Everolimus leads to a lower risk of BKV viremia than mycophenolic acid in de novo renal transplantation patients: a single-center experience

Background: There are limited published data concerning the effects of different immunosuppressive regimens on the development of polyomavirus (BKV) viremia. We examined the risk of developing BKV viremia in kidney transplant recipients receiving everolimus (EVR) or mycophenolic acid (MPA) as maintenance therapy. Methods: We observationally analyzed 296 patients who underwent renal transplantation at our center between 2005 and 2010: 58 were treated with EVR and low‐dose cyclosporine (LD‐CyA) (group 1) and 238 with MPA and standard‐dose CyA (group 2). All of the patients received induction therapy with basiliximab and maintenance steroids. BKV viremia (a whole‐blood viral load of >850 copies/mL) was measured by means of real‐time polymerase chain reaction at least once a month during a 12‐month follow‐up period. Results: BKV viremia was detected in 57 patients (19%), five (9%) in group 1 and 52 (22%) in group 2. Kaplan–Meier analyses showed that freedom from BKV viremia was significantly more frequent in group 1. The mean time of onset of BKV viremia was about four months after transplantation in both groups. The median viral load was greater in group 2 (12.5 ± 6.1 vs. 2.5 ± 1.8 × 104 copies/mL; p = 0.01). After the onset of BKV viremia, graft function significantly declined in group 2: 11 patients developed polyomavirus‐associated nephropathy (PVAN) and four presumptive PVAN; nine experienced an acute rejection after the discontinuation of MPA, and 11 (21%) lost their graft. There was no graft loss in group 1. Conclusion: These findings suggest that in comparison with MPA and Cya, an EVR and LD‐CyA regimen lowers the risk of BKV viremia after kidney transplantation and favorably alters outcomes.

BK virus; everolimus; interstitial nephritis; kidney transplantation; mycophenolic acid; polyomavirus

Polyomavirus (BKV) is a major cause of graft loss in renal transplant recipients (RTRs) [1] , [2] , with interstitial nephritis being the most represented histological lesion. The reported incidence of polyomavirus‐associated nephropathy (PVAN) in RTRs varies from 1% to 10% [2] , and although it has reduced the occurrence of acute rejection (AR), the introduction of more potent immunosuppressive agents such as tacrolimus (TAC) and mycophenolic acid (MPA) has been associated with the development of PVAN [1] . The diagnosis of PVAN is based on renal biopsy findings [2] but, as it may present as a focal disease, the false‐negative rate may be as high as 10–30%. BKV replication in the graft correlates with blood BKV levels [3] , [4] , and plasma polymerase chain reaction (PCR) is a sensitive and specific means of detecting BKV reactivation early [5] .

It has been reported that promptly reducing maintenance immunosuppression is safe and effective in RTRs with BKV viremia, but it may increase the risk of AR and predispose to chronic allograft nephropathy [2] . Antiviral agents such as cidofovir and other drugs with antiviral activity such as leflunomide, fluoroquinolones, and intravenous immunoglobulin have been increasingly studied, but their efficacy has not been confirmed. The effects of different immunosuppressive regimens on the activation of BKV have not been investigated in detail, and some authors have reported contradictory results in patients treated with the mammalian target of rapamycin inhibitors (mTORi) sirolimus (SRL) or everolimus (EVR). Hirsch et al. [6] described a case in which the discontinuation of sirolimus led to the clearance of BKV, whereas Wali [7] and Benavides [8] observed a reduction in the incidence of PVAN in RTRs treated with sirolimus. In a pooled analysis of three randomized clinical trials involving more than 2000 RTRs, Brennan et al. [9] found a significantly lower incidence of cytomegalovirus (CMV) infection in patients treated with EVR than in those treated with MPA.

The aim of this study was to evaluate the relationship between exposure to EVR and the risk of BKV viremia after kidney transplantation.

This was a single‐center, follow‐up observational cohort study of 319 patients who underwent kidney transplantation our institution between January 2005 and December 2010. The patients were followed up by our Renal Unit, with the visits being scheduled in accordance with the American Society of Nephrology guidelines [10] . All of the clinical and laboratory data concerning the RTRs were routinely recorded.

Immunosuppression was induced by EVR plus low‐dose cyclosporine (LD‐CyA) in 66 patients (group 1) and by MPA plus standard‐dose CyA in 253 (group 2). All of the patients in both groups received induction therapy with the anti‐CD25 monoclonal antibody basiliximab (Simulect, Novartis, Basel, Switzerland) and maintenance steroids.

Fifty‐three of the 66 patients in group 1 were given EVR in two clinical trials (18 in the Novartis CRAD001AIT12 “EVIDENCE” trial and 35 in the Novartis CRAD001AIT02 “EVEREST” trial), four because of a history of malignancy, and nine as a newly implemented standard therapy. EVR (Certican, Novartis) was started at 0.75 mg/bid/d, targeting trough blood levels of 3–8 or 8–12 ng/mL. CyA (Sandimmun Neoral, Novartis) was started at a dose of 2 or 4 mg/kg/d in two administrations, targeting a trough blood level of 100–150 ng/mL for the first month, which was then reduced to 50–100 ng/mL. Oral methylprednisolone was started at 16 mg/d in two administrations for the first month after transplantation and then tapered to 4 mg/d.

The 253 patients in group 2 started CyA at a dose of 6 mg/kg/d in two administrations, targeting trough blood levels of 150–300 ng/mL until month 6 and 70–150 ng/mL thereafter. MPA was administered as mycophenolate mofetil (Cell Cept; Roche, Basel, Switzerland) 2 g/d or mycophenolate sodium (Myfortic, Novartis) 1.44 g/d in two administrations. Oral methylprednisolone was started at 16 mg/d in two administrations and tapered to 8 mg/d after three months.

If BKV viremia was detected, maintenance immunosuppression was not changed in group 1, whereas MPA was discontinued in group 2.

The standard of care was a renal biopsy for graft dysfunction, which was defined as a >20% increase in serum creatinine levels. Acute rejection was assessed on the basis of Banff's classification [11] and treated with pulsed steroids; in the case of steroid resistance, antithymocyte globulin (ATG) was used.

To detect BKV DNA in blood, weekly real‐time PCRs were performed from month 1 to month 4 and monthly PCRs thereafter using the Invitrogen Platinum Quantitative PCR SuperMix‐UDG (Invitrogen Corporation, Carlsbad, CA, USA) and the following amplification profile: 2 min at 50°C and then 2 min at 95°C, followed by 40 cycles at 95°C for 8 s and 60°C for 34 s. The real‐time assay used a total reaction volume of 15 μL, of which 6 μL was used for DNA extraction. Pathological BKV viremia was defined as a blood viral load of >850 copies/mL. Definite PVAN was diagnosed by means of biopsy specimens (two cores obtained with a 16‐gauge needle) that were examined using light microscopy and immunohistochemistry (SV40 large T‐antigen). All of the graft biopsies contained tissue from the cortex and the medulla, and the findings were graded using the classification of Drachenberg et al. [12] . The patients with a BKV viremia load of >104 copies/mL of plasma for more than three wk but a negative biopsy, who were asymptomatic or had increased serum creatinine levels, were diagnosed as having presumptive PVAN.

As part of standard prophylaxis, the patients received a third‐generation cephalosporin intra‐operatively, cotrimoxazole 400 mg daily for six months, and oral nystatin for three months. From May 2006, three months' universal CMV prophylaxis with aciclovir was replaced by three months' valganciclovir given only to donor‐positive/recipient‐negative cases or those treated with AGT. A ureteral JJ catheter was placed at the time of transplantation and removed after one month.

Circulating lymphocyte subpopulations were determined by means of multiparameter flow cytometry. Peripheral blood was drawn on ethylenediaminetetraacetic acid (EDTA) and processed on the same day using a no‐wash, no‐lyse protocol: 100 μL of whole blood was labeled with fluorochrome‐conjugated monoclonal antibody combinations for 20 min at 4°C in the dark, followed by red blood cell lysis (Immunoprep and TQ. Prep; Beckman Coulter, Miami, FL, USA). The data were acquired and analyzed using a Cytomics FC500 (Beckman Coulter) instrument and CXP software.

Peripheral blood was drawn on EDTA and processed on the same day. High‐resolution agarose gel electrophoresis was used to define the subsets of serum proteins using an autoanalyzer (Hitachi, ATM, Manheim, Germany). The samples were assayed in duplicate on one d by the same operator.

Peripheral blood was drawn on EDTA and processed on the same day. The immunoglobulin classes were measured by nephelometry using reagents and an automated system (Siemens, Deerfield, IL, USA). The protein fraction and immunoglobulin analyses led to the exclusion of 26 patients from group 1 and 136 from group 2 who showed signs of concomitant non‐BKV‐related infection (i.e., EBV PCR, HCVAb or HBsAg positivity, and patients with urinary infection).

The data were statistically analyzed using the NCSS system (NCSS, LLC, Kaysville, UT, USA) and are expressed as mean values ± SD or median values and range, as appropriate. The correlations between the clinical parameters and BKV viremia were evaluated using Student's t‐test. The percentage of RTRs in each group who were BKV viremia free 12 months after transplantation was determined by means of Kaplan–Meier survival analysis, and the survival curves were compared using the log‐rank test. Cox's proportional hazard regression was used to estimate the risk of BKV viremia in the two groups. A p value of <0.05 was considered statistically significant.

A total of 319 RTRs were considered. Of the original 66 patients in group 1, three experienced early graft loss (one case of hemolytic uremic syndrome and two early deaths due to cardiac infarctions), and five were excluded because of reduced immunosuppression due to severe pulmonary infections; the final analysis therefore included 58 patients. Of the 253 patients originally enrolled in group 2, two experienced early graft loss (renal vein thromboses) and one primary non‐function, two were switched to other therapies because of leukopenia, and 10 were excluded because of pulmonary infection. The final analysis therefore included 238 patients (Fig. [NaN] ).

The two groups had similar baseline demographic and transplantation characteristics (Tables [NaN] and [NaN] ). The mean follow‐up was 17 ± 5.5 months in group 1 and 22 ± 8 months in group 2 (p = 0.73).

Demographic data and transplantation characteristics

1 No significant difference in the clinical characteristics of the two groups.

2 RT, renal transplantations; Tx, transplantation; BPAR, biopsy‐proven acute rejection; BKV, polyomavirus; MPD, methylprednisolone; ATG, antithymocyte globulin; CMV, cytomegalovirus; BD, before diagnosis.

Clinical course of transplantation before and after the diagnosis of BKV viremia. BKV‐related outcomes

• 3 In group 2, graft function was significantly worse at diagnosis of PVAN, BPAR occurred in nine patients, the mean viral load was significantly higher, and 11 patients (21%) lost their graft. In group 1, BKV viremia stopped one month after diagnosis.
• 4 PVAN, polyomavirus‐associated nephropathy; sC, serum creatinine; CCr, creatinine clearance; BPAR, biopsy‐proven acute rejection; BKV, polyomavirus.

The frequency of BKV viremia was significantly higher in group 2 (52.22% vs. 5.9%; p = 0.01): the two‐sided 95% confidence interval was 1.6–16% in group 1 and 17–27% in group 2, thus indicating that the sample sizes were sufficiently large to detect a difference between the two groups and provided a power of 82.4% to reject the null hypothesis of equal rates of BKV viremia. Kaplan‐Meier analyses showed that freedom from BKV viremia was significantly more frequent in group 1 (Fig. [NaN] ). The adjusted hazard ratio (HR) from the Cox model showed that group 2 was associated with a significantly increased risk of BKV viremia (HR = 1.71, 95% CI: 1.08–2.69; p = 0.02). The mean time to the onset of BKV viremia was 4.1 ± 1.5 months in group 1 and 3.8 ± 1 months in group 2 (p = 0.25), and the mean viral loads were, respectively, 2.5 ± 1.8 × 104 and 12.5 ± 6.1 × 104 copies/mL (p < 0.001) (Table [NaN] ). The time to the clearance of viremia was shorter in group 1 (Table [NaN] ). Clearance was obtained in all of the patients in group 1, but albeit lower levels of viremia persisted during the follow‐up in 15 group 2 patients (29%) (data not shown).

In comparison with the data recorded before the onset of BKV viremia, graft function was fairly stable at the time of diagnosis in group 1 (creatinine clearance [CCr] −0.9 mL/min) but reduced in group 2 (−17.4 mL/min; p = 0.02) (Table [NaN] ).

None of the affected patients in group 1 developed PVAN during the follow‐up. In group 2, the 15 cases of PVAN were classified as type A (7, 13%), type B (4, 8%), and presumptive PVAN (4, 8%) (Table [NaN] ) using the criteria of Drachenberg et al. [12] .

The rates of biopsy‐proven acute rejection (BPAR) before the onset of BKV viremia were 10.3% in group 1 and 10.1% in group 2 (p = 0.37); the histological grading [11] was Banff I in 24 patients (treated with steroid pulses) and Banff II–III in six treated with steroid pulses and ATG. New ARs occurred in nine group 2 patients after the discontinuation of MPA, as against none in group 1 throughout the follow‐up (Table [NaN] ). Eleven group 2 patients (21%) lost their grafts a mean of 16.5 months after the diagnosis of BKV viremia (range 10–39): five because of PVAN and six because of rejection. None of the affected group 1 patients lost a graft during the follow‐up (Table [NaN] ).

Blood trough levels of immunosuppressive drugs and circulating lymphocyte counts at diagnosis and six and 12 months after transplantation

• 5 Mean values ± SD.
• 6 CyA, cyclosporine; EVR, everolimus; MPA, mycophenolic acid; WBC, white blood cells.
• 7 For the calculation, mycophenolate sodium was transformed into equimolar doses of mycophenolate mofetil by multiplying it by 1.39.

When BKV viremia was detected, mean WBC, lymphocyte, total T‐cell, and CD4+ and CD8+ cell counts were below the normal range and similar in the two groups. In group 1, the mean lymphocyte, total T‐cell, and CD4+ and CD8+ cell counts remained low during the follow‐up, but all of these parameters increased in the affected patients in group 2 after the discontinuation of MPA. Mean CD19+ cell counts were normal in group 1, but below normal in group 2, although they increased after MPA discontinuation (Table [NaN] ).

Mean γ‐globulin concentrations were lower in group 2 than in group 1 at the time BKV viremia was diagnosed. After the discontinuation of MPA, they significantly increased in the BKV‐positive patients in group but remained below normal range in the BKV‐negative patients. This effect persisted even six months after transplantation and was not affected by the tapering of CyA administration (Table [NaN] ).

Mean IgG, IgA, and IgM levels were normal in group 1 and below normal in group 2. The appearance of BKV viremia was associated with a temporary increase in IgM only in group 1 that returned to normal after about three months. The withdrawal of MPA was followed by an increase in immunoglobulin levels toward the normal range in group 2 (Table [NaN] ).

Mean immunoglobulin and total γ‐globulin concentrations at diagnosis and six and 12 months after transplantation

• 8 Mean values ± SD.
• 9 BKV+ = affected patient; BKV− = unaffected patient.

A number of studies have found that mTOR inhibitors have an antiviral effect and preserve protective immunity [13] , [14] . EVR is a recently developed mTOR inhibitor that regulates cell protein synthesis [15] , is relatively devoid of nephrotoxicity [16] , inhibits neointimal and smooth muscle cell proliferation, and reduces fibrogenesis after the onset of graft damage [17] . All viruses are heavily dependent on cell protein synthesis for constituent proteins and genomic replication. Polyomavirus triggers host intracellular signaling and DNA replication [18] and activates the protein kinase Akt/mTOR pathway to facilitate its own replication. In a series of in vitro experiments using BKV‐infected renal epithelial cell lines, Liancini et al. [19] found that sirolimus can dose dependently reduce BKV large T‐antigen expression and its genome replication within a few days.

In line with previous findings [12] , our study shows a 19% incidence of BK viremia in a cohort of kidney transplant recipients followed up for 22 ± 8 months and that immunosuppressive therapy based on EVR and LD‐CyA is associated with a much lower incidence of BKV viremia than that associated with an MPA/CyA regimen (9% vs. 22%; p < 0.01). The reduced risk of BKV viremia does not seem to be a counterpart of less efficient immunosuppression because pre‐infection BPAR rates and T‐cell subset levels at the time of diagnosis (Table [NaN] ) were similar in the two groups.

The time to the onset of BKV viremia was similar in the two groups, but viral loads were higher in group 2 and the time to viral clearance was longer (Table [NaN] ). This can be explained by the mechanism of action of MPA, which is a reversible inhibitor of the enzyme inosine monophosphate dehydrogenase that catalyzes the formation of guanosine nucleotides from inosine and has relatively selective antiproliferative effects on lymphocytes. In vitro, MPA blocks the proliferation of T and B cells and inhibits the generation of cytotoxic T cells and antibody formation [20] . These effects are consistent with our findings of reduced CD4+, CD8+, and CD19+ cell counts and mean γ‐globulin and immunoglobulin concentrations (Tables [NaN] and [NaN] ). The immune system not only takes longer to clear BKV after MPA withdrawal, but in one‐third of patients, it does so incompletely and BKV viremia persists albeit at lower levels (data not shown).

At the time of the onset of BKV viremia, there was no graft function impairment in group 1, and no graft loss was observed during the follow‐up (Table [NaN] ); in group 2, renal function declined significantly and 11 patients (21%) lost their grafts.

After the appearance of BKV viremia, no BPAR occurred in group 1, but there were nine (17.3%) in group 2. These are the main findings of our study because there is no optimal strategy for the treatment of BKV viremia. It has been reported that a prompt reduction in immunosuppression is safe and effective in RTRs with BKV viremia, using strategies ranging from a reduction in or discontinuation of antiproliferative drugs (MPA or azathioprine) to a reduction in the calcineurin inhibitor (CyA or TAC), a 25–50% reduction in total immunosuppression or a switch from TAC to CyA [1] , [21] , [22] , [23] .

The adaptive immune response to an infective agent acts by means of humoral and cellular mechanisms. In the case of BKV infection, the humoral response increases IgG, IgM, and IgA levels, including neutralizing and subtype‐specific antibodies against the determinants of the BKV major capsid protein VP1 [24] . Randhhawa et al. found a correlation between the increase in IgG, IgM, and IgA levels and the onset of BKV viremia in defined categories of patients [25] , but little is known about the effects of different immunosuppressive drugs on humoral responses [26] .

The findings of studies of experimental animals [27] and human cells in vitro (30) suggest that MPA may inhibit antibody production [13] , and MPA‐induced suppression of the humoral response may be accompanied by a greater incidence of infections or more severe infections than those complicating previous immunosuppressive therapies. Our results suggest that MPA markedly decreased humoral responses and reduced immunoglobulin and γ‐globulin concentrations and CD19+ cell counts in the presence of BKV viremia, because these parameters remained normal in our group 1 patients. Another finding supporting this hypothesis is that immunoglobulin and γ‐globulin concentrations and CD19+ cell counts were not affected by per‐protocol CyA tapering and normalized only after the discontinuation of MPA (Tables [NaN] and [NaN] ). Unfortunately, it is not possible to test any difference in BK viremia onset between patients in group 1 with targeted EVR level of 3–8 compared with 8–12 ng/mL because the blood levels of EVR tended toward the higher limit of the range in the former and the lower limit of the range in the latter, so that no difference in EVR concentration at 6 months (p = 0.27) and at 12 months (p = 0.32) was reported by Salvadori et al. [28] .

The limitations of our study include its retrospective design, the small number of patients, and the fact that most of the patients in the EVR group were enrolled in two clinical trials. However, it has to be noted that EVR was always associated with a reduced dose of CyA and that the entry criteria were identical and inclusive, and so, their baseline characteristics were no different from those of the patients outside the protocols. Nevertheless, even if the screening and follow‐up were the same, patients participating in a clinical trial are generally followed more closely and are more compliant, and so, the outcome is usually in favor of this group (the “Hawthorne effect”).

Another bias is the different daily dose of steroids one month after transplantation (4 mg in group 1 and 8 mg in group 2). Hirsch et al. [29] have recently identified greater steroid exposure as an independent risk factor for high‐level viruria and viremia.

Finally, the determination of specific BKV IgG and IgM rather than total immunoglobulin subclasses would have allowed a more precise interpretation of their variations, but in an attempt to limit this bias, we excluded the patients with signs of non‐BKV infection from the analysis.

In conclusion, an EVR and LD‐CyA regimen is associated with a lower risk of BKV viremia than an MPA/CyA regimen and, at least in our study, prevented the evolution toward PVAN. Together with the reported inhibition of viral replication by the mTOR inhibitor, our data suggest that MPA increases the risk of BKV viremia by decreasing adaptive immune (humoral and cellular) responses to the virus. The risk of acute rejection after the discontinuation of MPA observed in our study warns against this therapy in patients on an MPA/CyA regimen who develop BKV viremia. The efficacy and safety of switching from MPA to EVR in this setting should be addressed in ad hoc trials, and larger studies are needed to confirm our results and allow a better understanding of the relationship between BKV viremia and immunosuppressive therapy.

Graph: Patient selection and disposition.

Graph: Kaplan–Meier plot of polyomavirus viremia‐free patients in the two groups.