Macrophage density in early surveillance biopsies predicts future renal transplant function

Inflammation impairs renal allograft survival but is difficult to quantify by eye at low densities. Here we measured leukocyte abundance in early surveillance biopsies by digital image analysis to test for a role of chemokine receptor genotypes and analyze the predictive value of leukocyte subsets to allograft function. In six-week surveillance biopsies, T-cell (CD3), B-cell (CD20), macrophage (CD68), and dendritic cell (CD209) densities were assessed in whole slide scans. Renal cortical CD3, CD20, and CD68 were significantly higher in histologic rejection. The CCR2 V64I genotype was associated with lower CD3 and CD209 densities. Above-median CD68 density was significantly associated with lower combined patient and graft survival with a hazard ratio of 3.5 (95% confidence interval 1.1-11.0). Both CD20 and CD68 densities inversely correlated with estimated glomerular filtration rate (eGFR) four years after transplantation. Additionally, CD68 correlated with eGFR loss. Among histological measurements including a complete Banff classification, only CD68 density was a significant predictor of an eGFR under 30ml/min after four years (odds ratio 7.4, 1.8-31.0) and part of the best eGFR prediction set in a multivariable linear regression analysis of multiple clinical and pathologic parameters. In a second independent cohort, the original CD68 median maintained its discriminative power for survival and eGFR. Thus, digital high-resolution assessment of CD68+ leukocyte infiltration significantly improves prognostic value of early renal transplant biopsies. Inflammation impairs renal allograft survival but is difficult to quantify by eye at low densities. Here we measured leukocyte abundance in early surveillance biopsies by digital image analysis to test for a role of chemokine receptor genotypes and analyze the predictive value of leukocyte subsets to allograft function. In six-week surveillance biopsies, T-cell (CD3), B-cell (CD20), macrophage (CD68), and dendritic cell (CD209) densities were assessed in whole slide scans. Renal cortical CD3, CD20, and CD68 were significantly higher in histologic rejection. The CCR2 V64I genotype was associated with lower CD3 and CD209 densities. Above-median CD68 density was significantly associated with lower combined patient and graft survival with a hazard ratio of 3.5 (95% confidence interval 1.1-11.0). Both CD20 and CD68 densities inversely correlated with estimated glomerular filtration rate (eGFR) four years after transplantation. Additionally, CD68 correlated with eGFR loss. Among histological measurements including a complete Banff classification, only CD68 density was a significant predictor of an eGFR under 30ml/min after four years (odds ratio 7.4, 1.8-31.0) and part of the best eGFR prediction set in a multivariable linear regression analysis of multiple clinical and pathologic parameters. In a second independent cohort, the original CD68 median maintained its discriminative power for survival and eGFR. Thus, digital high-resolution assessment of CD68+ leukocyte infiltration significantly improves prognostic value of early renal transplant biopsies. Kidney transplantation remains the most effective currently available form of renal replacement therapy, but graft survival is limited.1Nankivell B.J. Kuypers D.R. Diagnosis and prevention of chronic kidney allograft loss.Lancet. 2011; 378: 1428-1437Abstract Full Text Full Text PDF PubMed Scopus (247) Google Scholar, 2Stegall M.D. Gaston R.S. Cosio F.G. et al.Through a glass darkly: seeking clarity in preventing late kidney transplant failure.J Am Soc Nephrol. 2015; 26: 20-29Crossref PubMed Scopus (95) Google Scholar Risk factors for graft deterioration include infiltration by inflammatory cells.3Solez K. Racusen L.C. Abdulkareem F. et al.Adhesion molecules and rejection of renal allografts.Kidney Int. 1997; 51: 1476-1480Abstract Full Text PDF PubMed Scopus (37) Google Scholar, 4Mengel M. Sis B. Haas M. et al.Banff 2011 Meeting report: new concepts in antibody-mediated rejection.Am J Transplant. 2012; 12: 563-570Crossref PubMed Scopus (337) Google Scholar, 5Haas M. Sis B. Racusen L.C. et al.Banff 2013 meeting report: inclusion of c4d-negative antibody-mediated rejection and antibody-associated arterial lesions.Am J Transplant. 2014; 14: 272-283Crossref PubMed Scopus (1062) Google Scholar However, in early 6-week surveillance biopsies, the Banff total inflammation score did not correlate with inflammatory gene mRNA expression, and only gene expression was associated with graft function,6Mengel M. Chang J. Kayser D. et al.The molecular phenotype of 6-week protocol biopsies from human renal allografts: reflections of prior injury but not future course.Am J Transplant. 2011; 11: 708-718Crossref PubMed Scopus (83) Google Scholar whereas in another study,7Garcia-Carro C, Dorje C, Asberg A, et al. Inflammation in early kidney allograft surveillance biopsies with and without associated tubulointerstitial chronic damage as a predictor of fibrosis progression and development of de novo donor specific antibodies [e-pub ahead of print]. Transplantation. http://dx.doi.org/10.1097/TP.0000000000001216. Accessed March 4, 2017.Google Scholar inflammation according to the Banff criteria in early biopsies was associated with donor-specific antibody development at 1 year. Furthermore, in a number of studies in older grafts, the total amount of renal allograft inflammation in surveillance biopsies was a significant predictor of deterioration and failure.8Cosio F.G. Grande J.P. Wadei H. et al.Predicting subsequent decline in kidney allograft function from early surveillance biopsies.Am J Transplant. 2005; 5: 2464-2472Crossref PubMed Scopus (263) Google Scholar, 9Mengel M. Reeve J. Bunnag S. et al.Scoring total inflammation is superior to the current Banff inflammation score in predicting outcome and the degree of molecular disturbance in renal allografts.Am J Transplant. 2009; 9: 1859-1867Crossref PubMed Scopus (127) Google Scholar, 10Modena B.D. Kurian S.M. Gaber L.W. et al.Gene expression in biopsies of acute rejection and interstitial fibrosis/tubular atrophy reveals highly shared mechanisms that correlate with worse long-term outcomes.Am J Transplant. 2016; 16: 1982-1998Crossref PubMed Scopus (64) Google Scholar, 11Moreso F. Seron D. O'Valle F. et al.Immunephenotype of glomerular and interstitial infiltrating cells in protocol renal allograft biopsies and histological diagnosis.Am J Transplant. 2007; 7: 2739-2747Crossref PubMed Scopus (31) Google Scholar, 12Park W.D. Griffin M.D. Cornell L.D. et al.Fibrosis with inflammation at one year predicts transplant functional decline.J Am Soc Nephrol. 2010; 21: 1987-1997Crossref PubMed Scopus (174) Google Scholar, 13Shishido S. Asanuma H. Nakai H. et al.The impact of repeated subclinical acute rejection on the progression of chronic allograft nephropathy.J Am Soc Nephrol. 2003; 14: 1046-1052Crossref PubMed Scopus (192) Google Scholar Whether a low level of histologic inflammation in early surveillance biopsy specimens predicts functional outcome has not been reported. Studies addressing leukocyte subtypes in nonrejecting renal allografts revealed increases in CD3+ T cells, CD20+ B cells, and CD68+ macrophages in inflammatory infiltrates.11Moreso F. Seron D. O'Valle F. et al.Immunephenotype of glomerular and interstitial infiltrating cells in protocol renal allograft biopsies and histological diagnosis.Am J Transplant. 2007; 7: 2739-2747Crossref PubMed Scopus (31) Google Scholar, 12Park W.D. Griffin M.D. Cornell L.D. et al.Fibrosis with inflammation at one year predicts transplant functional decline.J Am Soc Nephrol. 2010; 21: 1987-1997Crossref PubMed Scopus (174) Google Scholar Increased B-cell density was associated with fibrosis and decreased function in surveillance biopsies from the first year after transplantation.11Moreso F. Seron D. O'Valle F. et al.Immunephenotype of glomerular and interstitial infiltrating cells in protocol renal allograft biopsies and histological diagnosis.Am J Transplant. 2007; 7: 2739-2747Crossref PubMed Scopus (31) Google Scholar However, most B-cell accumulation seems to occur at later stages.14Colvin R.B. Hirohashi T. Farris A.B. et al.Emerging role of B cells in chronic allograft dysfunction.Kidney Int. Suppl. 2010 Dec; : S13-S17Abstract Full Text Full Text PDF PubMed Scopus (27) Google Scholar Macrophage marker CD68+ myeloid cells have been detected in renal allografts during rejection and associated with functional decline.15Bergler T. Jung B. Bourier F. et al.Infiltration of macrophages correlates with severity of allograft rejection and outcome in human kidney transplantation.PloS One. 2016; 11: e0156900Crossref Scopus (49) Google Scholar, 16Copin M.C. Noel C. Hazzan M. et al.Diagnostic and predictive value of an immunohistochemical profile in asymptomatic acute rejection of renal allografts.Transplant Immunol. 1995; 3: 229-239Crossref PubMed Scopus (32) Google Scholar, 17Rowshani A.T. Vereyken E.J. The role of macrophage lineage cells in kidney graft rejection and survival.Transplantation. 2012; 94: 309-318Crossref PubMed Scopus (38) Google Scholar, 18Seron D. Moreso F. Protocol biopsies in renal transplantation: prognostic value of structural monitoring.Kidney Int. 2007; 72: 690-697Abstract Full Text Full Text PDF PubMed Scopus (73) Google Scholar In surveillance biopsies performed 1 year after transplantation, CD68+ cells correlated with fibrosis and loss of function.19Toki D. Zhang W. Hor K.L. et al.The role of macrophages in the development of human renal allograft fibrosis in the first year after transplantation.Am J Transplant. 2014; 14: 2126-2136Crossref PubMed Scopus (83) Google Scholar A recent study in for-indication biopsies suggests that in grafts with high levels of inflammation and a positive total inflammation score, myeloid dendritic cells characterized by CD209 in addition to CD68 may be superior indicators of graft prognosis.20Batal I. De Serres S.A. Safa K. et al.Dendritic cells in kidney transplant biopsy samples are associated with t cell infiltration and poor allograft survival.J Am Soc Nephrol. 2015; 26: 3102-3113Crossref PubMed Scopus (26) Google Scholar There is no currently available report of a systematic assessment as to whether the amount of any of these cell types in early surveillance biopsies is relevant for prognosis. The kidney contains a complex array of phagocytes with macrophage and antigen-presenting cell characteristics.21Kurts C. Panzer U. Anders H.J. et al.The immune system and kidney disease: basic concepts and clinical implications.Nat Rev Immunol. 2013; 13: 738-753Crossref PubMed Scopus (442) Google Scholar, 22Woollard K.J. Pusey C.D. The heterogeneous mononuclear phagocyte system of the kidney.Kidney Int. 2014; 85: 1011-1014Abstract Full Text Full Text PDF PubMed Scopus (4) Google Scholar Their recruitment is largely directed by chemokines and their cognate receptors.23Mannon R.B. Macrophages: contributors to allograft dysfunction, repair, or innocent bystanders?.Curr Opin Organ Transplant. 2012; 17: 20-25Crossref PubMed Scopus (66) Google Scholar Data mainly from mouse models indicate that both CCR2 and CX3CR1 contribute to phagocyte accumulation in renal inflammation.24Anders H.J. Romagnani P. Mantovani A. Pathomechanisms: homeostatic chemokines in health, tissue regeneration, and progressive diseases.Trends Mol Med. 2014; 20: 154-165Abstract Full Text Full Text PDF PubMed Scopus (61) Google Scholar, 25Kitching A.R. Dendritic cells in progressive renal disease: some answers, many questions.Nephrol Dial Transplant. 2014; 29: 2185-2193Crossref PubMed Scopus (22) Google Scholar The CCR2 ligand CCL2 (MCP-1) was highly elevated during renal graft rejection and has been proposed as therapeutic target.2Stegall M.D. Gaston R.S. Cosio F.G. et al.Through a glass darkly: seeking clarity in preventing late kidney transplant failure.J Am Soc Nephrol. 2015; 26: 20-29Crossref PubMed Scopus (95) Google Scholar, 26Ho J. Rush D.N. Nickerson P.W. Urinary biomarkers of renal transplant outcome.Curr Opin Organ Transplant. 2015; 20: 476-481Crossref PubMed Scopus (16) Google Scholar, 27Ho J. Wiebe C. Rush D.N. et al.Increased urinary CCL2: Cr ratio at 6 months is associated with late renal allograft loss.Transplantation. 2013; 95: 595-602Crossref PubMed Scopus (25) Google Scholar, 28Kalnins A. Thomas M.N. Andrassy M. et al.Spiegelmer inhibition of MCP-1/CCR2–potential as an adjunct immunosuppressive therapy in transplantation.Scand J Immunol. 2015; 82: 102-109Crossref PubMed Scopus (7) Google Scholar Its V64I mutation has been associated with a positive,29Abdi R. Tran T.B. Sahagun-Ruiz A. et al.Chemokine receptor polymorphism and risk of acute rejection in human renal transplantation.J Am Soc Nephrol. 2002; 13: 754-758PubMed Google Scholar, 30Krichen H. Khazen D. Sfar I. et al.Genetic polymorphisms of inflammatory molecules in Tunisian kidney transplantation.Transplant Proc. 2011; 43: 433-436Abstract Full Text Full Text PDF PubMed Scopus (6) Google Scholar, 31Omrani M.D. Mokhtari M.R. Tagizadae A. et al.Association of CCR5-59029 A/G and CCR2-V64I variants with renal allograft survival.Iran J Immunol. 2008; 5: 201-206PubMed Google Scholar, 32Singh R. Kapoor R. Srivastava A. et al.Impact of chemokine receptor CCR2 and CCR5 gene polymorphism on allograft outcome in North Indian renal transplant recipients.Scand J Immunol. 2009; 69: 51-56Crossref PubMed Scopus (10) Google Scholar negative,33Firasat S. Raza A. Abid A. et al.The effect of chemokine receptor gene polymorphisms (CCR2V64I, CCR5-59029G>A and CCR5Delta32) on renal allograft survival in Pakistani transplant patients.Gene. 2012; 511: 314-319Crossref PubMed Scopus (4) Google Scholar or no change in outcome after renal transplantation.34Azmandian J. Mandegary A. Saber A. et al.Chemokine receptor 2-V64I and chemokine receptor 5-Delta32 polymorphisms and clinical risk factors of delayed graft function and acute rejection in kidney transplantation.Iran J Kidney Dis. 2012; 6: 56-62PubMed Google Scholar, 35Kruger B. Schroppel B. Ashkan R. et al.A Monocyte chemoattractant protein-1 (MCP-1) polymorphism and outcome after renal transplantation.J Am Soc Nephrol. 2002; 13: 2585-2589Crossref PubMed Scopus (92) Google Scholar CX3CR1 is expressed on renal phagocytes36Hoffmann U. Bergler T. Segerer S. et al.Impact of chemokine receptor CX3CR1 in human renal allograft rejection.Transplant Immunol. 2010; 23: 204-208Crossref PubMed Scopus (26) Google Scholar, 37Kassianos A.J. Wang X. Sampangi S. et al.Fractalkine-CX3CR1-dependent recruitment and retention of human CD1c+ myeloid dendritic cells by in vitro-activated proximal tubular epithelial cells.Kidney Int. 2015; 87: 1153-1163Abstract Full Text Full Text PDF PubMed Scopus (30) Google Scholar and upregulated during fibrosis and renal graft rejection.36Hoffmann U. Bergler T. Segerer S. et al.Impact of chemokine receptor CX3CR1 in human renal allograft rejection.Transplant Immunol. 2010; 23: 204-208Crossref PubMed Scopus (26) Google Scholar, 38Segerer S. Hughes E. Hudkins K.L. et al.Expression of the fractalkine receptor (CX3CR1) in human kidney diseases.Kidney Int. 2002; 62: 488-495Abstract Full Text Full Text PDF PubMed Scopus (85) Google Scholar Although small studies on CX3CR1 polymorphisms did not show differences in renal graft survival,29Abdi R. Tran T.B. Sahagun-Ruiz A. et al.Chemokine receptor polymorphism and risk of acute rejection in human renal transplantation.J Am Soc Nephrol. 2002; 13: 754-758PubMed Google Scholar, 39Dabrowska-Zamojcin E. Dziedziejko V. Safranow K. et al.Association between the CX3CR1 gene V249I polymorphism and delayed kidney allograft function.Transplant Immunol. 2015; 32: 172-174Crossref PubMed Scopus (9) Google Scholar a recent study indicated that the V249I mutation protects from acute kidney injury in sepsis.40Chousterman B.G. Boissonnas A. Poupel L. et al.Ly6Chigh monocytes protect against kidney damage during sepsis via a CX3CR1-dependent adhesion mechanism.J Am Soc Nephrol. 2016; 27: 792-803Crossref PubMed Scopus (46) Google Scholar Functional studies demonstrate enhanced adhesiveness for the V249I mutation.40Chousterman B.G. Boissonnas A. Poupel L. et al.Ly6Chigh monocytes protect against kidney damage during sepsis via a CX3CR1-dependent adhesion mechanism.J Am Soc Nephrol. 2016; 27: 792-803Crossref PubMed Scopus (46) Google Scholar, 41Daoudi M. Lavergne E. Garin A. et al.Enhanced adhesive capacities of the naturally occurring Ile249-Met280 variant of the chemokine receptor CX3CR1.J Biol Chem. 2004; 279: 19649-19657Crossref PubMed Scopus (75) Google Scholar The role of CCR2 and CX3CR1 genotypes in renal graft leukocyte accumulation has not been evaluated. Manual assessment of cellular infiltrates in renal allograft biopsies is time-consuming, and quantification is difficult at low cell densities, increasing the probability of inter- and intraobserver variation.42Veronese F.V. Manfro R.C. Roman F.R. et al.Reproducibility of the Banff classification in subclinical kidney transplant rejection.Clin Transplant. 2005; 19: 518-521Crossref PubMed Scopus (57) Google Scholar, 43Farris A.B. Chan S. Climenhaga J. et al.Banff fibrosis study: multicenter visual assessment and computerized analysis of interstitial fibrosis in kidney biopsies.Am J Transplant. 2014; 14: 897-907Crossref PubMed Scopus (52) Google Scholar For research purposes, high-resolution determination of the inflammatory infiltrate (e.g., by gene array analysis) may be preferable. However, less costly and more widely available technology would be helpful in clinical practice. A number of digital imaging tools have been introduced into pathology.44Eccher A. Neil D. Ciangherotti A. et al.Digital reporting of whole-slide images is safe and suitable for assessing organ quality in preimplantation renal biopsies.Hum Pathol. 2016; 47: 115-120Abstract Full Text Full Text PDF PubMed Scopus (24) Google Scholar These require validation for each specific setting, not only in relation to current histopathologic standards, but also clinically meaningful outcomes.45Laurinavicius A. Plancoulaine B. Herlin P. et al.Comprehensive immunohistochemistry: digital, analytical and integrated.Pathobiology. 2016; 83: 156-163Crossref PubMed Scopus (9) Google Scholar, 46Shirinifard A. Thiagarajan S. Vogel P. et al.Detection of phenotypic alterations using high-content analysis of whole-slide images.J Histochem Cytochem. 2016; 64: 301-310Crossref PubMed Scopus (7) Google Scholar, 47Rosenberg A.Z. Palmer M. Merlino L. et al.The application of digital pathology to improve accuracy in glomerular enumeration in renal biopsies.PloS One. 2016; 11: e0156441Google Scholar Such studies on leukocyte densities in the renal allograft are not available so far. The aim of this project was to assess digital quantification of T-cell, B-cell, macrophage, and dendritic cell infiltration in early surveillance biopsy samples of renal allografts compared with conventional diagnosis of inflammation according to the Banff criteria. We investigated the role of common human chemokine receptor polymorphisms in their accumulation and the relevance of cell densities for graft functional development. Clinical characteristics of the 67 transplant recipients are given in Table 1. According to the revised Banff criteria, a rejection was present in 16% of grafts, mostly Banff category 3 borderline rejections (Supplementary Table S1). Whole-slide scans were annotated for renal cortex, medulla, and extrarenal tissues (Supplementary Figure S1). T-cell (CD3), B-cell (CD20), macrophage (CD68), and dendritic cell (CD209 = DC-SIGN) marker densities were quantified in whole specimens (methods, Supplementary Figure S1). We first analyzed whether leukocyte densities were associated with rejection. Indeed, renal cortical and medullary CD3, CD20, and CD68 positive infiltrates were significantly greater in grafts with rejection (Figure 1, Table 2), whereas CD209 showed a nonsignificant trend toward a decrease.Table 1Characterization of the patient cohortRecipient characteristics Age, yr51.9 ± 1.8 Sex26% (23) female Caucasian ethnicity99% (66) BMI25.2 ± 0.5 Dialysis before transplantation91% (60) Panel reactive antibodies at transplantation10% (mean reactivity to 35% of panel) Waiting time871 ± 105 daysUnderlying renal disease ADPKD27% (18) Hypertensive nephropathy12% (8) Diabetic nephropathy9% (6) Glomerulonephritis25% (17) Miscellaneous12% (8) Unknown15% (10)Graft characteristics Deceased donor67% (45) Cold ischemia time, h9.5 ± 0.9 HLA mismatch (A,B,DR)2.5 ± 0.2Immunosuppression at first outpatient visit Baseline steroid dose11.3 ± 0.8 mg/d CNI (tacrolimus and cyclosporine A)99% (66) Mycophenolate100% (67) mTOR inhibitor3% (2)Graft function at first outpatient visit Serum creatinine, μmol/l174 ± 10 Serum cystatin C, mg/l1.9 ± 0.1 GFR CKD-EPI (creatinine), ml/min42 ± 2 GFR CKD-EPI (cystatin C), ml/min39 ± 2 Proteinuria, g/l0.1 ± 0.02ADPKD, autosomal dominant polycystic kidney disease; BMI, body mass index; CKD-EPI, Chronic Kidney Disease Epidemiology Collaboration; CNI, calcineurin inhibitor; GFR, glomerular filtration rate; HLA, human leukocyte antigen; mTOR, mechanistic target of rapamycin.Values shown are % (N) or mean ± SEM. Open table in a new tab Table 2Univariate pathologic correlations of renal cortical leukocyte densitiesVariableCD3CD20CD68CD209Any rejection0.57 (<0.0001)0.56 (<0.0001)0.51 (<0.0001)Normal (Banff cat. 1)−0.24 (0.049)Antibody-mediated rejection (Banff cat. 2)Cellular borderline rejection (Banff cat. 3)0.48 (<0.0001)0.49 (<0.0001)0.24 (0.0004)T cell–mediated rejection (Banff cat. 4)0.27 (0.029)0.28 (0.02)Interstitial fibrosis and tubular atrophy (Banff cat. 5)0.26 (0.032)0.25 (0.025)0.34 (0.005)Other changes (Banff cat. 6)−0.39 (0.001)−0.33 (0.061)−0.33 (0.006)Interstitial infiltration score0.55 (<0.0001)0.57 (<0.0001)0.52 (<0.0001)Tubulitis score0.55 (<0.0001)0.57 (<0.0001)0.53 (<0.0001)Total inflammation score0.56 (<0.0001)0.60 (<0.0001)0.53 (<0.0001)C4d score−0.32 (0.009)−0.26 (0.034)Cortical fibrosis score0.31 (0.012)0.31 (0.010)0.38 (0.002)Tubular damage0.24 (0.049)CD3NACD200.58 (<0.0001)NACD680.51 (<0.0001)0.41 (0.0006)NACD2090.28 (0.024)−0.28 (0.02)NAcat., category; NA, not applicable.All significant correlations are given as Spearman’s r (P value). No significant correlations were observed for antibody-mediated rejection (Banff category 2) or glomerulitis and peritubular capillaritis scores. Open table in a new tab ADPKD, autosomal dominant polycystic kidney disease; BMI, body mass index; CKD-EPI, Chronic Kidney Disease Epidemiology Collaboration; CNI, calcineurin inhibitor; GFR, glomerular filtration rate; HLA, human leukocyte antigen; mTOR, mechanistic target of rapamycin. Values shown are % (N) or mean ± SEM. cat., category; NA, not applicable. All significant correlations are given as Spearman’s r (P value). No significant correlations were observed for antibody-mediated rejection (Banff category 2) or glomerulitis and peritubular capillaritis scores. CD3, CD20, and CD68 densities strongly positively correlated with each other (Table 2). In an analysis of rejecting and nonrejecting grafts separately, the association between T- (CD3) and B- (CD20) cell densities was significant for each group (Supplementary Figure S2). In nonrejecting grafts, positive associations of CD209 with CD20 and CD68 with CD3 and a negative association between CD68 and CD209 reached the significance level (Supplementary Figure S2). These data demonstrate that a marked increase of B-cell, T-cell and macrophage density in renal allograft rejection can be captured by computer-assisted histologic analysis. The role of human chemokine receptors CCR2 and CX3CR1 in renal allograft leukocyte accumulation was studied by comparing marker densities between the genotypes (Figure 2). All genotypes were in Hardy-Weinberg equilibrium and not significantly different from those of local healthy blood donors (Supplementary Table S2). The chemokine receptor CCR2 V64I allele in heterozygous and homozygous form was associated with significantly lower CD3+ T-cell and CD209+ dendritic–cell densities in the renal cortex (Figure 2a and d). Further analysis of grafts according to the CCR2 genotype revealed no difference in the frequency of rejection, and renal function did not differ significantly. There was no significant association of CX3CR1 genotype with any of the investigated leukocytes (Figure 2e–l). This suggests a limited role of human chemokine receptor genotypes in leukocyte accumulation in the kidney allograft. To test for relevance of the measured leukocyte markers for kidney graft function, we analyzed patient and graft survival in groups defined by high- and low-cell densities at the 6-week surveillance biopsy (i.e., above and below the median of each marker) (Figure 3). During the observation period, 4 patients recommenced renal replacement therapy, and 8 patients died with functioning grafts. High CD68 macrophage marker density in the 6-week surveillance biopsy was significantly associated with a higher rate of adverse events (Figure 3c), whereas a high CD209 density was associated with fewer events with borderline significance (Figure 3d). High CD3, CD20, CD68, and low CD209 densities were found in all patients with death-censored graft loss (Supplementary Figure S3). We further investigated the clinical course of the grafts in relation to leukocyte densities. Regarding the time before biopsy, delayed graft function significantly correlated with CD3, CD20, and CD68 density, whereas only CD68 also correlated with its duration (Supplementary Table S3). Within 4 years after transplantation, the total number of rejections significantly correlated with CD3 and CD68 densities, and there was a significant association of early CD3 with a total number of borderline and CD20 with total number of cellular rejections (Supplementary Table S4). To assess the impact of leukocyte markers on graft function, it was monitored by the estimated glomerular filtration rate (eGFR) according to the CKD-EPI (Chronic Kidney Disease Epidemiology Collaboration) formulas for creatinine and cystatin C.48Inker L.A. Schmid C.H. Tighiouart H. et al.Estimating glomerular filtration rate from serum creatinine and cystatin C.N Engl J Med. 2012; 367: 20-29Crossref PubMed Scopus (2486) Google Scholar Both CD20 and CD68 densities at week 6 significantly correlated with renal function in the surviving grafts after 4 years (Figure 4b,c,f,g). Similar results were obtained with serum creatinine, cystatin C, and urea concentrations (Supplementary Figure S4). There was no significant association of proteinuria with either leukocyte marker (Supplementary Figure S4). The association with excretory renal function was reflected by significantly lower eGFR over the whole observation period in patients with high renal cortical densities of either CD20 or CD68 (Figure 5a,b).Figure 5Association of CD20 and CD68 densities with estimated glomerular filtration rate (eGFR) development over time. Cystatin-based eGFR is given separately for the CD20 B-cell marker (a) and CD68 (b) macrophage marker densities above and below the median. Change of eGFR in relation to the initial outpatient presentation is given for high and low CD68 cortical densities separately (c) (2-way analysis of variance; ***P < 0.001, the number of patients [pts] at risk is indicated for each time point).View Large Image Figure ViewerDownload Hi-res image Download (PPT) In addition to lower absolute eGFR values, patients with a high CD68 density also lost significantly more eGFR than patients with low densities (Figure 5c). This association, which was unrelated to immunosuppressant use during follow-up (Supplementary Table S5), was not observed with any of the other studied leukocyte markers or pathologic scores. Given these correlations, we further characterized renal graft macrophage phenotype with regard to localization and polarization markers. CD68+ cells were localized in the cortical interstitium of all patients. Among other locations, tubular CD68 was more common in patients with borderline or cellular rejections or interstitial fibrosis according to the Banff criteria (Supplementary Table S6). With regard to phenotype and cortical densities of the M1 marker HLA-DR and M2 marker CD206,49Cao Q. Harris D.C. Wang Y. Macrophages in kidney injury, inflammation, and fibrosis.Physiology (Bethesda). 2015; 30: 183-194Crossref PubMed Scopus (210) Google Scholar, 50Murray P.J. Wynn T.A. Protective and pathogenic functions of macrophage subsets.Nat Rev Immunol. 2011; 11: 723-737Crossref PubMed Scopus (3320) Google Scholar both increased significantly with rejection. Additionally, the M2 marker CD206 significantly correlated with interstitial fibrosis (Banff category 5, Supplementary Table S7). Neither subtype marker correlated with previous delayed graft function or total or death-censored graft survival (data not shown). These data depict long-term functional implications of the graft CD68 and, to some degree, also CD20 cell densities 6 weeks after transplantation. Univariate regression analysis of multiple clinical and pathologic parameters with eGFR and eGFR loss 4 years after transplantation resulted in a number of significant variables including CD20, CD68, and CD206 densities (Table 3). Combinations of CD68 with macrophage subtype markers or CD209 were weaker than CD68 alone for long-term prediction of eGFR (Supplementary Table S7, data not shown).Table 3Univariate analysis of clinical parameters and pathologic findings with eGFR