Clinical Immunology 229 (2021) 108792 Contents lists available at ScienceDirect Clinical Immunology journal homepage: www.elsevier.com/locate/yclim Full Length Article Advantages of plasmatic CXCL-10 as a prognostic and diagnostic biomarker for the risk of rejection and subclinical rejection in kidney transplantation Olga Millán a, b, Jordi Rovira c, d, Lluis Guirado e, Cristina Espinosa a, Klemens Budde f, Claudia Sommerer g, Gaston J. Piñeiro h, Fritz Diekmann c, d, h, Mercè Brunet a, b, * a Pharmacology and Toxicology Section, CDB, IDIBAPS, Hospital Clinic of Barcelona, University of Barcelona, c/Villarroel, 170, 08036 Barcelona, Spain b Biomedical Research Center in Hepatic and Digestive Diseases (CIBERehd), Instituto de Salud Carlos III, c/Sinesio Delgado 4, 28029 Madrid, Spain c Laboratori Experimental de Nefrologia i Trasplantament (LENIT), IDIBAPS, Barcelona, Spain d Red de Investigación Renal (REDINREN), Plaza de las Cortes, 11, 28014 Madrid, Spain e Renal Transplant Unit, Nephrology Department, Fundació Puigvert, Carrer de Cartagena, 340, 08025 Barcelona, Spain f Medizinische Klinik mit Schwerpunkt Nephrologie, Charité Universitätsmedizin Berlin, Campus Charité Mitte Luisenstraße 13, 10117 Berlin, Germany g Department of Nephrology, University of Heidelberg, University Hospital of Heidelberg and Mannheim, Im Neuenheimer Feld 672, 69120 Heidelberg, Germany h Department of Nephrology and Kidney Transplantation, ICNU, Hospital Clínic de Barcelona, c/Villarroel 170, 08036 Barcelona, Spain A R T I C L E I N F O A B S T R A C T Keywords: This study evaluate the potential of plasmatic CXCL-10 (pCXCL-10) as a pre&post transplantation prognostic and Kidney transplantation diagnostic biomarker of T-cell-mediated rejection (TCMR), antibody-mediated rejection (ABMR) and subclinical CXCL-10 rejection (SCR) risk in adult kidney recipients considering BKV and CMV infections as possible clinical Rejection (TCMR confounder factors. ABMR SCR) Twenty-eight of 100 patients included experienced rejection (TCMR:14; ABMR:14); 8 SCR; 13 and 16 were BKV diagnosed with BKV and CMV infection, respectively. Pre-transplantation pCXCL-10 was significantly increased CMV in TCMR and ABMR and post-transplantation in TCMR, ABMR and SCR compared with nonrejectors. All CMV+ Biological matrix patients showed pCXCL-10 levels above the cutoff values established for rejection whereas the 80% of BKV+ patients showed pCXCL-10 concentration < 100 pg/mL. pCXCL-10 could improve pre-transplantation patient stratification and immunosuppressive treatment selec­ tion according to rejection risk; and after kidney transplantation could be a potential early prognostic biomarker for rejection. Clinical confounding factor in BKV+ and particularly in CMV+ patients must be discarded. 1. Introduction rejection and/or infections and improved ISP treatment selection. In this sense, the concentration of circulating IFN-γ-inducible protein The identification of potential noninvasive prognostic biomarkers (CXCL-10) has been proposed to be a promising biomarker of short- and could be a critical step towards the efficient selection and adjustment of long-term kidney graft function [1–3]. CXCL-10 is a potent chemo­ immunosuppressive (ISP) treatment and improvements in clinical graft attractant for several immune cells, including CD4 and CD8 T cells, to outcome [1]. Knowledge of the individual immune status of patients, the sites of inflammation. Before and after transplantation, increased especially before transplantation, would enable patient stratification concentrations of CXCL-10 in kidney and liver recipients have been according to the risk of suffering a specific clinical event, such as associated with an increased risk of rejection [4–6]. With the current ISP Abbreviations: AR, acute rejection; AUC, area under the curve; ABMR, antibody-mediated rejection; BKV, polyoma virus; BPAR, biopsy-proven acute rejection; CI, confidence interval; CMV, cytomegalovirus; Cr, creatinine; CXCL10, chemokine interferon-inducible protein 10; dnDSA, de novo donor-specific antigens; EVR, everolimus; GFR, glomerular filtration rate; IFN-γ, interferon-gamma; ISP, immunosuppressive; MMF, mycophenolate mofetil; MPA, mycophenolic acid; NPV, negative predictive value; PPV, positive predictive value; ROC, receiver operating characteristic; SCR, subclinical rejection; Tac, tacrolimus; TCMR, T cell-mediated rejection. * Corresponding author at: Sección de Farmacología y Toxicología (CDB), Hospital Clínico de Barcelona, C/Villarroel 170, 08036 Barcelona, Spain. E-mail addresses: omillan@clinic.cat (O. Millán), jrovira1@clinic.cat (J. Rovira), LGuirado@fundacio-puigvert.es (L. Guirado), cespinosa@clinic.cat (C. Espinosa), klemens.budde@charite.de (K. Budde), Claudia.Sommerer@med.uni-heidelberg.de (C. Sommerer), fdiekman@clinic.cat (F. Diekmann), mbrunet@clinic.cat (M. Brunet). https://doi.org/10.1016/j.clim.2021.108792 Received 27 January 2021; Received in revised form 18 May 2021; Accepted 29 June 2021 Available online 2 July 2021 1521-6616/© 2021 The Authors. Published by Elsevier Inc. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). O. Millán et al. Clinical Immunology 229 (2021) 108792 regimens, the incidence of acute rejection (AR) after kidney trans­ for the diagnosis of rejection, which was always confirmed by histo­ plantation is approximately 10–15%. AR is diagnosed by an increase in logical evaluation of graft biopsies (biopsy-proven acute rejection, serum creatinine (Cr) or the development of new-onset proteinuria and BPAR). The biopsy samples were read by the local pathologist and is confirmed by biopsy, an invasive method that may favor some severe classified according to the current Banff classification [18,19]. Blood adverse events. Currently, there is a lack of biomarkers for subclinical and urine samples were collected on the day of the biopsy in the morning rejection (SCR) apart from surveillance biopsy [7] being this clinical and prior to the biopsy. event of great relevance in function and graft lost [8,9]. BKV infection was defined as detectable virus >1000 copies/mL Previous observational studies have demonstrated that increased [20,21], and CMV infection was defined as detectable CMV virus >1000 urinary CXCL-10 levels can identify patients with an ongoing rejection copies/mL [22]. episode several days before a biopsy is indicated by rising serum Cr levels. In a previous European multicenter study (EudraCT number: 2.2. Pharmacokinetic monitoring 2013–001817-33) on kidney transplant recipients [5] coordinated by our group, urinary CXCL-10 was identified as a prognostic and diag­ For Tac and mycophenolic acid (MPA), the C0 and completed AUC nostic biomarker of AR and graft outcome in the early post-transplant (0, 30 min, 1, 1.5, 2, 3, 4, 6, 8, and 12 h after drug administration) were period [6,7,10]. For the first time, Rabant et al. [10] identified the as­ measured in the 1st week post-transplantation, and the simplified AUCs sociation between urinary CXCL-10 levels and the diagnosis of ongoing (pre-dose, 2 h and 4 h post-dose) were determined in the 1st, 2nd, 3rd antibody-mediated rejection (ABMR), a relevant cause of long-term and 6th months post-transplantation. From the 1st month to the 6th kidney allograft failure [11]. In some transplant centers, de novo month post-transplantation, the Tac and MPA complete AUC values donor-specific antigen (dnDSA) monitoring has been implemented as a were estimated [23]. biomarker for the risk of ABMR, but its association is a subject of con­ Blood samples were collected in EDTA-K3 tubes. Whole-blood Tac troversy. High within-patient variability of Tac blood concentrations and EVR concentrations were measured by liquid chromatography/ was associated with an increase in the development rate of dnDSA, with tandem mass spectrometry, and MPA plasma concentrations were a consequential increase in the risk of graft loss [12,13]. Importantly, measured by high-performance liquid chromatography with an ultra­ CXCL-10 is also elevated in the urine of kidney recipients with poly­ violet detector, as previously reported [24]. omavirus (BKV) viremia. However, no association between cytomega­ AUC estimation methods and intrapatient variability in Tac-C0 were lovirus (CMV) viremia and urinary CXCL-10 levels has been described estimated by means of the coefficient of variation (CV) (for details, see [4,6,14]. Supplementary Methods). Regarding serum CXCL-10, there are controversial results in terms of its utility as a potential biomarker for assessing the risk of rejection 2.3. Plasmatic and urinary CXCL-10 measurements [15,16], mainly due to the presence of clinical confounding factors. A few studies reported that high pre-transplantation serum CXCL-10 For the analysis of plasmatic CXCL-10, whole blood was collected in concentrations were associated with long-term graft loss after kidney EDTA-anticoagulant tubes before the morning dose of treatment and transplantation [16]. More recently, Xu et al. [17] showed that serum centrifuged, within the first 2 h post-extraction, at 3000 rpm for 10 min, CXCL-10 levels measured on the 4th and 7th days after kidney trans­ and the plasma was stored at − 70 ◦ C for batch analysis. For the analysis plantation were significantly higher in patients with AR than in patients of urinary CXCL-10, first morning urine samples were collected from all without AR. patients post-transplantation and centrifuged, within the first 2 h post- The aim of this European multicenter study was to evaluate the po­ extraction, at 3000 rpm for 10 min, and the supernatant was stored at tential and clinical event specificity of the plasmatic CXCL-10 concen­ − 70 ◦ C. Both kinds of samples were shipped to the Pharmacology and tration measured before and after transplantation as a prognostic and Toxicology Laboratory (CDB), Hospital Clinic of Barcelona, for central­ diagnostic biomarker of the risk of TCMR, ABMR and SCR in a cohort of ized CXCL-10 analysis in a blinded fashion. No samples (urine and adult kidney transplant patients and to consider possible clinical con­ plasma) remained frozen for more than a month until analysis, and they founding factors related to BKV and CMV infections. This prospective were only thawed once. CXCL-10 concentrations were measured by study could enlighten the benefits of CXCL-10 measurements in plas­ ELISA (R&D Systems, Minneapolis, MN, USA) according to the manu­ matic samples pre- (reported here for the first time in kidney trans­ facturer’s instructions. Further technical details are described in the plantation) and post-transplantation. Supplementary Methods. 2. Methods 2.4. Statistical analyses 2.1. Study design and patients Demographic data and molecular analysis results were collected in a unified database. All of the analyses were performed using SPSS soft­ In this prospective observational multicenter study, 100 adult kidney ware, version 23.0 (SPSS Inc., Chicago, IL, USA). The samples were transplant patients were recruited from four European centers, two in adjusted to fit a nonparametric distribution. Statistically significant Spain and two in Germany. Participants were de novo kidney transplant differences between groups were assessed with the Mann-Whitney test, recipients from a deceased or living donor, with no other transplanted and correlations between variables were assessed by Spearman’s rho organs. Recipients older than 70 years and those positive for hepatitis B test. We used receiver operating characteristic (ROC) curve analysis to or C or human immunodeficiency virus were excluded. The study was define the optimal biomarker cutoffs for distinguishing patients with approved by the Ethics Committees of the participating centers, and all and without TCMR, ABMR or SCR, and BKV or CMV infection was based patients provided their written informed consent (EudraCT number: on ROC curves and was calculated with the best Youden index (sensi­ 2013–001817-33). tivity + specificity – 1) [25,26]. All of the data are presented as the Ninety percent of patients received ISP therapy consisting of tacro­ median ± standard deviation (SD). P ≤ 0.05 was considered to indicate limus (Tac), mycophenolate mofetil (MMF) and methylprednisolone or statistical significance. Further details of the statistical analysis are 10% Tac, everolimus (EVR) and methylprednisolone. Further details of described in the Supplementary Methods. ISP therapy are described in the Supplementary Methods. Patients had a total of six visits during the study: pre-transplantation and during the 1st week and the 1st, 2nd, 3rd, and 6th months after transplantation. All of the participating centers used the same criteria 2 O. Millán et al. Clinical Immunology 229 (2021) 108792 3. Results 3.3. Plasmatic CXCL-10 concentration as a prognostic biomarker of rejection (TCMR, ABMR & SCR) 3.1. Patients Pre-transplantation, there were significant differences in the plas­ The main characteristics of the 100 kidney recipients are shown in matic CXCL-10 concentration between nonrejectors (n = 42) and pa­ Table 1. During the post-transplantation follow-up, 14 patients experi­ tients who experienced TCMR (n = 14; P = 0.000) or ABMR (n = 14; P = enced TCMR, 14 ABMR and 8 were diagnosed with SCR. Of the patients 0.001) (Fig. 1 & Table 2B); however, no significant differences were who suffered TCMR, plasma and urine samples were obtained for CXCL- observed in CXCL-10 levels between those free of rejection and SCR 10 measurements. The diagnosis of rejection was based on clinical and patients (P = 0.084). laboratory findings and was confirmed by histological evaluation of After transplantation, patients who experienced TCMR (n = 14) and graft biopsies. Table 1 shows the time post-transplantation of the clinical ABMR (n = 14) and those diagnosed with SCR (n = 8) had significantly event. For all patients who rejected, plasma and urine were always higher plasmatic CXCL-10 levels than nonrejectors (clinical event-free n collected for biomarker analysis before the ISP regimen was modified to = 42) throughout the study period (P < 0.05) (Fig. 1 & Table 2B). resolve the rejection episode. Plasmatic CXCL-10 median concentrations were similar in patients with Thirteen patients were diagnosed with BKV infection and 16 with BKV infection (n = 13) and in BKV-free patients during the entire study CMV infection (Table 1S) [27]. Of the 13 BKV-infected patients, 2, 1 and period (Fig. 1 and Table 2B), but high inter-patient variability was 3 patients experienced TCMR, ABMR and SCR, respectively. Two of the observed in the BKV+ group since, in 3 patients, BKV+ showed plasmatic 16 CMV-infected patients also suffered for ABMR, but in no case were CXCL-10 levels >200 pg/mL, all of them at the 6th month post- the two events coincident in time. transplantation (Table 1S-A). We noted that all of the patients with rejection had plasmatic CXCL-10 values greater than the specific cutoff 3.2. Pharmacokinetics established for each type of rejection, whereas for BKV infection, the 80% of patients showed CXCL-10 concentration values less than 100 pg/ Table 2S summarizes the C0 values for Tac, EVR and MPA, the AUC0- mL before and at the time of infection. Furthermore, in no case were the 12h values and the AUC0-12h/Cmin ratios for Tac and MPA, and the Tac, two events (rejection and BKV infection) coincident in time. MPA, EVR and prednisone doses. During the 6 months post- In addition, patients with active CMV infection (n = 16) had signif­ transplantation, the Tac C0 was 7–10 ng/mL. No statistically signifi­ icantly higher plasmatic CXCL-10 concentrations than CMV-free pa­ cant differences in Tac, MPA or EVR C0 or prednisone dose were tients (Fig. 1, Table 2B, Table 1S-B). Details about the time evolution observed between rejectors (TCMR, ABMR or SCR), patients with CMV pre- and post-transplantation of plasmatic CXCL-10 production in each or BKV infection and nonrejectors during the evaluation period. rejector patient are described in Supplementary Data and Figures. Although the AUC0-12h values for Tac and MPA were similar among the analyzed groups, TCMR patients showed a clear tendency towards lower 3.4. Pre- and post-transplantation prognostic values of plasmatic CXCL- Tac AUC0-12h values than nonrejectors (median Tac AUC0-12h value from 10 concentration for the risk of rejection (TCMR, ABMR, and SCR) and the 1st week to the 6th month post-transplantation: nonrejectors 177.2 possible confounding factors ± 18.2 ngxh/mL vs. TCMR patients 133.04 ± 9.03 ngxh/mL); however, no differences were observed in the Tac AUC0-12h/C0 ratios. The results Pre-transplantation cutoff values for predicting TCMR were deter­ of the analysis of intrapatient variability in Tac C0 are described in the mined based on AUC analysis of the ROC curve for plasmatic CXCL-10 Supplementary Data. levels, which were significantly higher in patients who experienced rejection (TCMR). ROC curve analysis showed that the plasmatic CXCL- Table 1 Characteristics of 100 de novo kidney transplant recipients. Total N = Free CE N = TCMR N = ABMR N = SCR N = 8 BKV N = CMV N = p 100 42 14 14 13 16 value Recipient Sex (male) 65(65%) 23 (55%) 9 (64%) 10 (71%) 6 (75%) 11 (85%) 10 (63%) 0.426 Recipient Age (years) 51.5 ± 12.9 51.5 ± 13.7 57.0 ± 12.6 54.0 ± 13.6 55.0 ± 51.0 ± 8.9 45.0 ± 0.411 13.6 12.5 Donor Age (years) 56 ± 14.1 53.5 ± 12.1 59 ± 14.9 65.0 ± 12.7 63.5 ± 9.5 50.5 ± 9.8 47.0 ± 0.270 19.8 Time on Dialysis (months) 34 ± 55.6 34.2 ± 49.9 45.0 ± 41.9 22.0 ± 25.9 33.5 ± 35.5 ± 37.5 ± 0.241 31.1 21.9 12.6 Cold Ischemia Time (hours) 10.1 ± 7.9 10.3 ± 6.5 11.5 ± 7.1 10.8 ± 8.6 15.7 ± 13.1 ± 2.0 19.3 ± 3.4 0.147 10.9 Type of Donor Cadaveric- 55 (55%) 21 (50%) 11 (79%) 10 (71%) 5 (63%) 5 (38%) 7 (44%) 0.574 Donor Living-Donor 45 (45%) 21 (50%) 3 (21%) 4 (29%) 3 (37%) 8 (62%) 9 (56%) Immunosuppressive Regimen Tac + MMF+ 90 (90%) 42 (100%) 13 (93%) 8 (57%) 5 (63%) 11 (85%) 15 (94%) 0.065 PDN Tac + EVR + 10 (10%) 0 (0%) 1 (7%) 6 (43%) 3 (37%) 2 (15%) 1 (6%) PDN Induction Therapy Yes 94 (94% 36 (86%) 14 (100%) 14 (100%) 8 (100%) 13 (100%) 16 (100%) 0.156 No 6(6%) 6 (14%) 0 (0%) 0 (0%) 0 (0%) 0 (0%) 0 (0%) Time Post-transplantation of the 1st week 3 5 2 0 0 Clinical Event 1st month 2 2 0 1 0 2nd month 2 1 0 0 2 3rd month 2 0 2 3 4 6th month 5 6 4 9 10 Quantitative variables are presented as the median and standard deviation. Count variables are presented as the raw and intragroup relative frequency. Free CE: free clinical event; TCMR: T-cell mediated rejection; ABMR: Antibody-mediated rejection; SCR: subclinical rejection; BKV: BK polyoma virus; CMV: Cytomegalovirus; Tac: Tacrolimus; PDN: Prednisone; MMF: Mycophenolate Mofetil; EVR: Everolimus. Kruskal-Wallis test groups. A value of P ≤ 0.05 was considered significant. 3 O. Millán et al. Clinical Immunology 229 (2021) 108792 Fig. 1. Plasmatic CXCL-10 concentration and clinical events. Differences in the plasmatic CXCL-10 concentrations among pa­ tients free of clinical events (gray boxes), BKV+ patients (white boxes), CMV+ patients (dotted boxes) and patients diagnosed with SCR (left striped boxes), TCMR patients (hatched boxes) and ABMR patients (right striped boxes) during the first six months post- transplantation. Significant differences between groups were assessed with the Mann-Whitney test. A value of P < 0.05 was considered to indicate statistical significance: *patients free of clinical events vs. TCMR patients; †patients free of clinical events vs. ABMR patients; ‡patients free of clinical events vs. patients diagnosed with SCR and §patients free of clinical events vs. CMV+ patients). 10 concentration had an outstanding capacity to discriminate TCMR month on, which is when there was active CMV infection) and patients patients and nonrejectors (AUC = 0.920; 95% CI = 0.805–1.000) free of both rejection and infection (AUC = 0.882; 95% CI = (Fig. 2A & B). The optimal cutoff value for predicting TCMR risk was 0.801–0.961). The optimal cutoff value for predicting the risk of CMV 156.89 pg/mL, with 90% sensitivity, 80% specificity, 90% PPV and 95% infection based on AUC analysis of the ROC curve for plasmatic CXCL-10 NPV. levels was 142.1 pg/mL, with 70% sensitivity, 89% specificity, 78% ROC curve analysis showed that post-transplantation plasma CXCL- PPV, and 92% NPV (Fig. 5A & B). 10 levels also had an outstanding capacity to discriminate TCMR pa­ tients and nonrejectors: AUC = 0.944 (95% CI = 0.909–0.979), and the 3.5. High urinary CXCL-10 concentration in kidney recipients with TCMR optimal cutoff value for predicting TCMR was 177.7 pg/mL with 84% sensitivity, 92% specificity, 80% PPV and 100% NPV (Fig. 2A & C). Patients who experienced a TCMR episode (n = 14) had consistently When we evaluated the prognostic capacity of plasmatic CXCL-10 higher urinary CXCL-10 levels than nonrejectors (n = 42) throughout concentrations concerning ABMR risk, ROC curve analysis revealed an the study period (Fig. 6 & Table 2A). Patients with BKV infection (n = excellent capacity to discriminate ABMR patients and nonrejectors 13) had significantly higher urinary CXCL-10 concentrations than BKV- before and after transplantation: before transplantation AUC = 0.826; free patients during the entire study follow-up period (Fig. 6 & 95% CI = 0.713–0.939, optimal cutoff value 140.4 pg/mL, 71% sensi­ Table 2A). There was no significant difference in the urinary CXCL-10 tivity, 73% specificity, 82.3% PPV and 97.7% NPV (Fig. 3A & B); after concentration between patients with TCMR and patients with BKV transplantation AUC = 0.906; 95% CI = 0.874–0.938, optimal cutoff infection. However, urinary CXCL-10 concentrations were similar in value 184.7 pg/mL, 82% sensitivity, 84% specificity, 84% PPV and patients with CMV infection (n = 16) and in patients free of rejection and 97.7% NPV (Fig. 3A & C). CMV infection during the entire follow-up period (Fig. 6 & Table 2A). Finally, the ability of pre-transplantation plasmatic CXCL-10 levels Details about the time evolution pre- and post-transplantation of urinary to predict the risk of SCR was evaluated; despite the results showing a CXCL-10 production in each rejector patient are described in Supple­ tendency for higher levels in SCR patients than in nonrejectors, the ROC mentary Data and Figures. curve analysis showed a no ability to discriminate these groups (AUC = 0.693; 95% CI = 0.713–0.939) (Fig. 4A & B). However, additional ROC curve analysis showed that post-transplantation plasmatic CXCL-10 3.6. Post-transplantation prognostic values of urinary CXCL-10 levels had an outstanding capacity to discriminate patients diagnosed concentration for TCMR risk and possible confounding factors with SCR and nonrejectors: AUC = 0.936 (95% CI = 0.893–0.979), and an optimal cutoff value for the prognosis of SCR of 131.0 pg/mL with a We evaluated the capacity of the urinary CXCL-10 concentration to 91.7% sensitivity, 82.2% specificity, 88% PPV and 89% NPV (Fig. 4A & predict the risk of TCMR. ROC curve analysis showed that the urinary C). CXCL-10 levels had excellent capacity to discriminate TCMR patients We attempted to determine whether active CMV infection is a con­ and nonrejectors (AUC = 0.885; 95% CI = 0.825–0.944). The optimal founding factor for the clinical utility of plasmatic CXCL-10 concentra­ cutoff value for predicting TCMR based on the AUC analysis of the ROC tions as a prognostic and diagnostic biomarker of TCMR, ABMR and curve for urinary CXCL-10 levels was 87.85 pg/mL, with 87% sensi­ SCR. ROC curve analysis revealed that the plasmatic CXCL-10 level tivity, 85% specificity, 92% PPV, and 86% NPV (Fig. 7A & C). could discriminate patients with active CMV infection (from the 2nd We sought to determine whether active BKV infection is a con­ founding factor for the clinical utility of the urinary CXCL-10 4 O. Millán et al. Clinical Immunology 229 (2021) 108792 Table 2 Urinary and plasmatic CXCL-10 production. 2A FREE CE (n = 42) TCMR (n = 14) BKV (n = 13) CMV (n = 16) FREE CE vs TCMR FREE CE vs BKV FREE CE vs CMV st st st st st st 1 Week 1 Week 1 Week 1 Week 1 Week 1 Week 1 st Week uCXCL-10 37.1 ± 21.5 146.6 ± 48.4 146.5 ± 59.7 41.6 ± 29.9 P = 0.000 P = 0.000 P = 0.216 FREE CE TCMR BKV CMV FREE CE vs TCMR FREE CE vs BKV FREE CE vs CMV 1 st Month 1 st Month 1 st Month 1 st Month 1 st Month 1 st Month 1 st Month uCXCL-10 22.9 ± 17.7 152.9 ± 34.6 124.8 ± 103.5 32.6 ± 16.2 P = 0.000 P = 0.000 P = 0.164 FREE CE TCMR BKV CMV FREE CE vs TCMR FREE CE vs BKV FREE CE vs CMV 2 nd Month 2 nd Month 2 nd Month 2 nd Month 2 nd Month 2 nd Month 2 nd Month uCXCL-10 23.9 ± 21.9 151.2 ± 32.1 120.4 ± 47.9 37.2 ± 19.8 P = 0.000 P = 0.000 P = 0.156 FREE CE TCMR BKV CMV FREE CE vs TCMR FREE CE vs BKV FREE CE vs CMV 3rd Month 3rd Month 3rd Month 3rd Month 3rd Month 3rd Month 3rd Month uCXCL-10 29.5 ± 20.6 178.7 ± 28.8 119.9 ± 54.2 38.3 ± 19.9 P = 0.000 P = 0.000 P = 0.405 FREE CE TCMR BKV CMV FREE CE vs TCMR FREE CE vs BKV FREE CE vs CMV 6th Month 6th Month 6th Month 6th Month 6th Month 6th Month 6th Month uCXCL-10 24.8 ± 22.4 190.7 ± 31.5 139.2 ± 48.9 56.4 ± 26.3 P = 0.049 P = 0.000 P = 0.069 2B FREE CE (n TCMR (n ABMR (n SCR (n = BKV (n = CMV n = FREE CE vs FREE CE vs FREE CE vs FREE CE vs FREE CE vs = 42) = 14) = 14) 8) 13) 16) TCMR ABMR SCR BKV CMV PRE-TX PRE-TX PRE-TX PRE-TX PRE-TX PRE-TX PRE-TX PRE-TX PRE-TX PRE-TX PRE-TX pCXCL- 97.5 ± 39.8 201.6 ± 163.6 ± 127.5 ± 66.3 ± 35 94.8 ± P = 0.000 P = 0.001 P = 0.084 P = 0.401 P = 0.395 10 78.6 79.4 77.7 30.5 FREE CE TCMR ABMR SCR BKV CMV FREE CE vs FREE CE vs FREE CE vs FREE CE vs FREE CE vs TCMR ABMR SCR BKV CMV 1 st Week 1 st Week 1 st Week 1 st Week 1 st Week 1 st Week 1 st Week 1 st Week 1 st Week 1 st Week 1 st Week pCXCL- 93.3 ± 53.3 231.2 ± 264.8 ± 192.6 ± 71.2 ± 79.7 ± P = 0.000 P = 0.000 P = 0.044 P = 0.203 P = 0.564 10 90.6 85.1 26.1 20.4 53.6 FREE CE TCMR ABMR SCR BKV CMV FREE CE vs FREE CE vs FREE CE vs FREE CE vs FREE CE vs TCMR ABMR SCR BKV CMV 1 st Month 1 st Month 1 st Month 1 st Month 1 st Month 1 st Month 1 st Month 1 st Month 1 st Month 1 st Month 1 st Month pCXCL- 65.6 ± 47.3 254.6 ± 372.6 ± – 51.6 ± 97.9 ± P = 0.000 P = 0.000 P = 0.000 P = 0.311 P = 0.318 10 135.8 169.7 18.4 62.4 FREE CE TCMR ABMR SCR BKV CMV FREE CE vs FREE CE vs FREE CE vs FREE CE vs FREE CE vs TCMR ABMR SCR BKV CMV nd nd nd nd nd nd 2 Month 2 Month 2 Month 2 2 2 Month 2 nd Month 2 nd Month 2 nd Month 2 nd Month 2 nd Month Month Month pCXCL- 69.1 ± 47.4 182.8 ± 177.9 ± 165.4 ± 78.5 ± 245.9 ± P = 0.001 P = 0.032 P = 0.041 P = 0,844 P = 0.001 10 155.4 50.4 13.4 23.6 151.1 FREE CE TCMR ABMR SCR BKV CMV FREE CE vs FREE CE vs FREE CE vs FREE CE vs FREE CE vs TCMR ABMR SCR BKV CMV 3rd Month 3rd Month 3rd Month 3rd 3rd 3rd Month 3rd Month 3rd Month 3rd Month 3rd Month 3rd Month Month Month pCXCL- 95.8 ± 55.6 292.3 ± 231.2 ± 217.2 ± 69.6 ± 160.7 ± P = 0.001 P = 0.021 P = 0.028 P = 0.149 P = 0.004 10 187.9 88.4 63.6 32.2 88.9 FREE CE TCMR ABMR SCR BKV CMV FREE CE vs FREE CE vs FREE CE vs FREE CE vs FREE CE vs TCMR ABMR SCR BKV CMV 6th Month 6th Month 6th Month 6th 6th 6th Month 6th Month 6th Month 6th Month 6th Month 6th Month Month Month pCXCL- 89.1 ± 66.2 329.6 ± 261.1 ± 206.8 ± 69.4 ± 295.4 ± P = 0.000 P = 0.001 P = 0.002 P = 0.157 P = 0.000 10 145.1 179.5 78.9 40.8 162.5 Data are expressed as the median and standard deviation. Mann-Whitney test groups. A value of P < 0.05 was considered significant. Free CE: free clinical event; TCMR: T-cell mediated rejection; ABMR: Antibody-mediated rejection; SCR: subclinical rejection; BKV: BK polyoma virus; CMV: Cytomegalovirus. Data are expressed as the median and standard deviation. Mann-Whitney test groups. A value of P < 0.05 was considered significant. Free CE: free clinical event; TCMR: T-cell mediated rejection; ABMR: Antibody-mediated rejection; SCR: subclinical rejection; BKV: BK polyoma virus; CMV: Cytomegalovirus. concentration as a prognostic and diagnostic biomarker of TCMR. The month). Another novel contribution by our study is the finding that ROC curve analysis showed that the urinary CXCL-10 level had an plasmatic CXCL-10 is a biomarker that can be sued to identify patients at excellent capacity to discriminate patients with active BKV infection and high risk of SCR, thereby providing physicians with an opportunity for patients free of both rejection and infection (AUC = 0.955; 95% CI = early intervention to modify these risk factors and improve graft and 0.922–0.988). The optimal cutoff value for predicting the risk of BKV clinical outcomes. infection based on AUC analysis of the ROC curve for urinary CXCL-10 Our results showed that there were significant pre-transplantation levels was 79.7 pg/mL, with 80% sensitivity, 93% specificity, 77% differences in the plasmatic CXCL-10 concentrations between non­ PPV, and 91% NPV (Fig. 7B & D). rejectors and patients who experienced TCMR or ABMR. For both events, the NPV was >95%, and as a consequence, those patients with 4. Discussion CXCL-10 levels below the established cut-off could be candidates for moderate ISP therapy, thus minimizing the adverse effects associated To the best of our knowledge, this European multicenter study is the with treatment over immunosuppression. If this finding is confirmed in first to demonstrate that the plasmatic CXCL-10 concentration could be a larger populations, it could be very useful since patients would be useful early prognostic and diagnostic biomarker for different types of stratified before transplantation according to the risk of rejection (for rejection (TCMR and ABMR) before transplantation and during the both TCMR and ABMR), which could allow for individual tailoring of the immediate post-transplantation period (from the 1st week to the 6th immunosuppressive regimen. 5 O. Millán et al. Clinical Immunology 229 (2021) 108792 Fig. 2. Pre- and post-transplantation ROC curves: plasmatic CXCL-10 concentration vs. TCMR risk. (A) Differences in the plasmatic CXCL-10 level between patients free of clinical events (gray boxes), BKV+ patients (white boxes) and TCMR patients (hatched boxes) during the first six months post-transplantation. Significant differences between groups were assessed with the Mann-Whitney test. P < 0.05 was considered to indicate statistical significance: †patients free of clinical events vs. TCMR patients. ROC curve analysis for discriminating TCMR rejectors and nonrejectors based on plasmatic CXCL-10 levels before (B) and after transplantation (C). Before trans­ plantation AUC = 0.920, 95% CI 0.805–1.000, cutoff value of 156.89 pg/mL, 90% sensitivity; 80% specificity, PPV = 90% and NPV = 95%; and after transplantation AUC = 0.944, 95% CI 0.909–0.979, cutoff value of 177.70 pg/mL, 84% sensitivity; 92% specificity, PPV = 80% and NPV = 100%. A previous study showed a significant increase in the urinary CXCL- levels for assessing the SCR risk has been previously reported but that 10 concentration in patients with ABMR [6], but this has not been re­ only one study evaluated plasmatic CXCL-10 production as a biomarker ported to occur in plasma. A recent publication by Mülbacher et al. [28] of subclinical (micro)vascular inflammation and did not find a robust reported that the diagnosis of ABMR was associated with significantly association14. SCR detection is usually conducted via surveillance bi­ higher levels of urinary and serum CXCL-10 in donor-specific antibody- opsies, but at present, there are very few centers that carry out protocol positive stable long-term kidney transplant recipients. Our results biopsies; therefore, having a noninvasive biomarker for SCR prediction showed that the plasmatic CXCL-10 levels also have the capacity to would be very useful, taking into account that SCR impacts the long- distinguish ABMR patients from nonrejectors or BKV-infected patients, term allograft outcome because it can lead to chronic tubulointerstitial not only in the immediate post-transplantation period but also before damage, chronic graft dysfunction, chronic rejection and graft loss. transplantation. Similar to the study conducted by Rabant et al. [6], we Furthermore, identifying these patients by monitoring plasmatic CXCL- obtained a high-value NPV (>97%) using plasma as a biological matrix, 10 would allow them to be excluded from the minimization protocols of which would enable the accurate identification of patients with a immunosuppressive therapy and for the immunosuppression to be decreased risk of experiencing ABMR and consequently avoid the need adjusted accordingly. Several studies have shown that SCR impacts long- for biopsies in DSA-positive kidney recipients. The availability of non- term allograft outcomes because SCR can lead to chronic tubulointer­ invasive biomarkers, before transplantation, with predictive capacity stitial damage, chronic graft dysfunction, chronic rejection and graft loss for the risk of both TCMR and ABMR will allow both patient stratifica­ [29,30]. The prognostic capacity of urinary CXCL-10 levels for assessing tion according to the risk of rejection (TCMR/ABMR) and improved the SCR risk has been previously reported [31,32]. Regarding SCR and immunosuppressive treatment selection. potential clinical confounding factors, our study shows that early post- It is important to note that prognostic capacity of urinary CXCL-10 transplantation patients with SCR had significantly higher plasmatic 6 O. Millán et al. Clinical Immunology 229 (2021) 108792 Fig. 3. Pre- and post-transplantation ROC curves: plasmatic CXCL-10 concentration vs. ABMR risk. (A) Differences in the plasmatic CXCL-10 level between patients free of clinical events (gray boxes), BKV+ patients (white boxes) and ABMR patients (hatched boxes) during the first six months post-transplantation. Significant differences between groups were assessed with the Mann-Whitney test. A value of P < 0.05 was considered to indicate statistical significance: *patients free of clinical events vs. ABMR patients. ROC curve analysis for discriminating ABMR patients and nonrejectors based on plasmatic CXCL-10 levels before (B) and after transplantation (C). Before trans­ plantation AUC = 0.826, 95% CI 0.713–0.939, cutoff value of 140.40 pg/mL, 71% sensitivity, 73% specificity, PPV = 82.3% and NPV = 97.7%; and after trans­ plantation AUC = 0.906, 95% CI 0.874–0.938, cutoff value of 184.70 pg/mL, 82% sensitivity, 84% specificity, PPV = 84% and NPV = 97.7%. CXCL-10 concentrations than patients without rejection or with BKV [14,33,34] have evaluated serum CXCL-10 as a biomarker for BKV+ risk infection but significantly lower levels than patients with TCMR or of infection in kidney recipients. All of them showed high inter-patient ABMR. The AUC (=0.936), PPV (=88%) and NPV (=89%) indicated variability like we did, but despite this fact, in our cohort, plasmatic good discriminatory capacity for identifying patients at high risk of CXCL-10 concentrations were significantly different in BKV-infected developing SCR. SCR detection is usually conducted via surveillance patients compared with patients with TCMR, ABMR or SCR, and in no biopsies, but at present, there are very few centers that carry them out; cases were the two events (rejection and BKV infection) coincident in therefore, having a noninvasive biomarker for SCR prediction would be time. We noted that all of the patients with rejection had plasmatic very useful. Identifying these patients would allow early intervention in CXCL-10 values greater than the specific cutoff established for each type terms of personal immunosuppressive drug adjustments with the goal of of rejection, whereas for BKV infection, 80% of patients had CXCL-10 controlling this maintained alloreactivity and preventing graft damage. concentration values below 100 pg/mL before and at the time of the Nevertheless, plasmatic CXCL-10 concentrations were increased in infection. Considering that in all the studies [14,33,34] (including the patients with CMV infection to levels similar to those observed in re­ present study), the number of patients evaluated with BKV infection is jectors. These results are in agreement with those previously published low, clearly, all of these findings must be confirmed in multicenter, by Ho et al. [14], who reported that patients with normal histology and randomized studies. concomitant CMV infection had higher serum CXCL-10 levels than pa­ Our results confirm previous findings that the urinary CXCL-10 tients with normal histology and no infection. Therefore, when plasma is concentration may also be a useful prognostic and diagnostic used as a biological matrix, active CMV infection is a potential clinical biomarker of TCMR; however, BKV infection also increases the CXCL-10 confounding factor and consequently, CMV viral load must be deter­ concentration [4,35]. For this reason, when urine samples are used to mined to identify patients with a decreased risk of rejection and those evaluate the risk of TCMR, BKV viral load must be measured to identify with increased CXCL-10 levels due to CMV infection. Few studies patients with a decreased risk of rejection and those with increased 7 O. Millán et al. Clinical Immunology 229 (2021) 108792 Fig. 4. Pre- and Post-transplantation ROC curves: plasmatic CXCL-10 concentration vs. SCR risk. (A) Differences in the plasmatic CXCL-10 level between patients free of clinical events (gray boxes), BKV+ patients (white boxes) and patients with SCR (hatched boxes) during the first six months post-transplantation. Significant differences between groups were assessed with the Mann-Whitney test. P < 0.05 was considered to indicate statistical significance:*patients free of clinical events vs. SCR patients. ROC curve analysis for discriminating patients with SCR and nonrejectors based on plasmatic CXCL-10 levels was evaluated before (B) and after transplantation (C). Before transplantation, no differences between patients free of clinical events and patients with SCR were observed: AUC = 0.693 (95% CI 0.713–0.939). After transplantation: AUC = 0.936, 95% CI 0.893–0.979, cutoff value of 131.0 pg/mL, 91.7% sensitivity, 82.2% specificity, PPV = 88% and NPV = 89%. CXCL-10 levels due to BKV infection. However, the presence of CMV episode. Unlike the results of some previous studies [13], our results infection does not affect urinary CXCL-10 levels, which were similar in showed no association between within-patient variability in Tac C0 and infected patients and in patients free of both rejection and infection. the risk of TCMR. This disagreement may be partially explained by the Concerning the relationship between drug exposure and the risks of different post-transplantation periods in which this biomarker was TCMR, ABMR and SCR, our results showed that although the differences evaluated. However, in patients with a %CV > 30%, we observed a between the evaluated groups were not statistically significant, patients higher risk of ABMR, which is not associated with dnDSA development who experienced rejection (TCMR) showed a clear tendency towards since the 67% of ABMR rejectors did not develop dnDSA. In our cohort, lower Tac AUC0-12h values than nonrejectors. Regarding the influence of all of the ABMR patients showed plasmatic CXCL-10 concentrations drug exposure on CXCL-10 changes, no correlation between plasmatic or greater than the cutoff established for the risk of this clinical event. urinary CXCL-10 levels and Tac and MPA exposure (C0) was observed. Therefore, plasmatic CXCL-10 measurements seem to be more effective High intrapatient variability in Tac exposure was recently shown to in identifying those patients at risk for developing ABMR. Its imple­ be associated with a high risk of dnDSA development [10] and to be a mentation in clinical routine may provide a means for patient stratifi­ risk factor for late kidney transplant failure [36]. The negative impact of cation to better prevent ABMR. dnDSA on long-term outcome after kidney transplantation has been In our opinion, the analysis of plasmatic CXCL-10 as a biomarker for demonstrated in many studies [37,38]. A possible cause for high intra­ identifying kidney transplant patients at high risk of rejection, either patient variability Tac exposure is the nonadherence to ISP treatment, TCMR or ABMR, has some advantages over the use of urinary CXCL-10. which contributes to ABMR as an independent risk factor for dnDSA It is often difficult or impossible to obtain urine samples before trans­ development [39]. In previous studies, this intrapatient variability in plantation, but collecting a pre-transplantation plasma sample permits Tac exposure was usually evaluated from 4 to 12 months after trans­ both patient stratification according to rejection risk and improved ISP plantation [12,40]. For the first the time, our study evaluated Tac treatment selection. BKV and CMV infections are frequent complications exposure before and very close to the time of the TCMR or ABMR after kidney transplantation, and current guidelines recommend 8 O. Millán et al. Clinical Immunology 229 (2021) 108792 Fig. 5. Post-transplantation ROC curves: plasmatic CXCL-10 concentration vs. CMV risk. (A) Differences in the plasmatic CXCL-10 concentration between patients free of clinical events (gray boxes), BKV+ patients (white boxes) and CMV+ patients (dotted boxes) during the first six months post-transplantation. Significant differences between groups were assessed with the Mann-Whitney test. P < 0.05 was considered to indicate statistical significance: †patients free of clinical events vs. CMV+ patients. (B) Post-transplantation ROC curve analysis for discriminating CMV+ patients and nonrejectors based on plasmatic CXCL-10 concentration: AUC = 0.882, 95% CI 0.801–0.961, cutoff value of 142.1 pg/mL, 70% sensitivity, 89% specificity, PPV = 78% and NPV = 92%. Fig. 6. Urinary CXCL-10 concentration and clinical events. Differences in the urinary CXCL-10 concentration between patients free of clinical events (gray boxes), BKV+ patients (white boxes), CMV+ patients (dotted boxes) and TCMR patients (hatched boxes) during the first six months post-transplantation. Significant differ­ ences between groups were assessed with the Mann-Whitney test. P < 0.05 was considered to indicate statistical significance: *patients free of clinical events vs. TCMR patients; †patients free of clinical events vs. BKV+ patients. systematic quantitative monitoring of BKV and CMV-DNA in plasma. the specific cutoff established for each type of rejection. However, in The effects of BKV infection on the graft outcome may be more severe cases of CMV infection, all patients (100%) had plasmatic CXCL-10 than those of CMV infection. Therefore, it is very important to determine levels above the cutoff values established for rejection; hence, CMV is whether the elevated CXCL-10 levels are caused by rejection or reac­ a much more powerful confounding factor than BKV infection. Finally, tivation of BKV infection; in contrast to the case with urine, both clinical the comparison of the performance analysis of plasma and urine in events seem to be differentiated by the use of plasma, and only a small predicting the risk of rejection after transplantation revealed that the percentage of patients would present plasmatic CXCL-10 levels above plasma matrix seems to be more robust. The 100% NPV indicated that 9 O. Millán et al. Clinical Immunology 229 (2021) 108792 Fig. 7. Post-transplantation ROC curves: urinary CXCL-10 concentration vs. TCMR/BKV risk. ROC curve analysis for discriminating TCMR patients (A) and BKV+ patients (B) from nonrejectors based on urinary CXCL-10 concentration. The capacity of post- transplantation urinary CXCL-10 levels to discriminate TCMR patients and BKV+ patients from nonrejectors was reported as follows: (A) AUC = 0.885, 95% CI 0.825–0.944, cutoff value 87.85 pg/mL, 87% sensitivity, 85% specificity, PPV = 92% and NPV = 86%; (B) AUC = 0.955, 95% CI 0.922–0.988, cutoff value 79.70 pg/ mL, 80% sensitivity; 93% specificity, PPV = 77% and NPV = 91%. Differences in the urinary CXCL-10 concentration between patients free of clinical events (gray boxes), CMV+ patients (dotted boxes) and TCMR patients (hatched boxes) during the first six months post-transplantation (C). Significant differences between groups were assessed with the Mann-Whitney test. P < 0.05 was considered to indicate statistical significance: *patients free of clinical events vs. TCMR patients. Differences in the urinary CXCL-10 concentration between patients free of clinical events (gray boxes), CMV+ patients (dotted boxes) and BKV+ patients (white boxes) during the first six months post-transplantation (D). Significant differences between groups were assessed with the Mann-Whitney test. P < 0.05 was considered to indicate statistical significance: *patients free of clinical events vs. BKV+ patients. plasma matrix analysis is capable of accurately identifying patients with In summary, chemokines are normally expressed at low levels and decreased risk of experiencing TCMR, and as a consequence, patient rapidly up-regulated at the onset of the immune response. During the candidates for more moderate ISP therapy, thus minimizing the adverse dynamic process of the interaction between the graft and immune sys­ effects associated with treatment. tem, changes in CXCL-10 concentrations allow the identification of pa­ A possible explanation for the discrepancies in CXCL-10 levels in tients at high risk of TCMR, ABMR and SCR and enable early patients with BKV or CMV depending on biological matrix (plasma or intervention to modify these risks factors and provide personalized urine) is that the specific inflammatory response to these viruses occurs immunosuppressive treatment accordingly, thus improving the graft in different compartments. BKV infection targets tubular epithelial cells and patient outcomes. It is well known that CXCL-10 modulates with subsequent hematogenic spreading, thereby increasing urinary chemotaxis during several immune inflammatory processes and in­ CXCL-10 levels, whereas CMV infection can affect circulating hemato­ fluences cytokine secretion, thus regulating the immune response poietic cells and potentially vascular endothelial cells, leading to an against the implanted graft. An intrinsic limitation of the use of urinary increased plasmatic CXCL-10 concentration. CXCL-10 levels to noninvasively detect rejection is that the level of this Our study has some limitations, including the relatively low number chemokine is also increased in other inflammatory conditions. In addi­ of events; the evaluation of the prognostic capacity of both biological tion, as the results show that this biomarker seems to have predictive matrices (plasma and urine) for the risk of only TCMR; and the inclusion capacity for the assessment of the risk of TCMR and ABMR before of a Caucasian population only; our findings should be validated in other transplantation and as it is often difficult or impossible to obtain urine ethnic populations. samples pre-transplantation, in this situation, plasma samples have a 10 O. Millán et al. Clinical Immunology 229 (2021) 108792 clear advantage over urine samples, allowing both patient stratification [8] W. Hoffman, R. Mehta, D.R. Jorgensen, P. Sood, P. Randhawa, C.M. Wu, et al., The impact of early clinical and subclinical T cell-mediated rejection after kidney according to rejection risk and improved immunosuppressive treatment transplantation, Transplantation 103 (7) (2019) 1457–1467. selection. Furthermore, plasmatic CXCL-10 can be used to identify pa­ [9] D.N. Rush, I.W. Gibson, Subclinical inflammation in renal transplantation, tients at high risk of SCR. Transplantation 103 (6) (2019) e139–e145. [10] M. Rabant, L. Amrouche, X. Lebreton, F. Aulagnon, A. Benon, V. Sauvaget, et al., Urinary C-X-C motif chemokine 10 independently improves the noninvasive 5. Conclusions diagnosis of antibody-mediated kidney allograft rejection, J. Am. Soc. Nephrol. 26 (2015) 2840–2851. [11] S. Davis, J. Cooper, Acute antibody-mediated rejection in kidney transplant The results of this European prospective, observational multicenter recipients, Transplant. Rev. (Orlando) (2017) 47–54. study suggest that the plasmatic CXCL-10 concentration could be a po­ [12] E. Rodrigo, D. San Segundo, G. Fernández-Fresnedo, et al., Within-patient tential early, noninvasive prognostic biomarker for rejection (TCMR and variability in tacrolimus blood levels predicts kidney graft loss and donor-specific antibody development, Transplanataion 100 (2016) 2479–2485. ABMR) in kidney transplant recipients. In addition, plasmatic CXCL-10 [13] R. Sapir-Pichhadze, Y. Wang, O. Famure, Y. Li, J.S. Kim, Time-dependent is a good biomarker for identifying patients at high risk of SCR, variability in tacrolimus trough blood levels is a risk factor for late kidney providing an opportunity for physicians to initiate early interventions to transplant failure, Kidney Int. 85 (2014) 1404–1411. [14] J. Ho, S. Schaub, C. Wiebe, A. Gao, et al., Urinary CXCL10 chemokine is associated modify these risk factors and improve the graft and clinical outcomes. As with Alloimmune and virus compartment-specific renal allograft inflammation, this biomarker can be evaluated before transplantation, its imple­ Transplantation 102 (2018) 521–529. mentation in clinical practice would allow better patient stratification [15] S. Heidt, S. Shankar, A.S. Muthusamy, D. San Segundo, K.J. Wood, Pretransplant serum CXCL9 and CXCL10 levels fail to predict acute rejection in kidney transplant according to rejection risk and improved decision-making regarding ISP recipients receiving induction therapy, Transplantation 91 (8) (2011) e59–e61. treatment. There is a need for prospective data from large, independent [16] E. Lazzeri, M. Rotondi, B. Mazzinghi, L. Lasagni, A. Buonamano, A. Rosati, et al., cohorts of kidney transplant patients in randomized, multicenter clinical High CXCL10 expression in rejected kidneys and predictive role of Pretransplant serum CXCL10 for acute rejection and chronic allograft nephropathy, trials to refine the cutoff values identified in the present study and to Transplantation 79 (2005) 1215–1220. evaluate the clinical usefulness of this biomarker. [17] C.X. Xu, B.Y. Shi, Z.K. Jin, J.J. Hao, W.L. Duan, F. Han, et al., Multiple-biomarkers provide powerful prediction of early acute renal allograft rejection by combination of serum Fractalkine, IFN-γ and IP-10, Transpl. Immunol. 50 (2018) 68–74. Funding [18] C. Roufosse, N. Simmonds, M.C. Groningen, M. Haas, K.J. Henriksen, C. Horsfield, et al., A 2018 reference guide to the Banff classification of renal allograft The present study was partially supported by a grant from Fundación pathology, Transplantation 102 (2018) 1795–1814. [19] M. Haas, A. Loupy, C. Lefaucheur, C. Roufosse, D. Glotz, D. Seron, B.J. Nankivell, et Mutua Madrileña (AP153362014) and by TEVA Pharmaceuticals. al., The Banff 2017 kidney meeting report: revised diagnostic criteria for chronic active T cell-mediated rejection, antibody-mediated rejection, and prospects for integrative endpoints for next-generation clinical trials, Am. J. Transplant. 18 Disclosure (2018) 293–307. [20] M.R. Jorgenson, J.L. Descourouez, B. Lyu, et al., Management of BK viremia is The authors of this manuscript have no conflicts of interest to associated with a lower risk of subsequent cytomegalovirus infection in kidney transplant recipients, Clin. Transpl. 34 (3) (2020), e13798. disclose as described by the Clinical Immunology. [21] H.H. Hirsch, P.S. Randhawa, AST Infectious Diseases Community of Practice, BK polyomavirus in solid organ transplantation-Guidelines from the American Society of Transplantation Infectious Diseases Community of Practice, Clin. Transpl. 33 (9) Acknowledgments (2019), e13528. [22] J. Maertens, C. Cordonnier, P. Jaksch, et al., Maribavir for preemptive treatment of The authors thank the physicians and patients from the four cytomegalovirus reactivation, N. Engl. J. Med. 381 (12) (2019) 1136–1147. [23] M. Miura, S. Satoh, T. Niioka, H. Kagaya, M. Saito, M. Hayakari, et al., Limited participating centers: Hospital Clinic, Barcelona; Fundació Puigvert, sampling strategy for simultaneous estimation of the area under the concentration- Barcelona; Charité-Universitätsmedizin, Berlin; and University Hospital, time curve of tacrolimus and mycophenolic acid in adult renal transplant Heidelberg. recipients, Ther. Drug Monit. 30 (2008) 52–59. CIBERehd is funded by the Instituto de Salud Carlos III. [24] J.M. Grinyo, H. Ekberg, R.D. Mamelok, F. Oppenheimer, J. Sanchez-Plumed, M. A. Gentil, et al., The pharmacokinetics of mycophenolate mofetil in renal transplant recipients receiving standard-dose or low-dose cyclosporine, low-dose Appendix A. Supplementary data tacrolimus or low-dose sirolimus: the symphony pharmacokinetic substudy, Nephrol. Dial. Transplant. 24 (2009) 2269–2276. [25] J.A. Hanley, B.J. McNeil, The meaning and use of the area under a receiver Supplementary data to this article can be found online at https://doi. operating characteristic (ROC) curve, Radiology 143 (1982) 29–36. org/10.1016/j.clim.2021.108792. [26] W. Youden, Index for ratting diagnosis test, Cancer Biol. Ther. 3 (1950) 32–35. [27] O. Ljungmann, M. Boeckh, H.H. Hirsch, et al., Definitions of cytomegalovirus infection and disease in transplant patients for use in clinical trials, Clin. Infect. Dis. References 64 (2017) 87–91. [28] J. Mühlbacher, K. Doberer, N. Kozakowski, H. Regele, S. Camovic, S. Haindl, et al., Non-invasive chemokine detection: improved prediction of antibody-mediated [1] M. Brunet, M. Shipkova, T. van Gelder, E. Wieland, C. Sommerer, K. Budde, et al., rejection in donor-specific antibody-positive renal allograft recipients, Front. Med. Barcelona consensus on biomarker-based immunosuppressive drugs management 7 (2020) 114. in solid organ transplantation, Ther. Drug Monit. 38 (Suppl. 1) (2016) S1–20. [29] B.J. Nankivell, J.R. Chapman, The significance of subclinical rejection and the [2] H.S. Ciftci, T. Tefik, M.K. Savran, E. Demir, Y. Caliskan, Y.D. Ogret, et al., Urinary value of protocol biopsies, Am. J. Transplant. 6 (2006) 2006–2012. CXCL9 and CXCL10 levels and acute renal graft rejection, Int. J. Organ TX Med. 10 [30] F. Moreso, M. Carrera, M. Goma, M. Hueso, J. Sellares, J. Martorell, et al., Early (2) (2019) 53–62. subclinical rejection as a risk factor for late chronic humoral rejection, [3] P. Hirt-Minkowski, S.A. De Serres, J. Ho, Developing renal allograft surveillance Transplantation 93 (2012) 41–46. strategies - urinary biomarkers of cellular rejection, Can. J. Kidney Health Dis. 2 [31] J. Ho, D.N. Rush, M. Karpinski, L. Storsley, I.W. Gibson, J. Bestland, et al., (2015) 28. Validation of urinary CXCL10 as a marker of borderline, subclinical, and clinical [4] J.A. Jackson, E.J. Kim, B.J. Begley Cheeseman, T. Harden, S.D. Perez, et al., Tubulitis, Transplantation 92 (2011) 878–882. Urinary chemokines CXCL9 and CXCL10 are noninvasive markers of renal allograft [32] T.D. Blydt-Hansen, I.W. Gibson, A. Gao, B. Dufault, J. Ho, Elevated urinary rejection and BK viral infection, Am. J. Transplant. 11 (2011) 2228–2234. CXCL10-to-creatinine ratio is associated with subclinical and clinical rejection in [5] O. Millán, K. Budde, C. Sommerer, I. Aliart, O. Rissling, B. Bardaji, et al., Urinary pediatric renal transplantation, Transplantation 99 (2015) 797–804. miR-155-5p and CXCL10 as prognostic and predictive biomarkers of rejection, graft [33] T. Schachtner, M. Stein, A. Sefrin, N. Babel, P. Reinke, Inflammatory activation and outcome and treatment response in kidney transplantation, Br. J. Clin. Pharmacol. recovering BKV-specific immunity correlate with self-limited BKV replication after 83 (12) (2017) 2636–2650. renal transplantation, Transpl. Int. 27 (2014) 290–301. [6] M. Matz, J. Beyer, D. Wunsch, M.F. Mashreghi, M. Seiler, J. Pratschke, et al., Early [34] L. Weseslindtner, L. Hedman, Y. Wang, R. Strassl, I. Helanterä, S.W. Aberle, et al., post-transplant urinary IP-10 expression after kidney transplantation is predictive Longitudinal assessment of the CXCL10 blood and urine concentration in kidney of short- and long-term graft function, Kidney Int. 69 (2006) 1683–1690. transplant recipients with BK polyomavirus replication-a retrospective study, [7] M. Rabant, L. Amrouche, L. Morin, R. Bonifay, X. Lebreton, L. Aouni, et al., Early Transpl. Int. 33 (2020) 555–566. low urinary CXCL9 and CXCL10 might predict immunological quiescence in [35] C. Tinel, A. Devresse, A. Vermorel, V. Sauvaget, D. Marx, V. Avettand-Fenoel, et al., clinically and histologically stable kidney recipients, Am. J. Transplant. 16 (2016) Development and validation of an optimized integrative model using urinary 1868–1881. 11 O. Millán et al. Clinical Immunology 229 (2021) 108792 chemokines for noninvasive diagnosis of acute allograft rejection, Am. J. [38] C. Lefaucheur, D. Viglietti, M. Mangiola, A. Loupy, A. Zeevi, From humoral theory Transplant. 20 (12) (2020) 3462–3476. to performant risk stratification in kidney transplantation, J Immunol Res 2017 [36] T. Vanhove, T. Vermeulen, P. Annaert, E. Lerut, D.R.J. Kuypers, High Intrapatient (2017) 5201098. variability of tacrolimus concentrations predicts accelerated progression of chronic [39] J. Sellarés, D.G. de Freitas, M. Mengel, J. Reeve, G. Einecke, B. Sis, et al., histologic lesions in renal recipients, Am. J. Transplant. 16 (2016) 2954–2963. Understanding the causes of kidney transplant failure: the dominant role of [37] B.D. Tait, C. Süsal, H.M. Gebel, P.W. Nickerson, A.A. Zachary, F.H. Claas, et al., antibody-mediated rejection and nonadherence, Am. J. Transplant. 12 (2012) Consensus guidelines on the testing and clinical management issues associated 388–399. with HLA and non-HLA antibodies in transplantation, Transplantation. 95 (2013) [40] M. Brunet, T. van Gelder, A. Asberg, V. Haufroid, D.A. Hesselink, L. Langman, et 19–47. al., Therapeutic drug monitoring of tacrolimus-personalized therapy: second consensus report, Ther. Drug Monit. (2019) 261–307. 12