In clinical renal rejection, the now widely used and reproducible Banff system for scoring pathology (1) employs tubulitis, the invasion of tubular epithelium by lymphoid cells, as one of the major diagnostic criteria. Tubulitis is certainly considered to be a better indicator of rejection than simple interstitial accumulation of mononuclear leukocytes. T lymphocytes represent the major cell type involved in development of tubulitis, and, in previous studies, it has been demonstrated that a large proportion are CD8+, capable of proliferating and of being fully activated in situ to produce cytotoxic effector molecules such as perforin (2, 3). Thus, tubulitis, which provides a well defined example of environmental homing of T cells, frequently results in tubular epithelial cell damage and graft dysfunction.
Chemotactic factors, including chemokines, play a central role in the mechanism of entry of leukocytes to the grafted kidney (4). Initially, leukocytes migrate from the circulation into the interstitial spaces. The ultimate destination of these cells almost certainly depends upon differential chemokine secretion by resident cells of the specific microenvironment, in this case tubular epithelium. Local production of individual chemokines and their retention in the extracellular matrix may lead to up-regulation of chemokine receptors on specific leukocyte subpopulations (5). This would result in homing “up” a chemotactic concentration gradient to establish an inflammatory focus, as seen in tubulitis.
The significance of the many cell and molecular interactions that have been shown to play a role in leukocyte chemotaxis in vitro remains to be demonstrated in situ. In particular, there is little in vivo evidence to support the specificity of individual chemokines for particular subsets of lymphocytes. The ultimate spin-off from such an investigation could be the development of new therapeutic approaches to control acute rejection.
In the present study, the involvement of β-chemokines in the evolution of tubulitis was investigated by combining indirect immunofluorescence staining of transplant renal biopsies with semiquantitative, confocal microscopic analysis (6). Representative levels of expression of different chemokines in the tubular epithelium were recorded and related to pathological features and clinical aspects of individual biopsies. Table 1 shows details of the antibodies used, and Figure 1 gives details of the analysis by scanning laser confocal microscopy.
Routinely fixed and paraffin-embedded transplant biopsy specimens were randomly selected and analyzed without knowledge of patient identity. After investigation, patient reference numbers were entered into the Renal Biopsy Register from which pathology grading was obtained. Nine biopsies were classified as non-rejection and were not used for statistical analysis. In addition, only results from the earliest available posttransplant biopsy per graft were included. As a result, an additional five biopsies were excluded from statistical analysis. The final analysis included data from 47 biopsies from different grafts. The clinical parameters thought to be relevant to the present study were biopsy time (i.e., number of days after transplant), routine background immunosuppression, and acute rejection therapy. Details of these parameters were obtained from the Transplant Registry.
Graft recipients in this series were aged between 10 and 72 years with a mean and median age of 39.7 and 39 years, respectively. First biopsies after transplant (29) were all taken from the graft before 30 days. Second and third biopsies (18 from different grafts) were taken between 9 and 48 days after transplant. In most cases, the routine background immunosuppression at time of biopsy was based on cyclosporine, but three patients were receiving tacrolimus. The standard anti-rejection therapy given during episodes of biopsy-confirmed acute rejection was high-dose methylprednisolone. In addition, approximately 50% of the patients with grades 2 and 3 acute rejection were treated with anti-thymocyte globulin (Pasteur-Merieux).
This study was undertaken using carefully standardized experimental conditions. The four polyclonal chemokine antibodies used were produced and purified by R&D Systems, Abingdon, Oxon, United Kingdom and shown to have similar sensitivities in Western blot experiments. A reference section, treated with non-immune goat IgG in place of primary antibody, was included for each biopsy. The level of expression of each of the chemokines in tubular epithelium was assigned a representative value for each biopsy between 50 and a maximal 250 pixel units. This was in accordance with background fluorescence within the corresponding reference section being normalized to 50 pixel units to allow day-to-day comparison of the chemokine staining results; this value provided the threshold for positive staining.
For all primary anti-chemokine antibodies, representative fluorescence in tubular epithelium in “normal” kidney sections (n=5) sometimes exceeded the reference fluorescence values: RANTES median, 50 (range, 50–75); MCP-1 median, 50 (range, 50–90); MIP-1α median, 50 (range, 50–100); and MIP-1β median, 60 (range, 50–80). These normal values may indicate some low-level constitutive expression or subclinical inflammation; however, the median values were consistently less than those recorded for rejection biopsies.
Having established the protocol for staining and analysis of sections, five biopsies were used as control material to monitor assay reproducibility. It was found that each biopsy produced essentially identical immunofluorescence results after staining on at least two separate occasions. In addition, there was very little difference in expression of any of the four chemokines between different regions of each biopsy.
A general staining pattern for β-chemokines was established. These cytokines were primarily located in graft tubular epithelium, with highest concentrations being observed at the basolateral surface and, occasionally, within the basement membrane (Fig. 1a). In addition, a small proportion of graft infiltrating cells was positive for all four chemokines at similar levels (≥200 pixel units;Fig. 1a). Chemokines were also diffusely distributed in the interstitial matrix; low-intensity staining was present in endothelial cells of some peritubular venules and frequently defined a network of interstitial fibroblast-like cells. Profiles of fluorescence intensity clearly illustrate peaks of chemokine expression in tubular epithelium (Fig. 1b).
Several in vitro studies have demonstrated the chemotactic/haptotactic properties of β-chemokines for T cells (7). It is now clear that chemokines, in common with cytokines such as interferon-γ, bind with relatively high affinity to the components of cell-surface and extracellular matrix proteoglycans. These glycosaminoglycan chains form a dense “scaffolding” for chemokine immobilization and presentation to leukocyte surface receptors. The general chemokine distribution, observed in the current study, is consistent with these in vitro findings and suggests a mechanism for microenvironment-specific homing and activation of T cells during the development of tubulitis.
The precise stimulus responsible for the induction of chemokine expression remains unclear. However, β-chemokines were observed in tubular epithelium in two out of three biopsies from primary nonfunction grafts showing widespread ischemic changes. This suggests that β-chemokine production by tubular epithelium can occur very soon after transplantation, or may even be induced before engraftment.
Linear regression analysis showed that the posttransplant time of the biopsy and the levels of β-chemokine expression in tubular epithelium were not significantly linked. In addition, there were no significant differences (Mann-Whitney U test) between the levels of expression of any of the chemokines in first and second/third posttransplant biopsies. These results suggest that β-chemokine expression by tubular epithelium is equally important in all episodes of rejection. Moreover, one might speculate that sustained expression and retention of chemotactic molecules could be factors influencing the development of serial episodes of rejection, especially in cases unresponsive to current anti-rejection therapy.
Figure 2 shows that the level of expression of each chemokine was generally lower in grade 1 rejection biopsies (including biopsies with mild focal interstitial infiltrate) than grade 2a biopsies, in which the moderate acute rejection diagnosis is based on the degree of tubulitis (1). Statistical analysis, using the Mann-Whitney U test, showed that differences in the expression of RANTES and MIP-1α, between grades, were not significant. However, there was a significant difference in expression of both MCP-1 (confidence interval 95%, P =0.006) and MIP-1β (confidence interval 95%, P =0.002) between grade 1 (19 cases) and grade 2a rejection (18 cases). Overall, the results of the present study confirm that β-chemokines play a key role in the development of acute cellular rejection and put into context our earlier observations on activation and proliferation of cytotoxic T cells in the tubular epithelium (2, 3).
We have observed that RANTES is expressed in tubular epithelium in all grades of acute cellular rejection. RANTES is well known to influence diverse T-cell functions such as adhesion molecule expression, proliferation, and cytokine release in vitro in addition to its chemotactic role (7). Our results appear to support a multifunctional role for this versatile chemokine in acute cellular rejection. However, protocol transplant renal biopsies are not available, and, in the absence of information from such biopsies, it is not possible to attribute RANTES expression to rejection alone.
Statistical analysis of the data obtained in this study suggests that MCP-1 and MIP-1β could have a more subtle, regulatory role to play during the development of tubulitis. Both of these chemokines were present at very low levels in tubular epithelium in grade 1 rejection when compared with grade 2a (Fig. 2). This is perhaps the most significant finding of the study.
In addition to the redundancy and overlapping activities of individual chemokines, there is considerable promiscuity of ligand/receptor binding. However, only one receptor, CCR5, has been identified for MIP-1β. Eitner et al. (8) recently reported the presence of CCR5-expressing mononuclear cells during allograft rejection. In addition, it has been demonstrated that the majority of T-cell clones generated from renal allografts undergoing acute rejection exhibit a polarized Th1 (Tc1) pattern of cytokine secretion regardless of their CD4/CD8 phenotype (9). CCR5 has been recognized by some as a marker of Th1/Tc1 lymphocytes, although Sallusto et al. (5) have suggested that it is also present on recently activated T cells regardless of their functional polarization. Our observations prompt speculation that MIP-1β may indeed be important, in vivo, in preferential recruitment of Th1/Tc1 cell subsets into the tubular microenvironment, where such cells would be responsible for proliferation and continued activation, including the induction of cytotoxic effector function.
Grandaliano et al. (4) reported that the expression of MCP-1 protein, during acute renal allograft rejection, was mainly localized to proximal tubules and infiltrating mononuclear cells. These authors speculated on a role for MCP-1, bound to extracellular matrix, in attracting and activating monocytes in the interstitium, which would be self-perpetuating and stimulatory with respect to tubular epithelium, as these monocytes would also release proinflammatory cytokines. Our results are in agreement with such a role for MCP-1. Moreover, CCR2 has been identified as the principal receptor for MCP-1 and is also present on differentiated T cells. Thus, a role for MCP-1 in the recruitment of T cells to the interstitium is feasible and supports the notion that “leukocytes can respond sequentially to chemokines in a multi-step navigation mode using different receptors” (5).
The precise role played by MIP-1α during allograft rejection remains enigmatic. It may be that MIP-1α plays a complementary role to RANTES and/or MIP-1β, as they all share CCR5 as a receptor. Steinhauser et al. (10) suggest that this chemokine is not produced by structural cells. However, in the present study, MIP-1α was associated with tubular epithelium in both primary nonfunction and acute rejection biopsies, although its level of expression was more variable that of the other β-chemokines. Steinhauser et al. found MIP-1α to be a potent chemoattractant and activator of monocytes in vitro. Furthermore, their work led to the suggestion that prolonged MIP-1α stimulation of macrophages could induce a chronic inflammatory state. As such, MIP-1α could be a factor in the cycle of events leading to chronic inflammation in allograft rejection. Further assessment of MIP-1α expression in biopsies will be required to substantiate this hypothesis.
In future, our studies will focus on the expression of chemokine receptors in renal allograft rejection, as recruitment of specific T-cell subpopulations to a discreet microenvironment almost certainly results from differential expression of a range of chemokines by structural cells and their corresponding receptors on infiltrating cells.
The authors thank Mostafa Mohamed for retrieval of clinical data, Dr. Trevor Booth for expert assistance with confocal microscopy, and Dr. Simi Ali for critically reading the manuscript.
1. Solez K, Racusen L, Rayner D, Olsen S, Halloran P. The Banff schema four years later. Transplant Proc 1996; 28 (1):450.
2. Robertson H, Wheeler J, Thompson V, Johnson JS, Kirby JA, Morley AR. In-situ lymphoproliferation in renal-transplant biopsies. Histochem Cell Biol 1995; 104 (4):331.
3. Robertson H, Wheeler J, Kirby JA, Morley AR. Renal allograft rejection: in situ demonstration of cytotoxic intratubular cells. Transplantation 1996; 61 (10):1546.
4. Grandaliano G, Gesualdo L, Ranieri E, Monno R, Stallone G, Schena FP. Monocyte chemotactic PEPTIDE-1 expression and monocyte infiltration in acute renal transplant rejection. Transplantation 1997; 63 (3):414.
5. Sallusto F, Lanzavecchia A, Mackay CR. Chemokines and chemokine receptors in T-cell priming and Th1/Th2-mediated responses. Immunol Today 1998; 19 (12):568.
6. Robertson H, Wheeler J, Morley AR, Booth TA, Talbot D, Kirby JA. β-Chemokine expression and distribution in paraffin-embedded transplant renal biopsy sections: analysis by scanning laser confocal microscopy. Histochem Cell Biol 1998; 110 (2):207.
7. Ward SG, Westwick J. Chemokines: understanding their role in T-lymphocyte biology. Biochem J 1998; 333 (Part 3):457.
8. Eitner F, Cui Y, Hudkins KL, et al. Chemokine receptor (CCR5) expression in human kidneys and in the HIV infected macaque. Kidney Int 1998; 54 (6):1945.
9. D’Elios M, Josien R, Manghetti M, et al. Predominant Th1 cell infiltration in acute rejection episodes of human kidney grafts. Kidney Int 1997; 51 (6):1876.
10. Steinhauser ML, Kunkel SL, Hogaboam CM, Evanoff H, Strieter RM, Lukacs NW. Macrophage/fibroblast coculture induces macrophage inflammatory protein-1 α production mediated by intercellular adhesion molecule-1 and oxygen radicals. J Leukocyte Biol 1998; 64 (5):636.