Loss of Solute Carriers in T CellMediated ... - Wiley Online Library

American Journal of Transplantation 2010; 10: 2241–2251 Wiley Periodicals Inc.

C 2010 The Authors C 2010 The American Society of Journal compilation Transplantation and the American Society of Transplant Surgeons

doi: 10.1111/j.1600-6143.2010.03263.x

Loss of Solute Carriers in T Cell-Mediated Rejection in Mouse and Human Kidneys: An Active Epithelial Injury–Repair Response G. Eineckea , D. Kaysera , J. M. Vanslambrouckb , B. Sisc , J. Reevec , M. Mengelc , K. S. Famulskic , C. G. Baileyb , J. E. J. Raskob,d and P. F. Halloranc,e , * a Department of Nephrology, Hannover Medical School, Hanover, Germany b Gene and Stem Cell Therapy Program, Centenary Institute, University of Sydney, New South Wales, Sydney, Australia c Alberta Transplant Applied Genomics Centre, Department of Laboratory Medicine and Pathology, University of Alberta, Edmonton, Canada d Cell and Molecular Therapies, Sydney Cancer Centre, Royal Prince Alfred Hospital, Camperdown, New South Wales, Australia e Department of Medicine, University of Alberta, Edmonton, Canada *Corresponding author: Philip F. Halloran, [email protected]

T cell-mediated rejection of kidney allografts causes epithelial deterioration, manifested by tubulitis, but the mechanism remains unclear. We hypothesized that interstitial inflammation triggers a stereotyped epithelial response similar to that triggered by other types of injury such as ischemia-reperfusion. We identified solute carrier transcripts with decreased expression in mouse allografts, and compared their behavior in T cell-mediated rejection to native kidneys with ischemic acute tubular necrosis (ATN). Average loss of solute carrier expression was similar in ATN (77%) and T cell-mediated rejection (75%) with high correlation of individual transcripts. Immunostaining of SLC6A19 confirmed loss of proteins. Analysis of human kidney transplant biopsies confirmed that T cell-mediated rejection and ATN showed similar loss of solute carrier mRNAs. The loss of solute carrier expression was weakly correlated with interstitial inflammation, but kidneys with ATN showed decreased solute carriers despite minimal inflammation. Loss of renal function correlated better with decreased solute carrier expression than with histologic lesions (r = 0.396, p < 0.001). Thus the loss of epithelial transcripts in rejection is not a unique consequence of T cell-mediated rejection but an active injury–repair response of epithelium, triggered by rejection but also by other injury mechanisms.

Key words: Acute tubular necrosis, allograft rejection, gene expression, injury, kidney transplantation, renal epithelium Abbreviations: ATN, acute tubular necrosis; KT2, solute carrier transcript set; NCBA, normal CBA mouse kidneys; PBT, pathogenesis-based transcript set; TCMR, T cell-mediated rejection. Received 23 April 2010, revised 28 July 2010 and accepted for publication 29 July 2010

Introduction A key change in kidney transplants during T cell mediated rejection (TCMR) is deterioration of the epithelium (1,2), forming a diagnostic feature in biopsies—tubulitis. Although tubulitis is associated with T cells expressing cytotoxic molecules, studies in knockout mice have shown that tubulitis in TCMR is independent of T cell cytotoxic mechanisms and is not mediated by cytotoxicity (3–6). Tubulitis is associated with loss of epithelial cadherins such as Ecadherin and kidney specific cadherin, suggesting that the epithelium de-differentiates and loses its ability to exclude inflammatory cells (5). Tubulitis is inherently nonspecific, occurring in many renal diseases and injuries and is frequent in atrophic tubules. Thus histopathologic changes in the tubular epithelium during TCMR are not specific for TCMR. The epithelial response to TCMR and other injuries in mice and humans is reflected in transcript changes, with increased expression of embryonic genes and genes related to fibrogenesis and extracellular matrix remodeling, and decreased expression of renal parenchymal genes (7,8). In mouse kidney allografts, TCMR triggers early progressive loss of parenchymal transcripts, preceding the development of tubulitis (7). Loss of parenchymal transcripts also occurs in human renal allografts during TCMR (9). In the present study, we hypothesized that the epithelial changes in TCMR are not a unique effect of antigen-specific T cells acting on the epithelium, but part of an active injury– repair response of the epithelium triggered by many forms of injury. We used the loss of solute carriers as an indicator of this epithelial response. Using microarrays, we 2241

Einecke et al.

identified a subset of renal solute carriers that are highly expressed in the healthy renal epithelium and highly sensitive to kidney injury. We explored expression of these solute carrier transcripts in mouse kidneys and human kidney transplant biopsies with TCMR and nonimmunologic stress.

Materials and Methods Mouse experiments Male CBA/J (CBA) and C57Bl/6 (B6) mice were obtained from Jackson Laboratory (Bar Harbor, ME). Mice were maintained in the Health Sciences Laboratory Animal Services at the University of Alberta. Maintenance and experiments conformed to approved animal care protocols. Non-mechanical-supporting renal transplants were performed as previously described (10,11). Host mice did not receive immunosuppression. Kidneys were recovered on days 5, 7 or 21 posttransplant as previously described (12). Native kidneys of the appropriate strain (CBA) and isografts (CBA kidneys transplanted into CBA hosts) served as controls. Kidneys with acute tubular necrosis (ATN kidneys) were obtained by clamping the vascular pedicle of the left CBA kidney for 60 min, kept moistened with PBS at 37◦ C and then released. Animals were kept for 7 days and then sacrificed and the kidneys were designated ATN D7, as detailed previously (13). After recovering, one-half kidney was snap-frozen in liquid nitrogen and stored at −70◦ C, the other half was formalin-fixed and paraffin embedded.

Clinical biopsy specimens The study population consisted of 234 consecutive biopsies for cause from 173 patients at the University of Alberta and the University of Illinois, Chicago. These were part of the ongoing Genome Canada study (9). Histologically normal tissue from eight nephrectomies performed for renal cell carcinoma served as controls. The study was approved by the University of Alberta Health Research Ethics Board (Issue # 5299). Written informed consent was obtained from all study patients. Biopsies were obtained under ultrasound guidance by spring-loaded needles (ASAP Automatic Biopsy, Microvasive, Watertown, MA, USA). In addition to cores for conventional assessment, one 18-gauge biopsy core was collected for gene expression analysis and placed immediately in RNALater, kept at 4◦ C for 4–24 h, then stored at −20◦ C.

Histology and immunofluorescence Paraffin-embedded tissue sections were stained with periodic acid-Schiff and subjected to histologic analysis as described previously (3); patient biopsies were graded according to Banff criteria by a renal pathologist (BS) (14–16). All samples had adequate cortical tissue for analysis by Banff criteria with the exception of six biopsies with less than two large arteries (four biopsies had only one artery and two biopsies had no arteries). Immunofluorescence of mouse kidney sections was performed according to previous protocols (17,18). Sections were stained with custom polyclonal primary antibody, 1:200 chicken anti-SLC6A19 (Aves Laboratories, OR, immunizing peptide: CZ-DPNYEEFPKSQK). Secondary antibody was goat anR tichicken Alexa Fluor 488 (Invitrogen, Carlsbad, CA). Proximal tubules were identified using a biotinylated lectin marker, Lotus tetragonolobus (LTA; Vector Laboratories, Burlingame, CA), followed by streptavidin-Alexa R Fluor 594 (Invitrogen). Digital immunofluorescence images were obtained using the 40X air objective installed on an Olympus FluoViewTM FV1000 confocal microscope (Olympus, Japan) utilizing the 488 nm and 561 nm lasers according to the fluorochrome. Images were equally processed using Adobe Photoshop 7.0 (Adobe Systems, San Jose, CA).

2242

Gene expression analysis RNA extraction, dsDNA and cRNA synthesis, hybridization to MOE430 2.0 R (mouse) or HG-U133 Plus 2 (human) arrays (GeneChip, Affymetrix ), washR ing and staining were carried out according to the Affymetrix Technical Manual (www.affymetrix.com) as previously described (19). Arrays were scanned and CEL files were obtained using GeneChip Operating Software R 1.2, Affymetrix ; data was then preprocessed using Robust Multichip Average (RMA) normalization (20) as described previously (21) and further analyzed using GeneSpringTM software (Version 7.2, Agilent, Palo Alto, CA). We performed microarray analysis on the following mouse samples: mouse allografts (CBA kidneys transplanted into B6 hosts) and isografts (CBA kidneys transplanted into CBA hosts) at days 5, 7 and 21, nontransplanted CBA kidneys with ATN and native kidneys of CBA (NCBA) mice. For each array, equal amounts of RNA from three mice were pooled. The following number of arrays were available for analysis: n = 2 for ATN kidneys, isografts day 7 and 21; n = 3 for allografts at day 5 and 7 and for control kidneys; n = 4 for isografts day 5; and n = 5 for allografts at day 21. Mouse expression data was normalized against gene expression in native kidneys. For clinical biopsies, samples were analyzed individually and gene expression data was normalized against histologically normal tissue samples of eight nephrectomies. We had 18 biopsies diagnosed as ATN, 33 biopsies diagnosed as borderline TCMR and 49 biopsies diagnosed as TCMR available for analysis. Raw Gene expression data (CEL files) for all samples used in this analysis is available at http://transplants.med.ualberta.ca/Nephlab/data.html.

Selection of solute carrier transcripts for analysis We previously defined pathogenesis-based transcript sets in experimental mouse kidney transplants and cultured cells that reflected major biological processes during allograft rejection or other forms of injury (7,21). To reflect changes in the renal epithelium during rejection, we derived a renal transcript set, which is comprised of 70 (mouse) or 64 (human) solute carrier transcripts with high renal parenchymal expression and known epithelial function (7); all of these are decreased in rejection. For the selection of KT2.1 transcript set we used five arrays from mice allografts D21, all of them with typical rejection histopathology (i.e. allografts fulfiling the Banff criteria for severe TCMR), including dedifferentiated but viable epithelium. We identified the most affected transcripts as a subset of the original solute carrier transcript set (10): only those solute carrier transcripts (n = 26) with >90% loss of expression in the day 21 kidneys with severe rejection changes. The KT2.1 transcripts with high expression in normal kidney and severely decreased expression in the selected rejecting but viable kidneys must be selective for normal differentiated epithelium: they are absent in the viable dedifferentiated epithelium that remains in the rejecting kidneys and are absent in all other cell types (stroma, vascular and inflammatory) present in the day 21 allografts with severe rejection. For analysis in human biopsies, we identified the human orthologs of the mouse transcripts using the automatic tool in the GeneSpringTM software. Expression of KT2.1 solute carrier transcripts in each condition was summarized as the gene set score, which was calculated based on the following algorithm: for each array representing an experimental condition (pooled RNA from three mice) or an individual patient, we calculated the fold change compared to normal kidneys for each gene within the gene set. In a second step, we calculated the geometric mean of the fold changes of all genes in the gene set for each array. In a third step, we summarized the gene set scores within one experimental group (i.e. all arrays from one time point in mice or all arrays from patients with the same diagnosis) by calculating the average scores within this group. For each experimental group, solute carrier scores are presented as the arithmetic mean. Similarly,

American Journal of Transplantation 2010; 10: 2241–2251

Epithelium in Allograft Rejection

Figure 1: Histopathology of normal kidneys, isografts (CBA into CBA), allografts (CBA into B6) and acute tubular necrosis. Control normal kidneys are shown in panel A and B. Isografts at day 21 show normal histology (Panel C) and normal tubular morphology with intact PAS positive brush borders (Panel D). Allografts at day 21 (Panel E) show severe interstitial infiltrate with severe tubulitis and shrinkage of tubules with thinning and/or loss of brush borders in many tubules (Panel F). Kidneys with acute tubular necrosis (ATN) at day 7 show tubules with granular casts (Panel G) and signs of acute tubular injury (Panel H) with flattening of tubular epithelium, loss of PAS positive brush borders, variation in cell size and shape and patchy desquamation of individual epithelial cells leaving bare basement membrane (PAS, original magnification ×20 or ×60).

expression of individual transcripts in an experimental group is shown as the arithmetic mean in the group. Groups were compared by two-tailed t-test. Continuous variables were correlated by Spearman correlation (SPSS 16.0 statistical software package, SPSS inc., Chicago, IL, USA). Calculation of the gene set scores was performed on nontransformed numbers. All statistical analyses (correlations, group comparisons) were done on log-transformed values.

Results Pattern of solute carrier transcript loss is similar in mouse isografts, allografts and ATN The histopathologic changes in kidneys with ischemic ATN, isografts and allografts have been described in detail previously (3,7,13). Mouse kidney isografts were normal by histopathology. Allografts even with severe TCMR at day 21 retained viable epithelium, despite tubulitis and severe morphologic abnormalities (Figure 1). American Journal of Transplantation 2010; 10: 2241–2251

We defined the KT2.1 set of solute carrier transcripts selective for normal differentiated epithelium (described in the methods section) and compared their expression in mouse kidney isografts and allografts to nontransplanted kidneys with ATN induced by vascular clamping 7 days previously. Results for individual transcripts are shown in Table 1. Compared to control kidneys, average expression of KT2.1 solute carrier transcripts was transiently reduced in isografts at day 5 then recovered (day 5: 59% loss, day 7: 28% loss and day 21: 10% loss). KT2.1 expression was lost in allografts at day 5 (68%), 7 (75%) or day 21 (94%). Average transcript loss in ATN kidneys was 77%, similar to that in allografts at day 7 (Figure 2A). We analyzed expression changes of individual solute carrier transcripts in isografts, allografts and ATN kidneys. Loss of individual solute carrier transcripts in ATN kidneys correlated with loss in isografts (day 5: r = 0.858, p < 0.001; day 7: r = 0.755, p = 0.001, day 21 = 0.482, p = 0.013) and 2243

Einecke et al. Table 1: Solute carrier transcripts showing >90% reduction in severe rejecting allografts 21 days after transplantation compared to native CBA kidneys (‘KT2.1 transcript set’)1 Affymetrix probeset ID 1416966_at 1423279_at 1417072_at 1437755_at 1451460_a_at 1449301_at 1422899_at 1439519_at 1422856_at 1419166_at 1418923_at 1420379_at 1419496_at 1417809_at 1419117_at 1426082_a_at 1451239_a_at 1438332_at 1426595_at 1428595_at 1422897_at 1417280_at 1418118_at 1425038_at 1441236_at 1419725_at

Gene symbol Slc22a8 Slc34a1 Slc22a6 Slc5a12 Slc22a7 Slc7a13 Slc6a20 Slc34a3 Slc12a3 Slc5a2 Slc17a3 Slco1a1 Slco1a6 Slc22a18 Slc22a2 Slc16a4 Slc26a1 Slc22a6 Slc18a1 Slc6a19 Slc22a12 Slc17a1 Slc22a1 Slc22a19 Slc9a3 Slc26a4

Signal in control kidneys

Ratio of signal in D7 rejection versus controls

Ratio of signal in D21 rejection versus controls

Ratio of signal in ATN kidneys versus controls

4894 13898 4846 3475 3124 11245 4569 1568 3190 2840 11593 5206 2301 5545 3857 1032 693 422 2164 1644 2235 7972 4905 2507 1153 938

0.07 (93% loss) 0.35 (65% loss) 0.23 (67% loss) 0.23 (77% loss) 0.26 (74% loss) 0.23 (77% loss) 0.21 (79% loss) 0.08 (92% loss) 0.45 (55% loss) 0.14 (86% loss) 0.32 (68% loss) 0,16 (84% loss) 0.23 (77% loss) 0.23 (67% loss) 0.23 (77% loss) 0.20 (80% loss) 0.26 (74% loss) 0.21 (69% loss) 0.25 (75% loss) 0.31 (69% loss) 0.36 (64% loss) 0.42 (58% loss) 0.29 (71% loss) 0.29 (71% loss) 0,40 (60% loss) 0.47 (55% loss)

0.01 (99% loss) 0.02 (98% loss) 0.02 (98% loss) 0.02 (98% loss) 0.02 (98% loss) 0.03 (97% loss) 0.03 (97% loss) 0.03 (97% loss) 0.04 (96% loss) 0.04 (96% loss) 0.05 (95% loss) 0,05 (95% loss) 0.05 (95% loss) 0.06 (94% loss) 0.06 (94% loss) 0.06 (94% loss) 0.06 (94% loss) 0.06 (94% loss) 0.07 (93% loss) 0.07 (93% loss) 0.08 (92% loss) 0.09 (91% loss) 0.09 (91% loss) 0.09 (91% loss) 0,09 (91% loss) 0.09 (91% loss)

0.15 (85% loss) 0.55 (45% loss) 0.16 (84% loss) 0.43 (57% loss) 0.04 (96% loss) 0.15 (85% loss) 0.23 (77% loss) 0.13 (87% loss) 0.49 (51% loss) 0.44 (56% loss) 0.31 (69% loss) 0.14 (86% loss) 0.35 (65% loss) 0.23 (77% loss) 0.21 (79% loss) 0.17 (83% loss) 0.26 (74% loss) 0.15 (85% loss) 0.18 (82% loss) 0.14 (86% loss) 0.19 (81% loss) 0,52 (48% loss) 0.39 (61% loss) 0.06 (94% loss) 0,44 (56% loss) 0.46 (54% loss)

1 Transcripts

were sorted by the degree of loss in day 21 rejection. Numbers represent the average fold change (arithmetic mean) of each transcript in the experimental group compared to the controls.

in allografts (day 5: r = 0.520, p = 0.006; day 7: r = 0.501, p = 0.009; day 21: r = 0.483, p = 0.012) (Figure 2B). Although the degree of loss varied among individual KT2.1 solute carrier transcripts, all KT2.1 transcripts showed loss, independent of the underlying condition.

SLC6A19 displays reduced protein expression and altered cellular distribution in mouse isograft, allograft and ATN kidneys To assess the corresponding protein changes, we selected one KT2.1 solute carrier, SLC6A19, for immunostaining. SLC6A19 mediates sodium-dependent transport of a broad range of neutral amino acids from the lumen into the epithelial cells, and when mutated results in Hartnup disorder (22,23). We examined tissue sections of normal mouse kidneys, ATN kidneys, isografts and allografts at day 5, 7 and 21 posttransplant using immunofluorescence confocal microscopy (Figure 3A–C). In normal kidneys, SLC6A19 was expressed on the apical membrane of S1 segments and within the cytoplasm of S3 segments of proximal tubules (Figure 3A). Subcellular distribution of transporters was altered in isograft, allograft and ATN kidneys, with loss of apical expression and increased cytoplasmic expression compared to controls. Isografts at days 5, 7 and 21, showed 2244

reduced apical SLC6A19 expression compared to normal kidney (Figure 3B i–iii), while apical staining was lost completely in allograft kidneys at days 5, 7 and 21 (Figure 3B iv–vi). ATN kidneys were similar to isograft kidneys, with a reduced number of tubules expressing apical SLC6A19 and positive tubules were more frequent in the outer cortex and less so in the medullary region compared to controls (Figure 3C).

Analysis of human kidney transplant biopsies confirms similarity between TCMR and ATN We analyzed expression of solute carrier transcripts in human kidney transplant biopsies with TCMR and ATN, using normal renal tissue from nephrectomies as controls. Nontransplant kidneys with ATN were not available because they are rarely biopsied in this center. We selected transplant biopsies with histological diagnoses of TCMR (n = 49), borderline TCMR (n = 33) and ATN (n = 18) (Figure 4A). TCMR and ATN biopsies showed similar loss of solute carrier transcripts (TCMR: 36% loss, p = 0.018; ATN: 30% loss, p = 0.197). Average solute carrier expression in borderline TCMR was decreased compared to controls but did not reach significance (26% loss, p = 0.20). The extent of individual solute carrier transcript loss in ATN American Journal of Transplantation 2010; 10: 2241–2251


A

Isografts

Controls DAY 5

DAY 7

Allografts

DAY 21

DAY 5

DAY 7

ATN

DAY 21

DAY 7

KT2.1 score (log2)

0

*

-1.0

*

-2.0

**

-3.0

**

**

-4.0 -5.0

**

B

Expression in ATN kidneys (log2 of fold change over controls) -5.0

correlated strongly with the extent of loss in TCMR (r = 0.810, p < 0.001) (Figure 4B), as in the mouse studies.

Solute carrier expression in other disease states and correlation with histologic lesions The similarity between TCMR and ATN suggested that the loss was a nephron response to injury, not a consequence of particular mechanisms of injury. We analyzed KT2.1 solute carrier expression in all human kidney transplant biopsies available regardless of diagnosis (n = 234) to assess whether loss of solute carrier expression is stereotyped not only in TCMR and ATN but also across all other disease states. Mean KT2.1 expression in individual biopsies (n = 234) is shown in Figure 5A, ordered by decreasing expression. The corresponding histologic diagnosis for American Journal of Transplantation 2010; 10: 2241–2251

-4.0

-3.0

-2.0

-1.0

0.0 -1.0

-1.5

-2.0

-2.5

-3.0

-3.5

Spearman r = 0.501 p = 0.009

Expression in allografts day 7 (log2 of fold change over controls)

Figure 2: Expression of solute carrier transcripts in mouse allografts, isografts and ATN kidneys. (A) Solute carrier expression in mouse isografts, allografts and ATN. Compared to normal CBA mouse kidneys, solute carrier expression was significantly decreased in isografts at day 7 and 21, in allografts and in ATN kidneys. Expression of solute carrier transcripts in the gene set was summarized as geometric mean; the numbers shown represent the logarithm of mean ± SEM for replicates. The generation of the gene set is described in materials and methods. (∗ ) indicates significant differences between allografts or isografts compared to normal CBA kidneys by two-tailed t-test (∗ p < 0.05, ∗∗ p < 0.01). (B) Individual solute carrier expression in allografts day 21 and ATN. Each triangle represents an individual solute carrier transcript. Expression is shown as the logarithm of mean expression in day 21 allografts and ATN, respectively.

-4.0

each biopsy is shown at the top of the graph. Diagnostic categories are scattered across the graph, indicating that the level of solute carrier expression was unrelated to the diagnosis. To exclude the possibility that the mean solute carrier expression might miss subtle differences between the diagnostic categories, we used principal component analysis of individual KT2.1 solute carrier expression values in various diseases to capture heterogeneity within different diseases. Thus disease-specific changes in individual transcripts can be explored rather than summary measures, and the variance can be expressed in two dimensions: component 1 captured the largest change, and component 2 provided additional detail. Principal component analysis confirmed that loss of KT2.1 solute carrier expression was 2245

Einecke et al.

Figure 3: Immunofluorescence confocal microscopy of slc6a19 in normal, isograft, allogaft and ATN mouse kidneys. (A) Immunofluorescence of a normal kidney section showing expression of SLC6A19 (I, green), the marker of S1-S3 proximal tubule segments, LTA (ii, red) and the merge of images i and ii (iii) which demonstrates colocalization (yellow) of SLC6A19 and LTA in S1 segments of proximal tubules. Arrow indicates an example of SLC6A19 expression and colocalization with the marker on the apical membrane of an S1 segment adjacent to the glomerulus (g). (B) Immunofluorescence of SLC6A19 (red) in isograft (i–iii) and allograft (iv–vi) kidney sections at days 5, 7 and 21 posttransplant. Arrows indicate examples of apical SLC6A19 staining (arrow 1) in an isograft kidney at day 5 and cytoplasmic SLC6A19 staining (arrow 2) in allograft and isograft kidneys. Insets for i–vi represent a single tubule from each image at higher magnification. (C) Expression of SLC6A19 (red) in the ATN mouse model. Arrow indicates an example of apical SLC6A19 staining. (AC) Confocal images, scale bars represent 30 lm.

GFR at biopsy (p < 0.0001); i-score (p = 0.002); mm-score (p = 0.008); ct-score (p = 0.04).

similar in many disease categories: (Figure 5B), with a large overlap between the diagnostic groups. Even biopsies with the greatest loss of solute carrier expression (the right end of the graph) were distributed equally across different diagnostic categories.

Discussion

We also analyzed the relationship of KT2.1 solute carrier expression with histologic lesions (Table 2). No lesions were strongly correlated. The highest correlations were interstitial inflammation (i-score, r = −0.329), followed by tubular atrophy (ct-sore) and interstitial fibrosis (ci-score). The correlation with tubulitis was low (r = −0.169). Correlation of KT2.1 solute carrier expression with renal function (GFR estimated by Cockroft Gault formula) was better (r = 0.396) than with any histologic lesions. With stepwise multiple regression using all histology lesion scores, GFR, and time posttransplant as input variables and p = 0.1 as cut-off, the variables (and p values) that make the final model are:

We studied epithelial changes in mouse allografts with TCMR to those in isografts and in native kidneys with ATN, and extended these observations to human kidney biopsies with TCMR, ATN and other diseases. We used a set of transcripts exclusive to epithelium, KT2.1, as a probe for the epithelial response. The pattern and degree of KT2.1 solute carrier loss was similar in allografts with TCMR and native kidneys with ATN, and occurred transiently in isografts. To examine whether protein products were affected, we selected slc6a19 as a well-characterized member of the renal solute carrier family. Transcript loss was accompanied by absent apical immunostaining as early as

2246



A

Histologic diagnosis Controls

Borderline

TCMR

ATN

KT2.1 score (log2)

-0.0 -0.2 -0.4 -0.6 -0.8

* Expression in ATN biopsies (log2 of fold change over controls)

B -2.0

day 5 in TCMR. The loss of apical staining probably reflects more than simply loss of transcripts, since there was staining in the cytoplasm, compatible with loss of insertion of the proteins into the apical membrane, indicating that the loss of transcripts was part of a general response of the epithelium. Thus TCMR and ATN both induce KT2.1 solute carrier loss by triggering an active nephron injury–repair response, which develops rapidly and affects both protein and mRNA. Human kidney transplant biopsies with TCMR or ATN also showed loss of solute carrier transcripts. The loss of individual solute carrier transcripts correlated in ATN and TCMR in both mice and humans, and in humans occurred across a wide spectrum of diagnoses, supporting the concept of a stereotyped nephron response. In humans, loss of solute carrier expression correlated with the loss of GFR, and weakly with inflammation. We conclude that the loss of transcripts for solute carriers represents an inherently reversible nephron response that can be trigAmerican Journal of Transplantation 2010; 10: 2241–2251

-1.5

-1.0

-0.5

0.0

0.5

0.4 0.2 0.0 -0.2 -0.4 -0.6 -0.8 Spearman r = 0.810, p < 0.001

Expression in TCMR biopsies (log2 of fold change over controls)

Figure 4: Expression of solute carrier transcripts human T cell mediated rejection and ATN. (A) Solute carrier transcript expression in human biopsies with TCMR and in biopsies showing ATN. Expression of transcripts in the gene set was summarized as geometric mean; the numbers shown represent the logarithm of mean ± SEM for replicates. The generation of the gene set is described in materials and methods. (∗ ) indicates significant differences between TCMR or ATN compared to nephrectomies by two-tailed t-test (∗ p < 0.05, ∗∗ p < 0.01). (B) Individual solute carrier expression in TCMR and ATN biopsies. Each triangle represents an individual solute carrier transcript. Expression is shown as the logarithm of mean expression in TCMR and ATN, respectively.

-1.0

gered by many perturbations, presumably as an adaption that permits repair, given its apparent complete reversibility in mouse isografts. The KT2.1 transcripts are specific for fully differentiated renal epithelium, and lost when the epithelium dedifferentiates. We selected the KT2.1 solute carrier transcript set by high expression in normal mouse kidney and 90–99% loss in selected viable day 21 mouse allografts with severe tubulitis and epithelial deterioration. These transcripts must be specific for fully differentiated renal epithelium because the rejecting kidneys retain many other cell types, including many viable epithelial cells that have lost their typical features such as brush borders as well as vascular, matrix and inflammatory cells. Moreover many of the KT2.1s such as SLC6A19 are well characterized in terms of tissue expression. We attempted to study the solute carrier transcripts in renal epithelial cell lines but they 2247

Einecke et al.

were poorly expressed, indicating that the renal epithelial cells lines under in vitro conditions are not fully differentiated, and probably simulate injured epithelium rather than normal epithelium (unpublished results). Thus the KT2.1 solute carrier transcripts are expressed in fully differentiated epithelium and are rapidly lost when the epithelium is disturbed. The loss of the KT2.1 solute carrier transcripts represents temporary return to an earlier developmental state, presumably reflecting a programed response inherent to the nephron, probably also involving replication and apoptosis. Transcript loss is selective, and some transcripts are 99% lost in viable kidneys with TCMR or ATN despite persistence of epithelial tissue by histology. Moreover, many genes are increased in these kidneys: thus loss of KT2.1 solute carrier transcripts in allografts, isografts, and ATN is accompanied by expression of injury transcripts and cell cycle and apoptosis-related transcripts, and reexpression of embryonic markers, for example, Wnt/notch (21). Other studies focusing on native kidney diseases indicate that many types of injury and disease can trigger a similar de-differentiation response (24). However, defining the stereotyped response does not preclude the possibility of additional variations specific to particular diseases. This will need to be assessed and further validated in independent patient cohorts, although the similarity between

the mouse and the human data provides validation of our results. The consistent loss of solute carrier transcripts (and slc6a19 protein) despite differences in types of injury and in the function and localization of the individual transporters indicates an active injury–repair response of the entire nephron, rather than a response of individual cells. The nephron is like a worm—a functioning unit in which thousands of cells are regulated through vasomotion, filtered load, workload, physical forces, moving columns of fluid, chemical gradients (25), intercellular junctions and changes in the extracellular matrix. Thus the observed changes reflect a higher level of control initiating a stereotyped ‘shutdown’ mechanism of the entire nephron, triggering de-differentiation, loss of specialized features such as brush border membrane and loss of the ability to exclude inflammatory cells, resulting in tubulitis, accompanied by suspension of physiological functions. In addition to renal graft function, the loss of solute carrier transcripts could have an impact on other functions of the kidney, like drug metabolism and transport. Some unexplained correlations between graft function and drug exposure could be explained by solute carrier expression alterations. The molecular mechanisms that underlie the loss of solute carrier transporters remain largely unknown. Selective loss of renal parenchymal solute carrier expression reflects a

Figure 5: Solute carrier expression in human biopsies in relationship to histologic diagnosis. (A) Average solute carrier transcript expression is shown for each individual biopsy, ordered from highest to lowest expression along the x-axis. The corresponding diagnosis for each biopsy is shown in the panel above, allowing the distribution of diagnoses to be assessed according to the degree of loss of solute carrier expression. TCMR = T cellmediated rejection; borderline = borderline TCMR; ATN = acute tubular necrosis; ABMR = antibody mediated rejection; GN = glomerulonephritis; CNIT = calcineurin inhibitor toxicity; IFTA = interstitial fibrosis and tubular atrophy; others include BK virus nephropathy, transplant glomerulopathy and biopsies with minor change by histology. (B) The distribution of solute carrier expression (assessed by principal component analysis) within each diagnostic category is shown.

2248



Figure 5: Continued.

large-scale stereotyped and regulated process reflecting intrinsic regulatory events that prevent dangerous tubular dysfunction with loss of solutes that could injure the host. If nephrons do not absorb filtered electrolytes, the consequences can be serious, imposing the obligation for an over-riding control mechanism. This is implicit in the ‘intact nephron hypothesis’ (26), which specifies that nephrons in diseased kidney are either functioning relatively normally or are shutdown, implying single nephron regulation in disease. We suggest that organ dysfunction in TCMR (as well as ATN and other renal disorders) should be reconceptualized as an organized nephron response, reflecting widespread parenchymal de-differentiation rather than parenchymal cell death by necrosis or apoptosis. Loss of the molecAmerican Journal of Transplantation 2010; 10: 2241–2251

ular mediators of epithelial functions could be the physical explanation for at least some of the function loss in TCMR and many renal disorders. This is in line with our demonstration that TCMR does not require T cell cytotoxic mechanisms (27). The mechanisms by which T cells trigger the active nephron injury–repair response will ultimately have to be solved in vivo: modeling in vitro is limited by the difficulty of achieving full differentiation of renal epithelium. Products of T cells and macrophages may act indirectly on the epithelium, via changes in the extracellular matrix or the microcirculation, and probably involve both direct responses of the epithelium to local changes and secondary changes reflecting intrinsic nephron regulation. The general principle that functional impairment in diseases reflects de-differentiation may also be applicable to many organs and disease states in which it is difficult to 2249

Einecke et al. Table 2: Correlation of solute carrier expression with renal function and histologic lesions Parameter GFR at Bx Change in GFR within 6 months postbiopsy Glomerulitis (g score) Transplant glomerulopathy (cg score) Interstitial inflammation (i score) Interstitial fibrosis (ci score) Tubulitis (t score) Tubular atrophy (ct score) Arteritis (v score) Fibrous intimal thickening (cv score) Arteriolar hyalinosis (ah score) Mesangial Matrix (mm score) Peritubular capillaritis (ptc score) Peritubular capillary basement membrane multilayering

Spearman r

p value (two-tailed)

0.396 −0.0881

≤0.0001 0.1931

−0.172 −0.0417

0.0084 0.5256

−0.329 −0.197 −0.169 −0.211 −0.0919 0.0212

≤0.0001 0.0024 0.0097 0.0012 0.1611 0.7467

0.0827 0.0267 −0.111 0.0154

0.2076 0.6849 0.0903 0.8147

explain how small numbers of necrotic or apoptotic cells could explain profound changes in normal functions.

Acknowledgments The authors wish to thank Vido Ramassar, Anna Hutton and Kara Allanach for technical support. Funding source: Genome Canada, Genome Alberta, the University of Alberta, Capital Health Edmonton Area, the University Of Alberta Hospital Foundation, Roche Molecular Systems, Hoffmann-La Roche Canada Ltd., Alberta Innovation & Science, the Roche Organ Transplantation Research Foundation, the Kidney Foundation of Canada, Astellas Canada, Cell and Gene Trust and the Australian Research Council. Dr. Halloran also held a Canada Research Chair in Transplant Immunology and the Muttart Chair in Clinical Immunology.

Disclosure The authors of this manuscript have no conflicts of interest to disclose as described by the American Journal of Transplantation.

References 1. Solez K, Racusen LC, Marcussen N et al. Morphology of ischemic acute renal failure, normal function, and cyclosporine toxicity in cyclosporine-treated renal allograft recipients. Kidney Int 1993; 43: 1058–1067. 2. Solez K, Axelsen RA, Benediktsson H et al. International standardization of criteria for the histologic diagnosis of renal allograft rejection: The Banff working classification of kidney transplant pathology. Kidney Int 1993; 44: 411–422.

2250

3. Jabs WJ, Sedlmeyer A, Ramassar V et al. Heterogeneity in the evolution and mechanisms of the lesions of kidney allograft rejection in mice. Am J Transplant 2003; 3: 1501–1509. 4. Halloran PF, Urmson J, Ramassar V et al. Lesions of T-cellmediated kidney allograft rejection in mice do not require perforin or granzymes A and B. Am J Transplant 2004; 4: 705–712. 5. Einecke G, Fairhead T, Hidalgo LG et al. Tubulitis and epithelial cell alterations in mouse kidney transplant rejection are independent of CD103, perforin or granzymes A/B. Am J Transplant 2006; 6: 2109–2120. 6. Kayser D, Einecke G, Famulski KS et al. Donor Fas is not necessary for T-cell-mediated rejection of mouse kidney allografts. Am J Transplant 2008; 8: 2049–2055. 7. Einecke G, Broderick G, Sis B, Halloran PF. Early loss of renal transcripts in kidney allografts: Relationship to the development of histologic lesions and alloimmune effector mechanisms. Am J Transplant 2007; 7: 1121–1130. 8. Einecke G, Mengel M, Hidalgo LG, Allanach K, Famulski KS, Halloran PF. The early course of renal allograft rejection: Defining the time when rejection begins. Am J Transplant 2009; 9: 483– 493. 9. Mueller TF, Einecke G, Reeve J et al. Microarray analysis of rejection in human kidney transplants using pathogenesis-based transcript sets. Am J Transplant 2007; 7: 2712–2722. 10. Halloran PF, Miller LW, Urmson J et al. IFN-gamma alters the pathology of graft rejection: Protection from early necrosis. J Immunol 2001; 166: 7072–7081. 11. Einecke G, Melk A, Ramassar V et al. Expression of CTL associated transcripts precedes the development of tubulitis in T-cell mediated kidney graft rejection. Am J Transplant 2005; 5: 1827– 1836. 12. Goes N, Urmson J, Ramassar V, Halloran PF. Ischemic acute tubular necrosis induces an extensive local cytokine response: Evidence for induction of interferon-g, transforming growth factorb1, granulocyte-macrophage colony-stimulating factor, interleukin-2 and interleukin-10. Transplant 1995; 59: 565–572. 13. Takeuchi O, Sims TN, Takei Y, Ramassar V, Famulski KS, Halloran PF. Differential usage of class II transactivator promoters PI and PIV during inflammation and injury in kidney. J Am Soc Nephrology 2003; 11: 2823–2832. 14. Racusen LC, Solez K, Colvin RB et al. The Banff 97 working classification of renal allograft pathology. Kidney Int 1999; 55: 713–723. 15. Racusen LC, Colvin RB, Solez K et al. Antibody-mediated rejection criteria: An addition to the Banff 97 classification of renal allograft rejection. Am J Transplant 2003; 3: 708–714. 16. Solez K, Colvin RB, Racusen LC et al. Banff 07 classification of renal allograft pathology: Updates and future directions. Am J Transplant 2008; 8: 753–760. 17. Kowalczuk S, Broer A, Tietze N, Vanslambrouck JM, Rasko JE, Broer S. A protein complex in the brush-border membrane explains a Hartnup disorder allele. FASEB J 2008; 22: 2880– 2887. 18. Broer S, Bailey CG, Kowalczuk S et al. Iminoglycinuria and hyperglycinuria are discrete human phenotypes resulting from complex mutations in proline and glycine transporters. J Clin Invest 2008; 118: 3881–3892. 19. Solez K, Colvin RB, Racusen LC et al. Banff ‘05 Meeting report: Differential diagnosis of chronic allograft injury and elimination of chronic allograft nephropathy (‘CAN’). Am J Transplant 2007; 7: 518–526. 20. Irizarry RA, Hobbs B, Collin F et al. Exploration, normalization, and summaries of high density oligonucleotide array probe level data. Biostatistics 2003; 4: 249–264.


Epithelium in Allograft Rejection 21. Famulski KS, Broderick G, Hay K et al. Transcriptome analysis reveals heterogeneity in the injury response of kidney transplants. Am J Transplant 2007; 7: 2483–2495. 22. Seow HF, Broer S, Broer A et al. Hartnup disorder is caused by mutations in the gene encoding the neutral amino acid transporter SLC6A19. Nat Genet 2004; 36: 1003–1007. 23. Kleta R, Romeo E, Ristic Z et al. Mutations in SLC6A19, encoding B0AT1, cause Hartnup disorder. Nat Genet 2004; 36: 999–1002. 24. Ishibe S, Cantley LG. Epithelial-mesenchymal-epithelial cycling in kidney repair. Curr Opin Nephrol Hypertens 2008; 17: 379–385.


25. Schnermann J, Homer W. The juxtaglomerular apparatus: From anatomical peculiarity to physiological relevance. J Am Soc Nephrol 2003; 14: 1681–1694 (Smith Award lecture). 26. Bricker NS, Morrin PAF, Kime SW. The pathologic physiology of chronic brights disease: An exposition of the intact nephron hypothesis. Am J Med 1960; 28: 77–98. 27. Mengel M, Reeve J, Bunnag S et al. Molecular correlates of scarring in kidney transplants: The emergence of mast cell transcripts. Am J Transplant 2009; 9: 169–178.

2251