždata not shown.. High urine osmolalities exerted a negative influence on UL viability due to the high concentrations of urea and glucose. In contrast, high protein concentrations in the urine protected the UL viability. The viability of the cell is a crucial problem in flow cytometric analysis ŽJorge et al., 1991; Lao et al., 1993; Marcussen et al., 1992; Roberti et al., 1995.. Nonlymphoid cells often change their lightscattering characteristics when they die and may be mistakenly counted as lymphocytes ŽJung et al.,
As the count of isolated UL was low compared to isolated blood lymphocytes, the flow cytometric analysis was confined to the ‘live gate’ instrument setting ŽFig. 3.. This lymphocyte gate cut off other events and focused only on the examined UL. The number of pure UL could be enhanced from 5000 to 10,000 collected events in the gate. The events detected in the areas outside the lymphocyte gate on the scattergram were characterised on a further 52 urine samples with monoclonal antibodies recognising urothelial and tubular epithelial cells ŽBaer et al., 1997; Fischer et al., 1990; Segasothy et al., 1989.. Apoptotic vs. dead cells were differentiated using Annexin-V-FITC and PI ŽVermes et al., 1995.. Tubular epithelial cells ŽDTC, PTC. were mainly seen in areas Ž4. and Ž5. ŽFig. 3B., and urothelial cells in area Ž6.. Apoptotic and dead cells were localised in different areas of the scattergram, especially in areas Ž1., Ž3., Ž4., but also in the lymphocyte area Ž2.. In freshly harvested urine sediments the percentage of apoptotic and dead cells in all counted events ranged from 7–9% and 5–7%, respectively. During the first 2 h after urine harvesting
4. Discussion
154
J. Stachowski et al.r Journal of Immunological Methods 213 (1998) 145–155
1993; Lan et al., 1996.. Nonviable cells often exhibit increased nonspecific staining. Therefore, major subpopulations of cells can be lost from the analysis due to a higher background ŽOwens and Loken, 1995.. Dead granulocytes and other cells may be included in a lymphocyte light-scatter gate and, thus, must be identified to correct the results. Nonviable cells may increase the double-negative population in certain reagent combinations or may cause difficulties in distinguishing the ‘dim’ Žlow RD. from negative cell populations. Thus, in the analysis of UL the preservation of high viability in the examined cells is an absolute requirement. An essential factor diminishing the UL yield was the number of erythrocytes layered onto the density gradient. Overloading the gradient with erythrocytes decreased lymphocyte recovery by trapping lymphocytes during sedimentation. Therefore, in cases with haematuria erythrocyte lysis was performed before density gradient centrifugation. Urine samples collected from renal transplant patients offered the possibility of assessing optimal UL isolation and the setting of flow cytometry parameters. It is known that the predominance of proximal tubular cells washed out with the urine may indicate acute tubular necrosis ŽATN. andror cyclosporin A nephrotoxicity ŽJorge et al., 1991; Kyo et al., 1992a,b; Kyo and Mihatsch, 1991.. These epithelial cells may also be present in e.g., transplant rejection ŽJorge et al., 1991; Lee et al., 1992; Roberti et al., 1995; Simpson et al., 1989.. The separation of UL from uroepithelial cells reduces the time needed for flow cytometric measurements and provides a rather fast method to assess marker proteins expressed by UL in order to diagnose rejection. Analysis of UL by the described means allows longitudinal, noninvasive monitoring of the intragraft immune response ŽBarth et al., 1996.. Based on the experiments performed, the following algorithm of adequate UL isolation and preparation for flow cytometric analysis is proposed: Ž1. morning urine harvesting directly onto FCS Ž30% vrv.; Ž2. UL isolation within 2 h; Ž3. erythrocyte lysis with subsequent two-step density gradient isolation of UL from residual erythrocytes, granulocytes ŽFicoll–Isopaque, 1.077 grcm3 . and from uroepithelial cells Ž30% methylglucamine 3,5-diacetomido2,4,6-triiodobenzoicum, 1.085 grcm3 .; Ž4. flow cy-
tometric analysis of UL using the ‘live gate’ setting in the area of blood lymphocyte cluster. Each urine sample has an individual composition but applying this protocol we were able to achieve optimal conditions for routine flow cytometric analysis of UL in all patients.
Acknowledgements The monoclonal antibodies against PTC ŽAPM, mAb CD13 and mAb 138H11. and against DTC ŽTHG; mAb 5A2. were a gift from Dr. Patrick Baer of the Department of Nephrology at the Johann Wolfgang Goethe-University in FrankfurtrMain in Germany. This study was supported by a grant from the Else-Kroener-Fresenius Foundation and Kuratorium fuer Dialyse und Nierentransplantation ŽKfH, Germany. as well as in part by grant No. KBN PB 4P05E 046 14 from the Committee for Scientific Research, Poland. The study was partially presented at the 29th Annual Meeting of the American Society of Nephrology ŽASN. in New Orleans, USA, November 3–6, 1996. The authors wish to acknowledge Prof. K. Wielckens for the generous access to flow cytometric equipment. We are very grateful to Ms. Elanne Smootz for her language assistance.
References Baer, P.C., Nockher, A.W., Haase, W., Scherberich, J.E., 1997. Isolation of proximal and distal tubule cells from human kidney by immunomagnetic separation. Kidney Int. 61, in press. Barth, C., Stachowski, J., Lammerding, P., Baldamus, C.A., 1996. Increased ratio of CD45R0rRA in urine lymphocytes ŽUL. heralds acute renal rejection: monitoring by flow cytometry. J. Am. Soc. Nephrol. 7, 1930. Dooper, I.M.M., Bogman, M.J.J.T., Hoitsma, A.J., Maass, C.N., Vooijs, P.G., Koene, R.A.P., 1991. Immunocytology of urinary sediments as a method of differentiating acute rejection from other causes of declining renal graft function. Transplantation 52, 266–271. Fischer, P., Scherberich, J.E., Schoeppe, W., 1990. Comparative biochemical studies on gamma-glutamyltransferases from human kidney and renal cell carcinoma applying monoclonal antibodies. Clin. Chim. Acta 191, 185–200. Gorski, A., Kardasiewicz, H., Wyzgał, J., Durlik, M., 1992. ´ Diminished integrin expression on granulocytes from recipients. Arch. Immunol. Ther. Exp. 40, 71–74.
J. Stachowski et al.r Journal of Immunological Methods 213 (1998) 145–155 Hammer, C., Reichenspurner, H., Klima, G., 1993. Immunologic parameters for diagnosis of graft rejection. Transplant. Proc. 25, 26–29, Suppl. 3. Jorge, J.T.G., Estrada, C., Gonzalez, Z.A., Morales-Otero, L.A., Lavergne, J., Santiago-Delpin, E.A., 1991. Flow cytometric analysis of urine sediment after kidney transplantation. Transplant. Proc. 23, 1764–1765. Jung, T., Schauer, U., Heusser, C., Neumann, C., Rieger, C., 1993. Detection of intracellular cytokines by flow cytometry. J. Immunol. Meth. 159, 197. Kyo, M., Mihatsch, M.J., 1991. Differential diagnosis of kidney transplant rejection and cyclospirin nephrotoxicity by urine cytology. Acta Urol. Jpn. 37, 1153–1158. Kyo, M., Gudat, F., Dalquen, P., Fujimoto, N., Ichikawa, Y., Fukunishi, T., Nagano, S., Mihatsch, M.J., 1992a. Differential diagnosis of kidney transplant rejection and cyclosporine nephrotoxicity by urine cytology. Transplant. Proc. 24, 1388– 1390. Kyo, M., Michatsch, M.J., Gudat, F., Dalquen, P., Huser, B., Thiel, G., 1992b. Renal graft rejection or cyclosporin toxicity? Early diagnosis by a combination of Papanicolaou and immunocytochemical staining of urinary cytology specimens. Transplant. Int. 5, 71–76. Lan, H.Y., Hutchinson, P., Tesch, G.H., Mu, W., Atkins, R.C., 1996. A novel method of microwave treatment for detection of cytoplasmic and nuclear antigens by flow cytometry. J. Immunol. Meth. 190, 1–10. Lao, M., Wyzgał, J., Nowaczyk, M., Gorski, A., 1993. Value of ´ urine sediment phenotyping in renal allograft recipients. Arch. Immunol. Ther. Exp. 41, 373–375. Lee, P.H., Huang, M.T., Lee, C.S., 1992. Analysis of urine cytology by flow cytometry in renal transplantation. Transplant. Proc. 24, 1543–1544. Marcussen, N., Schumann, J.L., Schumann, G.B., Parmar, M., Kjellstrand, C., 1992. Analysis of cytodiagnostic urinalysis
155
findings in 77 patients with concurrent renal biopsies. Am. J. Kidney Dis. 20, 618–628. Owens, M.A., Loken, M.R. ŽEds.., 1995. Flow Cytometry Principles for Clinical Laboratory Practice. Quantitative Assurance for Quantitive Immunophenotyping. Wiley-Liss, New York, p. 45. Radio, S.J., Stratta, R.J., Taylor, R.J., Linder, J., 1993. The utility of urine cytology in the diagnosis of allograft rejection after combined pancreas–kidney transplantation. Transplantation 55, 509–516. Roberti, I., Reisman, L., Burrows, L., Lieberman, K.V., 1995. Urine cytology and urine flow cytometry in renal transplantation—a prospective double blind study. Transplantation 59, 495–500. Segasothy, M., Birch, D.F., Fairley, K.F., Kincaid-Smith, P., 1989. Urine cytologic profile in renal allograft recipients determined by monoclonal antibodies. Diagnosis of allograft rejection. Transplantation 47, 482–487. Simpson, M.A., Madras, P.N., Cornaby, A.J., Etienne, T., Dempsey, R.A., Clowes, G.H.A., Monaco, A.P., 1989. Sequential determinations of urinary cytology and plasma and urinary lymphokines in the management of renal allograft recipients. Transplantation 47, 218–223. Teague, K., El-Naggar, A., 1994. Comparative flow cytometric analysis of proliferating cell nuclear antigen ŽPCNA. antibodies in human solid neoplasms. Cytometry 15, 21–24. Vermes, I., Hannen, C., Steffens-Nakken, H., Reutelingsperger, C., 1995. A novel assay for apoptosis. Flow cytometric detection of phosphatidilserine expression on early apoptotic cells using fluorescein labelled Annexin V. J. Immunol. Meth. 184, 39–51. Xiao, X., Li, Y., Cheng, Y., Ao, J., 1992. Urine active T-rosetteforming cells in the diagnosis of acute allograft rejection after cadaveric kidney transplantation. Transplant. Proc. 24, 1545– 1546.
Flow cytometric analysis of urine lymphocytes isolated from patients with renal transplants—purification of urine lymphocytes Jozef ´ Stachowski a,b,c,) , Claudia Barth a, Maria Lewandowska-Stachowiak b, Petra Lammerding a , Dariusz Runowski b, Conrad A. Baldamus a a Department of Nephrology, UniÕersity of Cologne, Joseph-Stelzmann Str. 9, 50924 Koeln, Germany Department of Clinical Chemistry, UniÕersity of Cologne, Joseph-Stelzmann Str. 9, 50924 Koeln, Germany Department of Nephrology, Clinic of Pediatric Diseases, UniÕersity School of Medical Sciences, Szpitalna Str. 27 r 33, 60372 Poznan, ´ Poland b
c
Received 24 March 1997; revised 6 October 1997; accepted 1 December 1997
Abstract Early diagnosis of rejection is a pivotal problem in renal transplantation. Recent advances in urinary cell analysis using flow cytometry are still burdened with difficulties concerning urine lymphocyte ŽUL. isolation. The analysis of lymphocytes washed out with the urine from the kidney transplant offers a tool to monitor noninvasively the intragraft immune response. However, the demand for optimal isolation of UL with high viability and good separation of other cell types has not, as yet, been met. The present study was undertaken to evaluate the optimal conditions for harvesting UL in order to perform adequate UL analysis by flow cytometry. We found that UL viability is mainly dependent on the time of urine harvesting. Low UL viability was caused by high urine osmolality due to high concentrations of urea and glucose. In contrast, high protein concentrations protected UL viability. Hence, the following algorithm of adequate UL isolation for flow cytometric analysis was established: Ž1. Collection of morning urine directly onto foetal calf serum ŽFCS: 30% vrv.; Ž2. UL isolation within 2 h; Ž3. Erythrocyte lysis with subsequent two-step density gradient isolation of UL from residual erythrocytes, granulocytes ŽFicoll-Isopaque, 1.077 grcm3 . and from uroepithelial cells Ž30% methylglucamine 3,5-diacetomido-2,4,6-triiodobenzoicum, 1.085 grcm3 .; Ž4. Flow cytometric analysis of UL using the ‘live gate’ setting in the area of blood lymphocyte cluster. Adequate UL isolation and special settings of the flow cytometer may provide a useful tool for early diagnosis and the noninvasive monitoring of renal transplant rejection. q 1998 Elsevier Science B.V. All rights reserved. Keywords: Urine lymphocytes; Flow cytometry; Kidney transplant rejection
Abbreviations: APM: aminopeptidase M; ATN: acute tubular necrosis; DTC: distal tubular cells; FCS: foetal calf serum; FIT: fluorescein-isothiocyanate; FSC: forward scatter; g-GT: g-glutamyltransferase; MFI: mean fluorescence intensity; PBS: phosphate buffered saline; PE: phycoerythrin; PI: propidium iodide; PS: phospholipid phosphatidylserine; PTC: proximal tubular cells; RD: receptor density; SSC: side scatter; THG: Tamm–Horsfall glycoprotein; UL: urine lymphocytes; Uropolinum: methylglucamine 3,5-diacetomido-2,4,6-triiodo benzoicum ) Corresponding author. Tel.: q49-221-460-1951; fax.: q49-221-460-1945; e-mail: [email protected]. 0022-1759r98r$19.00 q 1998 Elsevier Science B.V. All rights reserved. PII S 0 0 2 2 - 1 7 5 9 Ž 9 8 . 0 0 0 1 3 - 1
146
J. Stachowski et al.r Journal of Immunological Methods 213 (1998) 145–155
1. Introduction
2. Materials and methods
Early diagnosis of rejection episodes is still a pivotal problem in renal transplantation. Since mononuclear cells washed out with the urine seem to reflect the intragraft immune response, the analysis of activation markers on the cell surface or within the cell offers the possibility of noninvasive monitoring. Flow cytometry is the technique of choice for the analysis of cell surface antigen expression ŽHammer et al., 1993; Lao et al., 1993; Lee et al., 1992; Marcussen et al., 1992; Radio et al., 1993.. Recent developments in detecting target antigens within the cytoplasm and nucleus have made flow cytometry applicable not only to surface but also to intracellular molecules such as cytokines and nuclear antigens ŽJung et al., 1993; Lan et al., 1996; Teague and El-Naggar, 1994.. However, flow cytometry requires a complicated cell preparation process in which cell purification, staining and fixation are crucial. Optimal staining time depends on the cell type under analysis Že.g., blood vs. urine lymphocytes. and on the characteristics of the selected antigen or receptor. In order to perform adequate flow cytometric analysis the following requirements have to be met: high viability and good separation of the lymphocytes from dead cells, debris and other cell types Že.g., uroepithelial cells.. The cytological profile in the urine of renal allograft recipients has so far been determined by light microscopy using direct smear and monoclonal antibody staining ŽDooper et al., 1991; Kyo et al., 1992a,b; Kyo and Mihatsch, 1991; Segasothy et al., 1989; Simpson et al., 1989; Xiao et al., 1992.. Only a few publications have applied flow cytometry to the analysis of urine sediment ŽGorski et al., 1992; ´ Jorge et al., 1991; Lao et al., 1993; Marcussen et al., 1992; Roberti et al., 1995.. None of these studies controlled for lymphocyte viability and purity following urine harvesting. Moreover, optimal urine collecting conditions including osmolality, urea, glucose and protein concentration, which may influence adequate flow cytometric analysis of UL, were not defined. Therefore, the data available to date are not sufficient for reproducible results using flow cytometry as a routine diagnostic tool. This study optimises and defines the conditions of adequate UL isolation for flow cytometric analysis.
2.1. HarÕesting of urine samples Fifty-six renal allograft recipients were followed posttransplant. Morning urine samples Žcatheter or spontaneous midstream. were collected on alternate days for the first 3 weeks, twice weekly for the next 2 weeks, and weekly for the subsequent time course. All collected samples were examined immediately and after 1, 2, 3, 4 and 5 h in order to estimate the influence of time on the viability and cell surface receptor expression of UL. 2.2. Viability of urine lymphocytes dependent on time and on urinary concentration of protein, glucose and urea In order to determine the optimal time point for UL analysis by flow cytometry after urine harvesting, cell viability was examined. The viability of cells recovered from the interphase of density gradient centrifugation was assessed by trypan blue dye exclusion. Dead UL were distinguished from apoptotic UL directly after urine harvesting and before UL isolation Žsee Section 2.4.. The influence of the time interval between urine harvesting and flow cytometric analysis, as well as the influence of urine osmolality and different concentrations of protein, glucose and urea in the urine influencing UL viability, was analysed simultaneously. The pH value of the urine sample was adjusted to 7.0. Correlation coefficients for viability were calculated over time with respect to urine osmolality, urine urea, protein and glucose concentrations using SigmaStat for Windows, MATHCAD software or an in-house program. 2.3. Isolation of urine lymphocytes— separation from uroepithelial cells Collected urine specimens supplemented with FCS Ž30% vrv. were centrifuged for 10 min at 200 = g. The supernatant was decanted, the cells washed with PBS ŽDulbecco without Caq2 , Mgq2 . and resuspended in PBS containing 30% FCS ŽDifco, Detroit, MI, USA.. Subsequent isolation of the lymphocyte population and separation from granulocytes, as well
J. Stachowski et al.r Journal of Immunological Methods 213 (1998) 145–155
as uroepithelial cells Žproximal and distal tubular cells, collecting duct, urothelium., were performed using two-step density gradient centrifugation Ž300 = g, 30 min. on Ficoll-Isopaque ŽLymphoflot-Biotest Diagnostics, Germany; density 1.077 grcm3 and ca. 810 mosMrkg. and on 30% sodium methylglucamine 3,5-diacetomido-2,4,6-triiodobenzoicum ŽUropolinum, Polfa, Poland; density 1.085 grcm3 and ca. 880 mosMrkg.. In samples with massive haematuria, lysis of erythrocytes was performed using 0.84% NH 4 Cl before density gradient centrifugation Ž Ficoll or Uropolinum.. 2.4. Flow cytometry Experiments were performed on the FACScan flow cytometer ŽBecton-Dickinson, Heidelberg, Germany. with FACScan research and Consort 30 software and subsequent additional analysis using DATA-Mate ŽBecton-Dickinson. software. An argon ion laser operating at 488 nm was used to illuminate the cells. The filters used to measure fluorescein ŽFITC. and phycoerythrin ŽPE. fluorescence were 530-nm band-pass and 630-nm long-pass filters, respectively. For data acquisition we used a linear amplification mode for forward-scatter ŽFSC. and side-scatter ŽSSC. parameters and logarithmic amplification for fluorescence 1 ŽFL1-FITC. and fluorescence 2 ŽFL2 s PE.. Fluorescence histograms of UL of at least 5000 counts were generated using a ‘live gate’ set-up on the blood lymphocytes isolated from the same individual using FSC vs. SSC scattergram for the background Žcomparison.. The percentage of positive cells was measured from a cut-off set using an isotype matched nonspecific control antibody, while the mean channel fluorescence ŽMFI s mean fluorescence intensity presented in arbitrary units. was measured over the entire lymphocyte subpopulation. MFI was proportional to the number of receptor binding sites on the cell surface and was determined as the mean receptor density per cell ŽRD.. UL were stained in double-colour fluorescence with the monoclonal antibodies anti-CD45RA labelled with FITC and anti-CD4 conjugated with PE ŽBecton-Dickinson.. Data analysis was carried out using dot plot statistics of either Consort 30, FACScan or Data Mate software ŽBecton-Dickinson..
147
In order to characterise the events in those areas excluded from the ‘live gate’ setting, the whole urine sediment Ž n s 52. was stained with Annexin-V-FITC as a marker for apoptosis and propidium iodide ŽPI. as a marker for cell death ŽPharmingen. ŽVermes et al., 1995.. Annexin-V is a 35–36-kDa Ca2q-dependent phospholipid-binding protein that has a high affinity for the membrane phospholipid phosphatidylserine ŽPS.. In apoptotic cells, PS is translocated from the inner to the external cellular membrane; hence, Annexin-V-FITC bound to PS identifies cells undergoing apoptosis, especially in the earlier stages of apoptosis. PI is a standard flow cytometric viability probe distinguishing viable from nonviable cells ŽVermes et al., 1995.. Tubular epithelial and urothelial cells were detected by monoclonal antibodies URO8 and URO9, respectively, and were estimated in parallel by indirect fluorescence staining ŽSegasothy et al., 1989.. Improved analysis of tubular epithelial cells was performed using mAbs against aminopeptidase M ŽAPM, mAb CD13. and against g-glutamyltransferase Žg-GT, mAb 138H11. as markers of proximal tubular cells ŽPTC. ŽFischer et al., 1990.. The distal tubular cells ŽDTC. were identified using a mAb recognising Tamm–Horsfall glycoprotein ŽTHG; mAb 5A2., a specific antigen for the thick ascending limb and the early distal convoluted tubule ŽBaer et al., 1997.. Flow cytometric localisation of these cell types was performed using PAINT-A-GATE software. These parameters of events excluded from the ‘live gate’ were assessed on a separate set of 52 urine samples from renal transplant patients. Hence, these parameters are not included in this study.
3. Results 3.1. Viability of urine lymphocytes— dependence on the time interÕal after urine harÕesting, urine osmolality and concentration of urea, glucose, and protein in the urine Urine osmolality is mainly determined by the concentrations of urea, glucose and protein. These concentrations may vary widely even in physiological conditions ŽTables 1–3, Fig. 1.. UL viability
J. Stachowski et al.r Journal of Immunological Methods 213 (1998) 145–155
148
Table 1 Viability a of UL: influence of time and osmolality
Table 3 Viability of UL: influence of time and urine glucose concentration
Osmolality ŽmosM.
Time course after urine-sample harvesting 1h 2h 3h 4h 5h
Glucose Žmgrdl.
300–400 400–500 500–600 600–700 700–800 800–900 900–1000
100 100 96 45 50 20 11
- 20 20–50 50–100 100–200 200–400 400–600 600–800 800–1000 1000–2000 2000–4000
90 87 80 72 33 18 10
75 50 38 24 11 5 4
42 38 25 16 12 7 4
33 25 20 12 8 7 5
a
Viability is expressed as a percentage of all cells. Mean values for 1322 analysed urine samples are presented.
Time course after urine-sample harvesting 1h
2h
3h
4h
5h
100 98 94 83 80 71 45 32 20 9
95 90 92 87 80 82 70 33 18 5
60 50 37 24 12 11 6 5 4 3
20 22 21 17 14 15 13 11 6 3
22 22 20 22 11 10 7 7 4 2
a
Viability is expressed as a percentage of all cells. Mean values for 1322 analysed urine samples are presented.
is diminished by urea- and glucose-associated high osmolality after urine harvesting ŽTables 2 and 3.. The urine samples with high glucose concentrations were collected from diabetic patients, including a few diabetics with a poor blood-glucose adjustment, where the urine glucose reached values between 1000–4000 mgrdl ŽTable 3.. However, at the same time these patients showed a urine pH decrease, which may reveal additional negative influence on the UL viability in vitro Ždata not shown.. On the other hand, a high protein concentration in the urine had a highly protective effect on UL viability ŽFig. 1.. The separate influence of each parameter on UL viability does not reflect their mutual influence exerted at the same time under physiologic conditions. Hence, the urine-derived data were pooled and analysed by the six-parameter data matrix test taking into consideration the influence of protein, glucose and
Table 2 Viability of UL: influence of time and urea concentration Urea Žmgrdl.
Time course after urine-sample harvesting 1h
2h
3h
4h
5h
- 500 500–1000 1000–1500 1500–2000 2500–3000 3500–4000
98 96 96 91 60 40
85 90 70 60 56 32
68 50 25 20 18 8
40 30 20 10 8 4
30 20 10 5 2 0
a
Viability is expressed as a percentage of all cells. Mean values for 1322 analysed urine samples are presented.
urea concentrations, osmolality and time after urine collection on UL viability in vitro. This multiparameter regression analysis confirmed our findings of the protective effect of protein on UL viability. Computer simulation showed that the negative effect of a high urea concentration on UL viability Žup to 3590 mgrdl. was diminished by a urine protein concentration of 8100 mgrdl, which extended the viability to 96% for 2 h or to 76% for 3 h after urine harvesting. Therefore, the high correlation coefficient Žrange: 0.90–0.92. between protein concentration ŽG 8100 mgrdl. and UL viability was still present even when the concentration of urea or glucose was higher than 3600 mgrdl or 1000 mgrdl, respectively. Thus, in order to benefit from the protective effect of protein, urine samples were adjusted with FCS Ž10–30 grdl. as exogenous protein ŽFig. 1.. This procedure ensured that in the time course of 2–3 h following urine harvesting, UL viability ranged between 88% to 100% ŽFCS,30 grdl., independent of an unpredictable urine composition. 3.2. Algorithm of urine lymphocyte isolation The experiments concerning the influence of osmolality and different concentrations of protein, urea and glucose in the urine suggest the following schedule of UL collection in order to reach the optimal conditions of UL isolation. Urine samples Ž50–150 ml. have to be collected in the morning and should
J. Stachowski et al.r Journal of Immunological Methods 213 (1998) 145–155
149
Fig. 1. Viability of UL influenced by urine protein concentration and the time after urine harvesting. Urine samples Ž n s 1322. were aliquoted according to the protein concentration and incubated at room temperature for 5 h. At 1-h intervals UL were isolated by density gradient centrifugation and their viability was assessed. Some urine samples Ž n s 120. were adjusted to 10 grdl, 20 grdl or 30 grdl with exogenous FCS immediately after urine harvesting. The bar graph presents the mean values.
be directly supplemented with FCS Ž30% vrv.. The isolation procedure should be performed within a maximum of 2 h. A UL count of 5.5 = 10 3 –10.5 = 10 4rml can be obtained depending on the clinical conditions. UL were found to be mainly present in patients with acute or chonic rejection; the count was extremely low in patients with stable transplant function. The FCS-supplemented urine was centrifuged for 10 min at 200 = g and the sediment was resuspended in PBS q 30% FCS. This diluted urine sediment was centrifuged on Ficoll with subsequent centrifugation on 30% methylglucamine 3,5-diacetomido-2,4,6-triiodobenzoicum ŽUropolinum .. The density gradient of Uropolinum was determined
in a multiple series of experiments designed to optimise the procedure for maximal recovery of UL and minimal contamination by erythrocytes, epithelial cells and granulocytes ŽTable 4.. After this centrifugation procedure the isolated cell population contained ) 92% lymphocytes and 4%–6% monocytic cells as determined by peroxidase and May– Gruenwald–Giemsa staining. The epithelial cells were almost totally separated from the lymphocytes as demonstrated by phase-contrast microscopy ŽFig. 2; Table 4.. When only a small number of epithelial cells was present in urine samples Ž- 30rml., centrifugation on a Ficoll density gradient yielded sufficient cells for submission to flow cytometry analysis.
Table 4 Efficiency of separating urine sediment epithelial cells from lymphocytes using density gradient centrifugation on sodium methylglucamine 3,5-diacetomido-2,4,6-triiodobenzoicumrUropolinumr a Uropolinum
Interphase epithelial cells: lymphocytes Sediment epithelial cells: lymphocytes a
40% Ž1.096grcm3 , 1060mosMrkg.
30% Ž1.085grcm3 , 880mosMrkg.
25% Ž1.076grcm3 , 800mosMrkg.
20% Ž1.065grcm3 , 700mosMrkg.
15% Ž1.054grcm3 , 630mosMrkg.
10% Ž1.045grcm3 , 490mosMrkg.
1r2
1r4
1r3
1r3
0r1
0r0
2r0
3r0
3r1
3r1
4r3
1r1
Data are expressed as a ratio of uroepithelial cells to UL present in 20 representative high-power fields Ž40 = ..
150 J. Stachowski et al.r Journal of Immunological Methods 213 (1998) 145–155 Fig. 2. Phase-contrast microscopy analysis of sequential steps of the two-step density gradient centrifugation. ŽA. Interphase from the first-step density gradient centrifugation on the Ficoll–Isopaque. UL contaminated with uroepithelial cells; ŽB. Interphase of the second-step density gradient centrifugation on Uropolinum. Pure UL; ŽC. Sediment after Uropolinum centrifugation. Urothelial cells.
J. Stachowski et al.r Journal of Immunological Methods 213 (1998) 145–155
Fig. 3. Flow cytometry instrument setting in UL analysis. Simultaneous analysis of peripheral blood lymphocytes and UL was performed to set an appropriate lymphocyte gate, which was used as a ‘live gate’ in the UL studies. ŽA. Lymphocyte ‘live gate’ setting in peripheral blood lymphocytes. ŽB. Urine sediment before UL isolation Žapprox. 2 h after harvesting.; Ž1. cell debrisqerythrocytes, Ž2. UL, Ž3. apoptotic cellsqdead cells, Ž4. tubular epithelium ŽPTC.qapoptotic cellsqdead cells, Ž5. leucocytesqtubular epithelium ŽDTC., Ž6. urothelial cells. ŽC. UL after two-step centrifugation. ŽD. UL after two-step centrifugation—counting only in the ‘live gate’.
151
152 J. Stachowski et al.r Journal of Immunological Methods 213 (1998) 145–155 Fig. 4. Flow cytometric analysis of UL markers CD45RA and CD4 in relation to the time interval between urine harvesting and UL isolation. Dot blots of double-colour fluorescence of CD45RArFITC-FL1r and CD4rPE-FL2r. Percentage and mean fluorescence intensity ŽMFI. of CD45RA negative and CD45RA positive clusters are presented for one representative experiment performed with UL isolated from a patient with acute rejection of a kidney transplant.
J. Stachowski et al.r Journal of Immunological Methods 213 (1998) 145–155
153
The separation of UL from epithelial cells was necessary since some epithelial cells Že.g., urothelial cells s 25–80 m m. are much larger than lymphocytes and monocytes Ž6–8 m m and 12–20 m m, respectively.. In some transplant patients the number of uroepithelial cells was extremely high compared to the lymphocyte counts in urine. Therefore, in these cases the relative number of UL was very low after urine harvesting. We attempted to exclude uroepithelial cells by dual scatter gating. However, the cytometric measuring time was extremely prolonged when the urine lymphocytes were not first isolated by the proposed two-step centrifugation. It proved difficult to show the largest uroepithelial cells Žsize approx. 80 m m. as a separate cluster well demarcated from the right axis display in a dual scatter on a linear scale ŽFig. 3B.. A logarithmic scale, additionally used in dual scatter, did not improve the separation of this urothelial cluster from the lymphocyte cluster, dead cells and debris Ždata not shown..
this percentage could reach 10–25% and 12–38%, respectively, if the FCS was not adjusted Ždata not presented.. The analysis of UL surface markers was crucially influenced by the time interval between collection and preparation of UL. Adequate flow cytometric analysis was best up to 1 h following urine harvesting ŽFig. 4.. The receptor density of the surface molecules Že.g., CD45RA and CD4. vanished depending on the time interval of urine collection and subsequent UL isolation. Two separate clusters of CD45RAyCD4q and CD45RAqCD4q cells could be observed only in the first hour after urine harvesting. However, the collection of urine on PBS q 30% FCS extended the useful harvesting time up to 2–3 h. The same tendency was observed for other surface molecules such as CD3 and HLA-DRrIar Ždata not shown..
3.3. Phenotyping of urine lymphocytes using flow cytometry
Urine cytology is a noninvasive method. Large amounts of cell material can be harvested, but only small numbers of UL can be concentrated. Cell disintegration, depending on urine osmolality, makes the evaluation of cell surface markers difficult. Shrinkage and overstaining are possible. The present study is unique in that it is the first to define the optimal conditions of urine collection and flow cytometer settings in order to analyse the UL markers in renal transplant rejection. We have found that UL viability is mainly dependent on the time interval after urine harvesting and isolation. Directly after urine harvesting only a few UL revealed signs of death Ž5–7%. or apoptosis Ž7–9%.. This portion increased proportionally with time after voiding Ždata not shown.. High urine osmolalities exerted a negative influence on UL viability due to the high concentrations of urea and glucose. In contrast, high protein concentrations in the urine protected the UL viability. The viability of the cell is a crucial problem in flow cytometric analysis ŽJorge et al., 1991; Lao et al., 1993; Marcussen et al., 1992; Roberti et al., 1995.. Nonlymphoid cells often change their lightscattering characteristics when they die and may be mistakenly counted as lymphocytes ŽJung et al.,
As the count of isolated UL was low compared to isolated blood lymphocytes, the flow cytometric analysis was confined to the ‘live gate’ instrument setting ŽFig. 3.. This lymphocyte gate cut off other events and focused only on the examined UL. The number of pure UL could be enhanced from 5000 to 10,000 collected events in the gate. The events detected in the areas outside the lymphocyte gate on the scattergram were characterised on a further 52 urine samples with monoclonal antibodies recognising urothelial and tubular epithelial cells ŽBaer et al., 1997; Fischer et al., 1990; Segasothy et al., 1989.. Apoptotic vs. dead cells were differentiated using Annexin-V-FITC and PI ŽVermes et al., 1995.. Tubular epithelial cells ŽDTC, PTC. were mainly seen in areas Ž4. and Ž5. ŽFig. 3B., and urothelial cells in area Ž6.. Apoptotic and dead cells were localised in different areas of the scattergram, especially in areas Ž1., Ž3., Ž4., but also in the lymphocyte area Ž2.. In freshly harvested urine sediments the percentage of apoptotic and dead cells in all counted events ranged from 7–9% and 5–7%, respectively. During the first 2 h after urine harvesting
4. Discussion
154
J. Stachowski et al.r Journal of Immunological Methods 213 (1998) 145–155
1993; Lan et al., 1996.. Nonviable cells often exhibit increased nonspecific staining. Therefore, major subpopulations of cells can be lost from the analysis due to a higher background ŽOwens and Loken, 1995.. Dead granulocytes and other cells may be included in a lymphocyte light-scatter gate and, thus, must be identified to correct the results. Nonviable cells may increase the double-negative population in certain reagent combinations or may cause difficulties in distinguishing the ‘dim’ Žlow RD. from negative cell populations. Thus, in the analysis of UL the preservation of high viability in the examined cells is an absolute requirement. An essential factor diminishing the UL yield was the number of erythrocytes layered onto the density gradient. Overloading the gradient with erythrocytes decreased lymphocyte recovery by trapping lymphocytes during sedimentation. Therefore, in cases with haematuria erythrocyte lysis was performed before density gradient centrifugation. Urine samples collected from renal transplant patients offered the possibility of assessing optimal UL isolation and the setting of flow cytometry parameters. It is known that the predominance of proximal tubular cells washed out with the urine may indicate acute tubular necrosis ŽATN. andror cyclosporin A nephrotoxicity ŽJorge et al., 1991; Kyo et al., 1992a,b; Kyo and Mihatsch, 1991.. These epithelial cells may also be present in e.g., transplant rejection ŽJorge et al., 1991; Lee et al., 1992; Roberti et al., 1995; Simpson et al., 1989.. The separation of UL from uroepithelial cells reduces the time needed for flow cytometric measurements and provides a rather fast method to assess marker proteins expressed by UL in order to diagnose rejection. Analysis of UL by the described means allows longitudinal, noninvasive monitoring of the intragraft immune response ŽBarth et al., 1996.. Based on the experiments performed, the following algorithm of adequate UL isolation and preparation for flow cytometric analysis is proposed: Ž1. morning urine harvesting directly onto FCS Ž30% vrv.; Ž2. UL isolation within 2 h; Ž3. erythrocyte lysis with subsequent two-step density gradient isolation of UL from residual erythrocytes, granulocytes ŽFicoll–Isopaque, 1.077 grcm3 . and from uroepithelial cells Ž30% methylglucamine 3,5-diacetomido2,4,6-triiodobenzoicum, 1.085 grcm3 .; Ž4. flow cy-
tometric analysis of UL using the ‘live gate’ setting in the area of blood lymphocyte cluster. Each urine sample has an individual composition but applying this protocol we were able to achieve optimal conditions for routine flow cytometric analysis of UL in all patients.
Acknowledgements The monoclonal antibodies against PTC ŽAPM, mAb CD13 and mAb 138H11. and against DTC ŽTHG; mAb 5A2. were a gift from Dr. Patrick Baer of the Department of Nephrology at the Johann Wolfgang Goethe-University in FrankfurtrMain in Germany. This study was supported by a grant from the Else-Kroener-Fresenius Foundation and Kuratorium fuer Dialyse und Nierentransplantation ŽKfH, Germany. as well as in part by grant No. KBN PB 4P05E 046 14 from the Committee for Scientific Research, Poland. The study was partially presented at the 29th Annual Meeting of the American Society of Nephrology ŽASN. in New Orleans, USA, November 3–6, 1996. The authors wish to acknowledge Prof. K. Wielckens for the generous access to flow cytometric equipment. We are very grateful to Ms. Elanne Smootz for her language assistance.
References Baer, P.C., Nockher, A.W., Haase, W., Scherberich, J.E., 1997. Isolation of proximal and distal tubule cells from human kidney by immunomagnetic separation. Kidney Int. 61, in press. Barth, C., Stachowski, J., Lammerding, P., Baldamus, C.A., 1996. Increased ratio of CD45R0rRA in urine lymphocytes ŽUL. heralds acute renal rejection: monitoring by flow cytometry. J. Am. Soc. Nephrol. 7, 1930. Dooper, I.M.M., Bogman, M.J.J.T., Hoitsma, A.J., Maass, C.N., Vooijs, P.G., Koene, R.A.P., 1991. Immunocytology of urinary sediments as a method of differentiating acute rejection from other causes of declining renal graft function. Transplantation 52, 266–271. Fischer, P., Scherberich, J.E., Schoeppe, W., 1990. Comparative biochemical studies on gamma-glutamyltransferases from human kidney and renal cell carcinoma applying monoclonal antibodies. Clin. Chim. Acta 191, 185–200. Gorski, A., Kardasiewicz, H., Wyzgał, J., Durlik, M., 1992. ´ Diminished integrin expression on granulocytes from recipients. Arch. Immunol. Ther. Exp. 40, 71–74.
J. Stachowski et al.r Journal of Immunological Methods 213 (1998) 145–155 Hammer, C., Reichenspurner, H., Klima, G., 1993. Immunologic parameters for diagnosis of graft rejection. Transplant. Proc. 25, 26–29, Suppl. 3. Jorge, J.T.G., Estrada, C., Gonzalez, Z.A., Morales-Otero, L.A., Lavergne, J., Santiago-Delpin, E.A., 1991. Flow cytometric analysis of urine sediment after kidney transplantation. Transplant. Proc. 23, 1764–1765. Jung, T., Schauer, U., Heusser, C., Neumann, C., Rieger, C., 1993. Detection of intracellular cytokines by flow cytometry. J. Immunol. Meth. 159, 197. Kyo, M., Mihatsch, M.J., 1991. Differential diagnosis of kidney transplant rejection and cyclospirin nephrotoxicity by urine cytology. Acta Urol. Jpn. 37, 1153–1158. Kyo, M., Gudat, F., Dalquen, P., Fujimoto, N., Ichikawa, Y., Fukunishi, T., Nagano, S., Mihatsch, M.J., 1992a. Differential diagnosis of kidney transplant rejection and cyclosporine nephrotoxicity by urine cytology. Transplant. Proc. 24, 1388– 1390. Kyo, M., Michatsch, M.J., Gudat, F., Dalquen, P., Huser, B., Thiel, G., 1992b. Renal graft rejection or cyclosporin toxicity? Early diagnosis by a combination of Papanicolaou and immunocytochemical staining of urinary cytology specimens. Transplant. Int. 5, 71–76. Lan, H.Y., Hutchinson, P., Tesch, G.H., Mu, W., Atkins, R.C., 1996. A novel method of microwave treatment for detection of cytoplasmic and nuclear antigens by flow cytometry. J. Immunol. Meth. 190, 1–10. Lao, M., Wyzgał, J., Nowaczyk, M., Gorski, A., 1993. Value of ´ urine sediment phenotyping in renal allograft recipients. Arch. Immunol. Ther. Exp. 41, 373–375. Lee, P.H., Huang, M.T., Lee, C.S., 1992. Analysis of urine cytology by flow cytometry in renal transplantation. Transplant. Proc. 24, 1543–1544. Marcussen, N., Schumann, J.L., Schumann, G.B., Parmar, M., Kjellstrand, C., 1992. Analysis of cytodiagnostic urinalysis
155
findings in 77 patients with concurrent renal biopsies. Am. J. Kidney Dis. 20, 618–628. Owens, M.A., Loken, M.R. ŽEds.., 1995. Flow Cytometry Principles for Clinical Laboratory Practice. Quantitative Assurance for Quantitive Immunophenotyping. Wiley-Liss, New York, p. 45. Radio, S.J., Stratta, R.J., Taylor, R.J., Linder, J., 1993. The utility of urine cytology in the diagnosis of allograft rejection after combined pancreas–kidney transplantation. Transplantation 55, 509–516. Roberti, I., Reisman, L., Burrows, L., Lieberman, K.V., 1995. Urine cytology and urine flow cytometry in renal transplantation—a prospective double blind study. Transplantation 59, 495–500. Segasothy, M., Birch, D.F., Fairley, K.F., Kincaid-Smith, P., 1989. Urine cytologic profile in renal allograft recipients determined by monoclonal antibodies. Diagnosis of allograft rejection. Transplantation 47, 482–487. Simpson, M.A., Madras, P.N., Cornaby, A.J., Etienne, T., Dempsey, R.A., Clowes, G.H.A., Monaco, A.P., 1989. Sequential determinations of urinary cytology and plasma and urinary lymphokines in the management of renal allograft recipients. Transplantation 47, 218–223. Teague, K., El-Naggar, A., 1994. Comparative flow cytometric analysis of proliferating cell nuclear antigen ŽPCNA. antibodies in human solid neoplasms. Cytometry 15, 21–24. Vermes, I., Hannen, C., Steffens-Nakken, H., Reutelingsperger, C., 1995. A novel assay for apoptosis. Flow cytometric detection of phosphatidilserine expression on early apoptotic cells using fluorescein labelled Annexin V. J. Immunol. Meth. 184, 39–51. Xiao, X., Li, Y., Cheng, Y., Ao, J., 1992. Urine active T-rosetteforming cells in the diagnosis of acute allograft rejection after cadaveric kidney transplantation. Transplant. Proc. 24, 1545– 1546.