Glomerular Infiltration by CXCR3+ ICOS+ Activated T Cells in Chronic Allograft Nephropathy with Transplant Glomerulopathy


* Enver Akalin,


The pathogeneses of chronic allograft nephropathy (CAN), a leading cause of allograft failure, and one of its complications, transplant glomerulopathy (TGP), are unknown. Immunohistologic analysis of human renal transplant biopsies showed expression of inducible costimulator (ICOS), the chemokine receptor CXCR3, and its ligands, Mig and IP-10, by intraglomerular and periglomerular leukocytes in biopsies with CAN and TGP but not CAN alone. ICOS and CXCR3 are both characteristics of activated, effector T cells, suggesting different pathogenetic mechanisms underlying TGP vs. CAN. We conclude that targeting of specific chemokine and chemokine receptor pathways and/or ICOS may have clinical application in the prevention and treatment of TGP.


Chronic allograft nephropathy (CAN) is a leading cause of kidney allograft failure and a major problem in renal transplantation, given that the incidence of late graft failure from CAN has not changed appreciably despite successes in reducing the incidence of acute rejection. The mechanisms and extent of immune involvement in the pathogenesis of CAN are unknown. Histopathologic features of CAN include interstitial fibrosis, tubular atrophy, and fibrous intimal thickening of arteries, with variable glomerular lesions (1). Two pathologic features, transplant glomerulopathy (TGP), and splitting and lamination of the peritubular capillary basement membrane as detected by electron microscopy, are considered to be specific for chronic rejection, and may represent the immune component of CAN (2). TGP occurs in approximately 15% of CAN patients. Recent studies suggest the importance of antibody-mediated immune mechanisms in the pathogenesis of CAN as indicated by C4d deposits in peritubular capillaries and donor-specific alloantibodies in the serum (3). The presence of activated lymphocytes in TGP has not yet been determined, but could be important in that there is currently no effective therapy for CAN, and it is difficult to decide whether to increase or to decrease immunosuppression based on solely on Banff classification by light microscopy.

Chemokines are chemotactic proteins, which mediate inflammatory responses by promoting leukocyte migration, adhesion and effector functions. The role of various chemokines and their receptors has been extensively studied in rodent models of allograft rejection (4). Moreover, limited studies in human renal allograft recipients demonstrate increased expression of several chemokines and their receptors during acute rejection episodes, though the extent to which any or all of these chemokine/chemokine receptor pathways contribute significantly to clinical rejection is unknown. In the only previous study of the role of chemokines in human CAN, increased CCR5 expression was detected on infiltrating cells in 9 biopsy samples (5). In addition to chemokines, the CD28 homolog, ICOS, is a newly recognized costimulatory molecule involved in ongoing activation and effector functions of T cells (6). Allograft survival in ICOS–/– mice or recipients treated with an anti-ICOS mAb is significantly enhanced compared to controls, and use of an anti-ICOS mAb therapy can prevent development of chronic rejection (7).

We undertook an IRB-approved, blinded immunohistologic analysis of human renal transplant biopsies using monoclonal antibodies (Pharmingen, San Diego, CA, USA) to T (CD3) and B (CD19) cells, ICOS, chemokines (IP-10, Mig, MCP-1, MIP-1α, MIP-1β, MIP-3β, RANTES, TARC) and chemokine receptors (CCR1, CCR2, CCR3, CCR4, CCR5, CXCR3) on serial cryostat sections (8). C4d localization was done on paraffin sections using rabbit anti-C4d antibody (Biomedica, Vienna, Austria) (9). We studied 17 biopsies from 16 allograft recipients with CAN; one patient with TGP had a second biopsy 16 months after the first biopsy. Transplant kidney biopsies were done to investigate deteriorating renal function or proteinuria.

Six biopsies from five patients showed TGP (Figure 1a), with reduplication of the glomerular basement membrane, widening of the subendothelial space, interposition of mesangial matrix and endothelial swelling, in addition to CAN (Figure 1b). As shown in Table 1, recipient sex, age, transplant type, pretransplant PRA levels and other clinical features were comparable, except for the following: (i) incidence of a previous acute rejection episode; (ii) timing of biopsy post-transplant; (iii) incidence of proteinuria; and (iv) creatinine levels at the time of biopsy, especially as three patients with TGP had creatinine levels <1.5 mg/dL and were biopsied to investigate nephrotic-range proteinuria. The results of histopathologic analysis of the biopsies, with grading using Banff criteria, are listed in Table 1, and differed principally with respect to the extent of glomerulopathy. Peritubular capillary C4d deposition was detected in more than half the biopsies from each group, whereas glomerular C4d deposition was only infrequently observed (Table 1), suggesting that humoral alloreactivity was not a central feature of TGP, at least in this small series.

Figure 1.

Histologic and immunohistologic features of a representative renal transplant biopsy from a patient with CAN plus TGP. (a) H&E-stained paraffin section showing reduplication of the glomerular basement membrane and mesangial interposition. (b) Corresponding transmission electron microscopy showing glomerular capillary wall thickening and widening of the subendothelial space.

Table 1. : Clinicopathologic features of patients with CAN with or without TGP
ParameterCAN+ TGP– (n = 11)CAN+ TGP+ (n = 5)
  • a

    Months after transplantation.

  • b

    At biopsy, mean ± SEM.

  • c

    > 3.5 g/24 h.

  • d

    d Months after biopsy.

Recipient sex, male6 (55%)3 (60%)
Recipient age, median (range)47 (24–63)44 (23–57)
Transplant type, cadaveric6 (55%)3 (60%)
Biopsy time, median (range)a26 (9–210)44 (23–57)
Serum creatinine level (mg/dL)b2.9 ± 0.32.4 ± 0.6
Nephrotic range proteinuriac0 (0%)5 (100%)
Pretransplant PRA > 20%4 (36%)2 (40%)
Previous acute rejection episode2 (18%)3 (60%)
Graft failure after diagnosis6 (55%)2 (40%)
Graft failure time, median (range)d7 (1–20)10 (1–20)
CAN score (Banff), mean ± SEM:
 Allograft glomerulopathy (cg)0.36 ± 0.282.67 ± 0.21
 Interstitial fibrosis (ci)2.00 ± 0.271.50 ± 0.50
 Tubular atrophy (ct)1.18 ± 0.301.33 ± 0.42
 Vascular fibrous thickening (cv)0.91 ± 0.320.83 ± 0.31
Peritubular C4d+7 (64%)4 (80%)
Glomerular C4d+1 (9%)1 (20%)

In contrast to C4d staining, no intraglomerular T-cell, B-cell or ICOS staining was seen in CAN alone (Table 2, Figure 2), whereas all TGP samples showed intraglomerular CD3+ and ICOS+ staining (Figure 3) and lacked B-cell infiltration. ICOS is expressed by resting B cells as well as by activated but not resting CD4 and CD8+ T cells; other leukocytes lack ICOS expression (7). Hence, although double-labeling was not performed, the glomerular infiltration by T cells and associated ICOS staining suggests the presence of ICOS+ T cells in TGP, and indeed, the tight linkage between ICOS expression within kidney allografts with TGP vs. CAN alone argues for the importance of ongoing immune activation in the development of TGP.

Table 2. : Leukocyte labeling of renal biopsies from patients with CAN
CAN+ TGP+66/66/66/64/60/50/6
CAN+ TGP–110/90/70/90/90/110/11
Figure 2.

Immunohistologic labeling of serial cryostat sections of a biopsy from a patient with CAN alone (representative of 11 such biopsies analyzed), demonstrating lack of control mAb staining (asterisks indicate glomeruli present within serial sections), as well as lack of CD3, ICOS, CXCR3 or more than trace staining for Mig or IP-10. (Hematoxylin counterstain, × 100 original magnifications).

Figure 3.

Immunohistologic labeling of serial cryostat sections of a biopsy from a patient with CAN plus TGP(representative of 6 such biopsies analyzed), demonstrating lack of control mAb staining(asterisks indicate glomeruli present within serial sections), and glomerular and periglomerular infiltration by CD3+ T cells expressing ICOS and CXCR3, in association with local production of the CXCR3 chemokine ligands, Mig and IP-10. (Hematoxylin counterstain, × 100 original magnifications).

We also found selective chemokine and chemokine receptor expression by intraglomerular and periglomerular mononuclear cells in biopsies with TGP vs. CAN alone (Table 2) (p < 0.01). All TGP+ cases showed intraglomerular CXCR3+ CD3+ T cells (Figure 3), and concomitant labeling for the CXCR3 ligands, Mig (100%) and IP-10 (67% cases). No glomerular or periglomerular staining for CCR1, CCR2, CCR4 or CCR5 or their chemokine ligands was observed, and only minor staining of interstitial leukocytes in the rest of each biopsy was noted. Our CXCR3/Mig/IP-10 data are notable because both ligands are induced by IFN-γ, and CXCR3 is predominantly expressed by activated T cells and NK cells (10). CXCR3-deficient mice or mice treated with anti-CXCR3 mAb have markedly prolonged cardiac allograft survival compared to controls (11), as do allograft recipients following targeting of CXCR3 ligands (12), consistent with a key role of this pathway in mediation of experimental allograft injury. Intragraft expression of CXCR3 and its ligands also occurs during human cardiac allograft rejection (8,13). Moreover, studies using high-density oligonucleotide array (GeneChip) technology showed that only a small set of genes was persistently up-regulated in acutely rejecting human kidney biopsy samples, as compared to normal control transplant biopsies (14). Of this subset of genes associated with renal allograft rejection, Mig expression was up-regulated in all acute rejection biopsy samples, and had the highest hybridization intensity among the 10 gene transcripts detected. The current findings that all kidney allografts with TGP were positive for Mig and CXCR3, and that 4 out of 6 had IP-10 expression, whereas no CAN biopsies without TGP showed CXCR3, Mig or IP-10 expression, suggest the importance of this pathway in the pathogenesis of TGP.

In summary, the selective expression of ICOS and CXCR3, which are functionally defined markers of activated T cells, as well as expression of CXCR3 ligands in the context of CAN/TGP vs. CAN alone suggest different pathologic mechanisms underlie TGP vs. CAN. An ongoing effector T-cell response to glomerular antigens may be present, leading to persistent generation of chemokines, which attract and arm host effector cells. We conclude that the targeting of specific chemokine and chemokine receptor pathways may have clinical application in the prevention and/or treatment of TGP.