Acute rejection episodes leading to treatment refractory early graft loss are increasingly rare events in living related renal transplantation today. Pathophysiologic pathways often remain unsolved. We report on tubulointerstitial and vascular rejection developing within 2 weeks after transplantation in a 12-year-old boy treated with cyclosporine, mycophenolate, steroids and double blinded basiliximab. Despite steroid pulses, switch to tacrolimus and ATG serum creatinine peaked at 347 μmol/L with imminent graft loss and ongoing C4d negative cellular vascular rejection. Permanent gain of function was only achieved after a single dose of rituximab. Retrospectively CD20+ nodular B-cell aggregates could be demonstrated in all three biopsies obtained prior to rituximab and resolved concomitantly with functional improvement. Our case for the first time demonstrates resolution of nodular CD20+ infiltrates and decline of OX40, NF-κB and CTL transcription shortly after rituximab indicating a B-cell facilitated C4d negative pathway. Single dose rituximab may effectively reverse even long-lasting refractory rejection.
Acute rejection (AR) after renal transplantation is a complication known to have an impact on long-term graft survival. Imminent graft loss due to acute rejection is getting an exceptionally rare situation with current potent anti-T-cell mediated immunosuppression.
Mechanisms of both steroid- and antilymphocyte-globulin-resistant rejection episodes are poorly understood. Recent studies have shown a high density of intragraft CD-20+ B cells in patients who experience steroid-resistant rejection episodes (1). Immunosuppressive therapy focusing on T cells is not effective in these cases. Anecdotal data indicate that the pattern of intragraft gene expression in treatment refractory AR in fact may differ from steroid responsive AR (2).
We hereby present the course of a 12-year-old boy receiving a living related kidney transplant from his father. The boy rapidly developed an acute cellular tubulointerstitial and C4d negative vascular rejection with presence of CD-20+ cells in the graft refractory to steroids, tacrolimus and ATG. Imminent loss of the graft function was only prevented by administration of rituximab, a specific anti-CD20 antibody, achieving an ongoing gain of function. Intragraft events were closely followed by four sequential biopsies obtained during the course. Both histopathological presentation and gene expression were evaluated during AR and after rituximab treatment.
A 12-year-old boy with end-stage renal disease due to focal segmental glomerulosclerosis (FSGS) received a living donor renal transplantation from his father. He was on hemodialysis for 2 months and received only one pack of red blood cells prior to transplantation. HLA mismatch on A/B/DR locus was 0/1/1 and the pre-transplant cross match was negative with panel reactive antibodies of 0%. The recipient was negative for EBV (IgM, IgG and PCR) and CMV. IgG titers for both viruses were detectable in the donor.
The initial post-operative course was uneventful. Immunosuppressive medication consisted of cyclosporine, mycophenolate and methylprednisolone. The patient was included in a double blinded basiliximab study. Randomization data are unblinded. Serum creatinine (Scr) declined to 92 μmol/L on day 5.
On day 12, Scr increased to 164 μmol/L and a transplant renal biopsy was performed on the same day. Histology showed a focal infiltrate of mononuclear cells and with mild tubulitis classified as Banff borderline. Treatment remained unchanged with high cyclosporine levels aimed at 250 ng/mL (EMIT assay). As Scr did not decline substantially until day 25, a second transplant biopsy was obtained. Histology now showed acute cellular tubulo-interstitial rejection graded as Banff IA. Anti-rejection therapy was initiated with methylprednisolone pulses (300 mg/m2) for 3 days and in addition, immunosuppressive medication was switched from cyclosporine to tacrolimus. Despite this treatment renal function worsened and Scr reached 300 μmol/L on post-operative day 33. A third biopsy revealed continued signs of acute cellular rejection, but with additional vascular involvement (Banff grade IIA). Anti-thymocyte globulin (20 mg/kg/day) was given intravenously for 6 days with transient improvement of renal function and subsequent rise of Scr to 347 μmol/L at day 52 (Figure 1).
A single dose of rituximab (375 mg/m2) was given intravenously without side effects. At this time, the previously negative EBV-PCR increased to 4000 copies/mL (Norm <250, in overt PTLD >100 000 copies). A fourth renal biopsy was taken on day 57 because creatinine did not lower 5 days after rituximab treatment. It revealed complete resolution of AR. The boy never experienced clinically overt bacterial or fungal or viral (CMV, EBV, Polyoma) infection. Proteinuria was limited to levels of 0.14, 0.06, 0.21 and 0.06 g albumin/mmol creatinine at the time points biopsies were taken. At 6 months graft function remained stable with a Scr of 175 μmol/L.
A month later he presented with fatigue, loss of appetite, anemia, abdominal pain and hepatosplenomegaly. No enlarged peripheral lymph nodes were detectable. PET and gastroscopy revealed a monomorphic post-transplant lymphoproliferative disease (PTLD) located within the stomach later characterized as a CD20+, CD30+, BCL-2+EBV+, CD10, TdT, CD3 negative B-cell lymphoma with a Ki-67-index of 70%. EBV virus load was 3000 copies only. In PET no enhancement of glucose metabolism within the transplant was notable. PTLD was successfully treated with six administrations of rituximab within 3 months according to a consensus protocol. Immunosuppression was reduced withholding tacrolimus and initiation of sirolimus thereafter. Renal function further improved and Scr remains stable at 150 μmol/L 13 months post-transplantation.
Histopathological Biopsy Evaluation
Biopsies were obtained at days 12, 25, 33 and 57 post-operatively for clinical purposes. All four re-reviewed biopsies had sufficient material according to the Banff criteria with at least one arterial cross section and seven glomeruli. All biopsies were negative for C4d in peritubular capillaries.
Retrospectively further characterization of lymphocytic infiltrates in the four biopsies was done. Staining for B lymphocytes by anti-CD20 revealed perivascular dense nodular CD20+ B-cell aggregates within the cortex localized in the first three biopsies (Figure 2A). To exclude retrospectively, an early intragraft manifestation of PTLD allograft biopsies with nodular B-cell infiltrates as well as a tissue specimen (stomach) leading to the diagnosis of PTLD were immunohistochemically stained for light chains and EBV-LMP. Additionally, an in situ hybridization for EBV (EBER) and a clonality analysis for the Ig-heavy chain rearrangement were carried out. The results of these assays demonstrated clearly that the nodular B-cell aggregates in the allograft biopsies were not part of the PTLD (Figure 2A). These infiltrates were polyclonal, EBV negative and showed no light chain restriction, while the PTLD specimen was clonal for the rearranged Ig-heavy chains and EBV positive. After rituximab application B-cell infiltrates were not detectable anymore in the transplant. At no time histological signs of FSGS recurrence were detectable and no biopsy showed podocyte fusion in electron microscopy.
Intragraft Gene Expression
In biopsies obtained at days 25 and 57 intragraft gene expression was determined for GAPDH, fas-ligand, granzyme B, nuclear factor κB, Foxp3 and OX40. In addition these genes were evaluated in a native kidney biopsy obtained 2 years earlier verifying FSGS. Harvesting of biopsy tissue, RNA isolation and reverse transcription was performed as previously described (3). Transcripts were quantified by Taqman PCR using the ΔΔCT-Method and GAPDH as a housekeeping gene. Data show high gene expression levels for cytotoxic lymphocyte effectors granzyme B and fas-ligand and T-cell activation genes OX40 and nuclear factor κB in this vascular cellular AR within the range seen in other cellular AR biopsies (unpublished data). These transcription rates decline substantially (30–50%) after histopathological resolution of the rejection process by rituximab. In contrast the immunregulatory gene Foxp3 upregulates its previously low transcription with resolution of rejection. In the previous biopsy of the boy's native kidney with FSGS very high levels of Foxp3 and NF-κB were observed with almost undetectable OX40, fas-ligand and granzyme B-gene expression supporting the histological diagnosis that there is no recurrence of FSGS.
Graft loss due to irreversible AR is becoming an increasingly rare event in renal transplantation today and is especially uncommon in living related recipients without prior donor transfusions and no panel reactive antibodies.
AR under most circumstances is regarded as a T-cell mediated event. Currently immunosuppressive protocols are designed to specifically inhibit this alloreactive T-cell response with few drugs as steroids, azthioprine, mycophenolate and rapamycin aiming at a broader anti-proliferative response.
In our patient intensification of anti-T-cell treatment by steroid pulses, tacrolimus rescue and anti-lymphocyte antibodies failed to stop AR. Pathophysiological pathways of C4d negative AR episodes refractory to anti-T-cell treatment are poorly understood. UNOS data indicate that those patients developing AR despite current immunosuppression are unlikely to resolve their immunological injury completely and that they carry a 5-fold higher risk of chronic allograft nephropathy (4). Therefore, the character of AR episodes evolving despite today's T-cell focused immunosuppression in fact may differ from AR episodes seen 15–20 years ago.
B cells have not been implicated in bioptic readout of AR after RTX. B-cell infiltrates are not recognized in the current Banff classification of AR, but graft injury potentially mediated by B cell is rapidly gaining focus.
Our patient did not demonstrate any signs of humoral AR and both preformed antibodies and C4d remained negative at any time. Sarwal et al. have demonstrated that the density of CD20+ B cells detectable in glucocorticoid resistant AR strongly correlates with poor outcome (78–89% graft loss) and that gene arrays of biopsies in these patients show ‘an overriding signature of B-cells’ (1). Kerjaschki et al. reported that nodular infiltrates with KI67+CD20+ B cells among dendritic and T cells in rejecting grafts are associated with lymphatic neoangenesis and accumulation of CCR7+ cells. They claim that ‘by virtue of their cellular composition, nodular infiltrates have the potential to launch and perpetuate specific immune responses to graft alloantigens and thus could contribute to recurrent episodes of AR, support humoral rejection, pave the way for chronic rejection and eventual loss of transplant function….’ (5). Nine of 10 graft losses occurred in patients with such infiltrates. However, graft failure developed significantly later than in our patient. Hippen et al. looked at the long-term impact of CD20+ infiltrates in AR on graft function and demonstrated that patients with CD20+ infiltrates were more likely to have steroid resistant rejection and a significantly reduced graft survival within 48 months (6).
The efficacy of rituximab in treatment refractory AR has evolved in recent years. The sequence of biopsies obtained in our patient for the first time identified nodular B-cell clusters in all three specimens with ongoing AR and showed resolution of these infiltrates after rituximab. This implies that CD20+ cells actually disappear and do not internalize the CD20 antigen. Destruction of CD20+ cells by rituximab, not internalization of the receptor, has been shown as the mechanism of action in lymphoma patients (7). Except for the misinterpretation due to a sampling error, this goes along with the data, that rituximab directly inhibits B-cell proliferation and induces cellular apoptosis by inhibiting bcl-2 (8). Without rituximab the clinical course of our patient may well have resulted in graft loss.
However, pathophysiological pathways of a B-cell facilitated nonhumoral graft injury remain highly speculative. Our gene expression data show enhanced transcription of T-cell genes transcription factor NF-κb, the B-T cell costimulation sustaining molecule OX40 and CTL effector genes granzyme B and fas-ligand in AR. This gene expression is down-regulated after rituximab. OX40, enhanced by B7-CD28 costimulatory binding, is a regulatory gene sustaining T-cell dependant B-cell proliferation. OX40, present on activated CD4+ T cells, may be one of the crosstalk links between T- and B-cell-driven graft destruction. In fact, OX40+CD4-cells are detectable in nodular B-cell formations similar to those present in our patient and OX40L on vascular endothelial cells has been implicated initiating vascular injury (9).
Rivera et al. showed in mice that antigen-specific B cells are essential for systemic T-cell responses to low antigen concentrations (10). It could be argued that graft infiltrating B cells may function as efficient antigen-presenting cells for indirect allorecognition and provide repetitive costimulation. Demirci et al. proved that the OX40-OX40L pathway is capable to restore the ability to mount rejection even if CD28 and CD154 pathways are blocked (11). Dynamics and impact of Foxp3 gene activation in regulatory T cells controlling development and resolution of AR are less well investigated. Baan et al. very recently showed a decline of transcription and heightened expression of Foxp3 in successful treatment of cardiac allograft rejection (12). Expression in our patient seems to follow a similar line. The effect of rituximab on OX40-OX40L or Foxp3 pathways is unknown.
Other causes of graft injury to be considered are highly unlikely to have set off the loss of function in our patient. FSGS remained undetectable in light and electron microscopy. Continued sampling errors are possible, but no gross proteinuria occurred and gene expression data differed from native FSGS in the same patient and previously published data (13). Early intragraft PTLD abolished by single-shot rituximab may be considered. However, the immunohistological signature of B-cell infiltrates differs substantially from later PTLD. The graft never showed enhanced glucose signaling in PET and EBV PCR was negative with low lactate dehydrogenase activity.
Combining current information nodular B-cell clusters in predominantly vascular AR, CD20+ infiltrates, anti-T-cell treatment irresponsiveness, long-term efficacy of rituximab and a gene expression pattern to be verified characterize a form of graft destruction currently not well defined.
In AR refractory to anti-T-cell treatment rituximab may be considered when infiltrating CD20+ B-cell clusters are detectable and according to the literature data may be especially beneficial in vascular AR. Narrowing prophylactic immunosuppression to modern drugs specifically inhibiting T-cell activation may in fact facilitate B-cell activation finally contributing to previously unrecognized forms of graft injury and development of PTLD.
In conclusion our case demonstrates that even late anti-B-cell treatment can result in reversal of longstanding AR unresponsive to anti-T-cell therapy. Nodular B-cell clusters were an abundant feature during ongoing rejection and resolved only after reversal by rituximab. Diagnostic criteria for such refractory AR episodes may differ from anti-T-cell treatment reactive AR. Options are recognition of CD20+ B-cell clusters in histopathology and to detect altered gene expression profiles (namely OX40 and Foxp3).
Extended evaluation of potential pathways in treatment of refractory graft injury is mandatory to understand the potential role of nodular B-cell clusters. Prospective studies are needed to clarify the correlation of B-cell nodules with refractory AR. This might lead to a further refinement of the Banff-classification adding B-cell facilitated AR rejection to T-cell and antibody-mediated rejection.