Pathology of Kidney Transplantation

Tubulointerstitial Rejection (Type I)

The prominent microscopic feature of TCMR is a pleomorphic interstitial infiltrate of mononuclear cells, accompanied by interstitial edema and sometimes hemorrhage ( Fig. 25.1 ). The infiltrate is typically patchy, both in the cortex and medulla. The infiltrating cells are primarily T cells and macrophages. Activated T cells (lymphoblasts) with increased basophilic cytoplasm, nucleoli, and occasional mitotic figures indicate increased synthetic and proliferative activity. Granulocytes are not uncommonly present but rarely prominent. When neutrophils are conspicuous, the possibility of AMR or pyelonephritis should be considered. Eosinophils are present in about 30% of biopsies with rejection and can be abundant, but are rarely more than 2% to 3% of the infiltrate. Abundant eosinophils (10% of infiltrate) are associated with endarteritis (Banff type II). Mast cells increase, as judged by tryptase content, and correlate with edema. Acute rejection with abundant plasma cells has been described as early as the first month associated with poor graft survival. Infiltrating T cells express cytotoxic molecules, namely perforin, FasL, granzyme A and B, and TIA-1/GMP-17, and tumor necrosis factor-β (lymphotoxin). Apoptosis of the infiltrating T cells can be demonstrated with the TdT-uridine-nick end label (TUNEL) technique, probably as a result of activation-induced cell death, and would thereby serve to limit the immune reaction.

Mononuclear cells invade tubules and insinuate between tubular epithelial cells, a process termed “tubulitis” (see Fig. 25.1 inset), which is best appreciated in sections stained with PAS or a silver stain to delineate the tubular basement membrane (TBM). All cortical tubules (proximal and distal) as well as the medullary tubules and the collecting ducts may be affected. Tubular cell apoptosis occurs, which correlates with the number of cytotoxic cells and macrophages in the infiltrate. Tubular epithelial cells express human leukocyte antigen–DR (HLA-DR), intercellular adhesion molecule 1 (ICAM 1), and vascular cell adhesion molecule 1 (VCAM-1) in increased amounts in TCMR and express the costimulatory molecules CD80 and CD86. Tubules also synthesize tumor necrosis factor-α, transforming growth factor-β1, IL-15, osteopontin, and vascular endothelial growth factor (VEGF). Increased expression of S100A4 may signal the process of epithelial to mesenchymal transition (EMT), which may actually consist of an in situ epithelial response rather than a true emigration of tubular epithelial cells into the interstitium. Thus, as suggested by proceedings at a Banff conference on allograft pathology, EMT may be better thought of as an epithelial to mesenchymal phenotype (EMP). Some tubular cell-derived molecules have the potential to inhibit acute rejection, such as protease inhibitor-9 (PI-9), the only known inhibitor of granzyme B and IL-15, which inhibits expression of perforin.

CD8+ and CD4+ cells invade tubules. Intratubular T cells with cytotoxic granules, and CD4+FOXP3+ cells accumulate selectively in the tubules, compared with the interstitial infiltrate. T cells proliferate once inside the tubule, as judged by the marker Ki67 (MIB-1), which contributes to their concentration within tubules, in addition to selective invasion. Increased tubular HLA-DR, tumor necrosis factor (TNF)α, IFNγ receptor, IL-2 receptor, and IL-8 are detectable by immunoperoxidase study in TCMR. Several adhesion molecules are increased on tubular cells during rejection, including ICAM-1 (CD54) and VCAM-1, and correlate with the degree of T cell infiltration.

Signs of tubular cell injury can be detected by TUNEL apoptosis assay. Increased numbers of TUNEL+ tubular cells are present in acute rejection, compared with normal kidneys. The frequency was significantly lower in cyclosporin A (CsA) toxicity or ATN. The degree of apoptosis correlates with the cytotoxic cells in the infiltrate, consistent with a pathogenetic relationship. Prominent apoptosis of the infiltrating T cells has also been detected at a frequency comparable to that in the normal thymus (1.8% of cells). Others have described occasional TUNEL+ lymphocytes. Apoptosis probably occurs in infiltrating T cells as a result of activation-induced cell death and would thereby serve to limit the immune reaction. Little, if any, immunoglobulin deposition is found by immunofluorescence in TCMR, which is characterized primarily by extravascular fibrin accumulation in the interstitium and not uncommonly increased C3 along the TBM. The C3 is largely derived from tubular cells. C3 may have a role in the pathogenesis of acute rejection, because C3-deficient mouse kidneys have prolonged survival. C4d deposition in PTCs indicates an antibody-mediated component.

Gene expression studies of graft tissue have revealed that transcripts for proteins of cytotoxic T lymphocytes (CTLs), such as granzyme B, perforin, and Fas ligand and the master transcription factor for CTLs, T-bet, are characteristic of TCMR. Graft CTL-associated transcripts (CATs) precede tubulitis in mouse kidney grafts. Treatment of rejection is followed by a measurable decrease of CATs. However, knockout of either granzyme or perforin does not prevent acute rejection, suggesting they are not essential. IFNγ mRNA is detectable in fine needle aspirates 1 week before the clinical onset of rejection. Other genes associated with acute rejection are IFNγ, TNFβ, TNFα, RANTES (regulated on activation, normal T cell expressed and secreted, also known as also Chemokine (C-C motif) ligand 5 [CCL5]), and macrophage inflammatory protein 1-alpha (MIP-1-alpha, also known as Chemokine [C-C motif] ligand 3 [CCL3]); no elevation of TGFβ or IL-10 is detected.

Endarteritis (Type II Rejection)

Infiltration of mononuclear cells under arterial and arteriolar endothelium is the pathognomic lesion of TCMR ( Fig. 25.2 ). Many terms have been used for this process, including “endothelialitis,” “endothelitis,” “endovasculitis,” “intimal arteritis,” or “endarteritis.” We prefer the last term, which emphasizes the type of vessel (artery vs. vein) involved and the site of inflammation. Mononuclear cells that are sometimes attached to the endothelial surface are insufficient for the diagnosis of endarteritis; however, they probably represent the early phase of this lesion. Endarteritis in TCMR must not be confused with fibrinoid necrosis of arteries. The latter is characteristic of acute AMR and can also be seen in thrombotic vasculopathy. Regrettably, some still do not separate these lesions, regarding all “vascular rejection” as predominately humoral.

Endarteritis has been reported in 35% to 56% of renal biopsies with TCMR. Many do not find the lesion as often, which may possibly be ascribed to inadequate sampling, overdiagnosis of rejection (increasing the denominator), patient population with respect to medication adherence (severity of rejection), or the timing of the biopsy with respect to antirejection therapy. Endarteritis lesions affect arteries of all sizes including the arteriole, although the lesions affect larger vessels preferentially. For example, in a detailed analysis, 27% of the artery cross sections were affected, vs. 13% of the arterioles. A sample of four arteries would have an estimated sensitivity of about 75% in the detection of type II rejection. Thus a sample may not be considered adequate to rule out endarteritis unless several arteries are included. “Arteriolitis” has the same significance as endarteritis. Endarteritis can occur in cases with little or no interstitial infiltrate or tubulitis, arguing that it has a distinct pathogenetic mechanism, and even in cases with “isolated endarteritis,” that finding is an independent risk factor for kidney transplant failure. In severe cases, a transmural mononuclear infiltrate affects the media, with focal necrosis of the myocytes, features that constitute type III rejection (transmural inflammation or fibrinoid necrosis). Although this occasionally occurs in the absence of demonstrable antibodies, it is more typical of AMR.

Endothelial cells are typically reactive with increased cytoplasmic volume and basophilia. The endothelium shows disruption and lifting from supporting stroma by infiltrating inflammatory cells. Occasionally endothelial cells are necrotic or absent, however, thrombosis is rare. Endothelial apoptosis occurs and increased numbers of endothelial cells appear in the circulation. The media usually shows little change. In severe cases a transmural mononuclear infiltrate may be seen (termed “type III rejection”). The cells infiltrating the endothelium and intima are T cells and monocytes, but not B cells. Both CD8+ and CD4+ cells invade the intima in early grafts, but later CD8+ cells predominate, suggesting that class I antigens are the primary target. Vascular endothelial cell apoptosis can be detected in sites of endarteritis.

Normal arterial endothelial cells express class I antigens, weak ICAM-1, and little or no class II antigens, or VCAM-1. During acute rejection the endothelium of arteries expresses increased HLA-DR and ICAM-1 and VCAM-1. This adhesion molecule upregulation occurs in association with CD3+ ⁸² and CD25+ ⁸⁰ infiltrating mononuclear cells. Endothelial cells also have decreased endothelin expression in rejection with endarteritis, but not in tubulointerstitial rejection.

Glomerular Lesions

In most TCMR cases, glomeruli are spared or show minor changes, typically a few scattered mononuclear cells (T cells and monocytes) and occasionally segmental endothelial damage ( Fig. 25.3 ). A severe form of this glomerular injury, termed “transplant glomerulitis” or “acute allograft glomerulopathy,” develops in a minority of cases (<5%), manifested by hypercellularity, injury, and enlargement of endothelial cells; infiltration of glomeruli by mononuclear cells; and by webs of PAS-positive material. Crescents and thrombi are rare. Endarteritis often accompanies the transplant glomerulitis. The glomeruli contain numerous CD3+ and CD8+ T cells and monocytes. Fibrin and scant immunoglobulin and complement deposits are found in glomeruli. This variant of cellular rejection has been associated with certain viral infections, such as cytomegalovirus (CMV) infection and hepatitis C virus, although viral antigens are not in the glomerular lesions.

Atypical Rejection Syndromes

Unique patterns of rejection have been observed under novel immunosuppression regimens. For example, following pronounced lymphocyte depletion from alemtuzumab (CAMPATH-1H), TCMR with a prominent monocyte population (i.e., an acute monocytic rejection) has been described. In these cases, much of the interstitial rejection infiltrate stains for CD68, correlating with renal dysfunction and tubular stress, shown by HLA-DR staining of the tubules. Under these conditions, T cells did not correlate with renal dysfunction or HLA-DR staining.

Studies have included simultaneous bone marrow and kidney transplantation protocols in attempt to induce tolerance to the transplanted organ. In these studies, human leukocyte antigen (HLA)-mismatched renal transplants have been performed; withdrawal of maintenance immunosuppression has been accomplished in some of the patients with relatively preserved renal function. In several of these patients, a capillary leak or engraftment syndrome has been observed around 10 days after a simultaneous kidney/bone marrow transplant preceded by a nonmyeloablative conditioning regimen. In this “engraftment syndrome,” acute tubular injury is accompanied by congested PTCs containing mononuclear cells and red blood cells. IHC shows that the cells are primarily CD68+MPO+ mononuclear cells and CD3+CD8+ T cells, the latter with a high proliferation index (Ki67+). XY chromosome fluorescence in situ hybridization (FISH) has been used to demonstrate that the PTC cells are recipient derived, correlating with chimerism studies showing a simultaneous decline in circulating donor cells and recovery of recipient circulating cells. PTC endothelial injury can also be seen on EM in these cases. The etiology of the syndrome remains undefined, and others have performed combined kidney and bone marrow transplants without observing this phenomenon. With modifications in the combined kidney and bone marrow transplantation protocol, it is possible that the “engraftment syndrome” can be eliminated or at least attenuated; this suggests that “engraftment syndrome” may not be an accurate term for what may actually just be a form of transient acute kidney injury.

Differential Diagnosis

TCMR typically has a diffuse, interstitial mononuclear cell infiltrate, whereas patients with CNI toxicity (CNIT) and those with stable function have only focal mononuclear cell infiltrates ( Table 25.3 ). Endarteritis or C4d+ is found extremely rarely, if ever, in CNIT and if either is present, is the most discriminating feature for acute rejection. Prominent tubulitis favors acute rejection, because it is less prominent in acute tubular necrosis, particularly in the proximal tubules. However, tubulitis has been documented in renal transplants with dysfunction due to lymphoceles (obstruction) and in urine leaks, possibilities that need to be considered and excluded by other techniques. Acute obstruction typically has some dilation of the collecting tubules, especially in the outer cortex. Edema and a mild mononuclear infiltrate are also common.

Interstitial mononuclear inflammation and tubulitis occur in a variety of diseases other than acute rejection, such as drug-induced (allergic) or infectious tubulointerstitial nephritis. When eosinophils are more abundant than usual for rejection and eosinophils invading tubules are identified, then drug allergy may be favored over rejection. The presence of endarteritis permits a definitive diagnosis of active rejection. Lymphocytes commonly surround vessels (without medial involvement), a nonspecific feature, and must not be confused with endarteritis. Tubulitis is often present in atrophic tubules and does not indicate acute rejection. The diagnosis of acute pyelonephritis should be raised when active inflammation and abundant intratubular neutrophils are present. A note of caution though because in acute AMR, neutrophilic tubulitis with neutrophil casts can be seen; a C4d stain will help in distinguishing between these. A positive urine culture will also separate infection from rejection.

Polyoma virus interstitial nephritis (BK virus) is often diagnosed by the presence of the enlarged, hyperchromatic tubular nuclei with lavender viral nuclear inclusions, often in collecting ducts. However, these may be inconspicuous, and diligent study of multiple sections may be required. Other clues are prominent apoptosis of tubular cells and abundant plasma cells, which invade tubules. IHC for the polyoma SV40 large T antigen or in situ hybridization for BK polyomavirus and EM (even of paraffin) will confirm the diagnosis. Sometimes BK virus infection, with its exuberant plasmacytic infiltration and activated immunoblasts may be confused with the plasmacytic hyperplasia form of posttransplant lymphoproliferative disease, which also should be considered in the differential diagnosis of acute cellular rejection. Rarer infections, including microsporidia, should also be considered in biopsies with interstitial inflammation.

Diagnostic Criteria

The three diagnostic criteria for acute AMR are (1) histologic evidence of acute injury (neutrophils in capillaries, acute tubular injury, fibrinoid necrosis), (2) evidence of antibody interaction with tissue (typically C4d in PTCs), and (3) serologic evidence of circulating antibodies to antigens expressed by donor endothelium (typically HLA). Criteria for the diagnosis of acute AMR have been refined over the years. Generally speaking, if only two of the three major criteria are established (e.g., when antibody is negative or not done), the diagnosis can be considered suspicious for acute AMR. Biopsies meeting criteria for both acute AMR and TCMR type I or II are considered to have both forms of rejection. Biopsies with C4d and no pathology are likely a manifestation of “accommodation” (see later).

Pathologic Features

Histologic findings are typically scant to moderate mononuclear interstitial infiltrates, sometimes with prominent neutrophils and increased numbers of macrophages (see Fig. 25.3 ). The extent of mononuclear infiltration often does not meet the criteria for TCMR. PTCs have neutrophils in about 50% of cases and are classically dilated ( Fig. 25.4A ). Interstitial edema and hemorrhage can be prominent. Glomeruli have accumulations of macrophages (∼50% of cases) and neutrophils (∼25% of cases; see Fig. 25.3 ) and occasionally fibrin thrombi or segmental necrosis. Acute tubular injury, sometimes severe, can be identified in many cases and may be the only initial manifestation of acute AMR. Focal necrosis of whole tubular cross sections, similar to cortical necrosis has been reported; 38% to 70% of acute AMR cases may have patchy infarction. Little mononuclear cell tubulitis is found, although a neutrophilic tubulitis with or without neutrophil casts may be prominent, resembling acute pyelonephritis. Plasma cells can be abundant in acute AMR, either early or late after transplantation, sometimes associated with severe edema and increased IFNγ production in the graft. B cells can be also present, but have no apparent diagnostic value.

In about 15% of cases small arteries shows fibrinoid necrosis, with little mononuclear infiltrate in the intima or adventitia but with neutrophils and karyorrhectic debris ( Fig. 25.5 ). Arterial thrombosis can be found in 10%, and a pattern resembling TMA has also been reported. Around 75% of cases with fibrinoid necrosis are C4d positive. Presumably the C4d-negative cases had T-cell-mediated rejection or TMA. Antibodies to the angiotensin II type 1 receptor have been detected in a few cases with arterial fibrinoid necrosis, in the absence of capillary C4d deposition. The presence of mononuclear endarteritis in cases of acute AMR strongly suggests a component of T-cell-mediated rejection.

By EM the PTCs are dilated, containing neutrophils. The endothelium is reactive and shows loss of fenestrations. The glomerular endothelium is separated from the GBM by a widened lucent space with endothelial cell swelling and loss of endothelial fenestrations, indicative of injury. Platelets, fibrin, and neutrophils are found in glomerular and PTCs. The small arteries with fibrinoid necrosis show marked endothelial injury and loss, smooth muscle necrosis, and deposition of fibrin.

C4d Interpretation

Feucht and colleagues first drew attention to C4d as a possible marker of an antibody-mediated component of severe rejection. C4d, a fragment of complement component C4, is released during activation of the classical complement pathway by antigen–antibody interaction. Because C4d contains a thioester bond, it binds covalently to tissues at the local site of activation. The covalent linkage explains why C4d remains for several days after alloantibody disappears, because antibody binds to cell surface antigens that can be lost by modulation, shedding, or cell death.

Although immunoglobulin deposition is found in only a minority of cases, C4d is characteristically detected in a widespread, uniform ring-like distribution in the PTCs by immunofluorescence in cryostat sections (see Fig. 25.4B ). Deposition occurs in both the cortex and medulla. Using IHC in formalin-fixed, paraffin embedded tissue, C4d has a similar pattern, although the intensity is variable. “Serum staining” is an artifact of C4d IHC, so PTCs must show clear circumferential staining to be called positive by this technique. Glomerular capillary staining also occurs but is hard to distinguish from C4d normally found in the mesangium in frozen sections stained by immunofluorescence. Formalin fixation eliminates this background staining and demonstrates glomerular C4d in about 30% of acute AMR cases.

Grafts with focal C4d (<50% of PTC) are of uncertain significance and the patient should be monitored closely for donor-reactive antibodies. Two of three studies have failed to show any significant clinical or pathologic difference between cases with focal and diffuse C4d staining. C4d deposition can precede histologic evidence of acute AMR by 5 to 34 days. C4d in 1-week protocol biopsies was followed by clinical acute rejection in 82% of cases and was associated with donor-reactive antibodies.

In the setting of acute rejection, C4d is a specific (96%) and sensitive (95%) marker of circulating antidonor HLA-specific antibodies by the antihuman globulin cytotoxicity test. PTC C4d deposition is associated with concurrent circulating antibodies to donor HLA class I or II antigens in 88% to 95% of recipients with acute rejection. Moreover, C4d deposition and the severity of histologic injury by antibody correlates with the serum DSA level in acute humoral rejection. False negative antibody assays may be due to absorption by the graft as shown by elution from rejected grafts in patients who had no detectable circulating antibody, or it may be due to differences in detection of antibody directed against different HLA alleles and sensitivity of solid-phase methods for particular alloantigens. Alternatively, non-HLA antigens may be the target. C4d-negative acute rejection may show flow cytometry evidence of antidonor reactive antibodies as frequently as 50%, due in part to noncomplement fixing antibodies. Cell based assays have a false positive rate of <10%.

In a comparison of methods for C4d, the triple layer immunofluorescence technique proved the most sensitive, although the difference with IHC in paraffin embedded tissue was small. With fixed tissue, plasma in the capillaries and interstitium may stain for C4d, which interferes with interpretation.

Other components of the complement system have been sought. C3d, a degradation product of C3, was found in PTCs in 39% to 60% of biopsies from HLA-mismatched grafts with diffuse C4d. C3d was usually but not always associated with C4d. C3d correlated with acute AMR in all studies, and was associated with increased risk of graft loss in two series, compared with C3d-negative cases, but C3d provided no convincing additional risk compared with C4d+. The interpretation of C3d stains is complicated by the common presence of C3d along the TBM. Even though C3d should indicate more complete complement activation, it added no diagnostic value to C4d in grafts showing histologic features of acute AMR, except in the setting of ABO-incompatible grafts. Other complement components, such as C1q, C5b-9, and C-reactive protein (CRP) are not conspicuous in PTCs in acute rejection. Lectin pathway components, which activate C4 by binding to microbial carbohydrates, are sometimes detected. Among 18 biopsies with C4d, 16 had diffuse H-ficolin deposition along the PTCs, whereas none of the 42 cases without C4d had H-ficolin. No Mannan-binding lectin serine protease 1 (MASP-1, also known as mannose-associated serine protease 1) or MASP-2 was detectable. The significance of this observation is not clear, because MASP proteins are required to activate C4 via the ficolins or Mannose-binding lectin (MBL).

Natural killer (NK) cells have been the focus of recent research in graft injury, particularly regarding AMR. Microarray analysis has indicated that several DSA-specific gene transcripts show high expression in NK cells, and IHC also shows prominent numbers of peritubular capillary NK cells in these cases. Depletion of NK cells with anti-NK1.1 significantly reduced DSA-induced chronic allograft vasculopathy in a murine cardiac allograft model.