UK Guideline on Management of Bk Polyomavirus (BKPyV) Infection and Disease Following Kidney Transplantation

Research recommendations:

1. UK-wide multicentre randomised controlled trials are strongly recommended.
2. Trial strategy to harmonise immunosuppression including dose reduction strategy.
3. In those who develop BKPyV infection or disease following solid organ transplantation; is introduction of tacrolimus/sirolimus or replacement with sirolimus associated with better outcomes?
4. Assess the trajectory of response using current regimens including establishing the frequency of monitoring.

Audit Measures:

1. Current centre-specific practice of immunosuppression reduction against local guidelines.
2. Where variation in local practice for immunosuppression reduction exists, audit effect on BKVAN outcomes.
3. Frequency of graft loss in spite of immunosuppression reduction.

Rationale

Immune response to BKPyV infection

Immune responses to BKPyV infection, through both the innate and adaptive immune system are critical for control of viral replication and the prevention of pathogenesis.

The innate immune system provides the first line of defence for viral control. However, these early defence mechanisms are not always robust enough to clear viral infection [88]. Neutrophils perform part of the early response to viral infections but their role in fighting BKPyV remains unclear [89]. Likewise eosinophils may be recruited to the site of infection but their role has attracted limited research [90]. Macrophages are known to be recruited to BKPyV infection sites although the expression of aryl hydrocarbon receptor may indicate these cells take on a less inflammatory M2 phenotype [91].

Natural Killer (NK) cells form part of the innate response (although possessing some adaptive qualities) and recognise virally infected cells, triggering cytotoxicity to control infections. The tempo of NK cell response is dictated in part by signalling through killer-cell immunoglobulin-like receptors (KIRs; reviewed [92]) where lower expression of activating KIR3DS1 has been associated with severe BKVAN in kidney transplant recipients [93]. The BKPyV miRNA can suppress the expression of ULBP3, a stress-induced ligand recognised by the NK receptor “NKG2D”, reducing NK cell-mediated cytotoxicity [94]. Specific HLA loci may have a protective role when present kidney transplant donors and recipients however require further research [23].

An effective immune response requires a precise crosstalk among the various immune cells as an intercellular signalling network. These cytokines and chemokines control the magnitude and tempo of the immune response. Tumour necrosis factor alpha (TNFα) and interferon gamma (IFNγ) have been demonstrated as major cytokines for controlling the spread of the BKPyV in kidney-transplanted recipients with BKVAN [95]. However, TNFα has also been suggested to stimulate BKPyV replication in renal proximal tubular epithelial cells [96]. Both interferons alpha and gamma have been shown to restrict BKPyV replication in renal proximal epithelial cells [97,98] and for gamma this has been reproduced in urothelial cells [77].

Adaptive immunity

Dendritic cells (DCs) play a critical role in clearance of virally infected cells (reviewed [99]) and a reduced number of DCs have been observed in the peripheral blood of kidney transplant recipients with BKVAN [100]. It is likely that the reduction in DC leads to a weaker BKPyV adaptive response given DC are potent antigen-presenting cells important for the induction of antiviral cytotoxic T-cell responses. In mouse models, the magnitude of murine polyomavirus specific CD8+ T cell responses correlated directly with the amount of DC [101,102]. Interestingly, unlike other polyomaviruses, the VP1 protein of BKPyV does not induce DC maturation [103].

IgG seropositivity can reach 95% in the general population [10] and IgG seroreactivity tends to increase in those with active viral replication [104]. The presence of BKPyV-specific antibodies alone does not provide protection against BKPyV reactivation and its associated pathology [104]. The lack of neutralisation despite high antibody titre could be due to mutations in the viral capsid genes; as the main BKPyV genotypes can reciprocally escape neutralisation by antibodies raised against the other genotypes [2]. More studies focussed on anti-BKPyV neutralising antibodies are needed.

The role of the T cell response in clearance of BKPyV infection post kidney transplantation has been extensively reported [105,106]. There is a direct correlation between circulating levels of BKPyV-specific CD8+ T cells and effective immune control [107] suggesting a potential role for monitoring BKPyV-specific T cells in the peripheral blood of kidney transplant recipients to predict BKPyV-reactivation as done for CMV [108]. In the post-transplant period prompt recovery of BKPyV-specific T cells protects against reactivation where an insufficient BKPyV-specific T cell immunity is associated with continued viral replication and inflammation [105].

Both CD4+ and CD8+ T cell responses have a role in control of BKPyV infection [109]. Some reports have suggested that the CD4+ immune response plays a greater role in controlling BKPyV infection than CD8+ T cells [110] as evidenced by the ability of the CD4+ T helper cells to control BKPyV reactivation in the absence of a CD8+ response [95]. This CD4+ mediated control of BKPyV replication is through the expression of IFNγ, TNFα and granzyme B [95]. Polyomavirus miRNA activity has also been shown to have a role in mediating immune evasion by making virally infected cells less susceptible to recognition by cytotoxic T cells and reducing the production of cytokines by these cells [111].

BKPyV-specific T cell responses have been demonstrated against the BKPyV T antigens and capsid proteins in both seropositive healthy individuals and transplant patients [105,112–115]. By contrast, the BKPyV agnoprotein is abundantly expressed but largely ignored immunologically [116]. Data from kidney transplant recipients with active BKPyV replication suggested that CD8+ T cells were predominantly LT-Ag specific, while VP1 elicited a mainly CD4+ T cell response [117]. VP3 has been shown to be an antigenic target eliciting both CD4+ and CD8+ T cell responses [118]. Within the renal tissue itself the presence of BKPyV-specific T cells has been associated with both BKPyV viremia and BKVAN [119]. Critically, in the same study, renal biopsy specimens contained 7.8-times more alloreactive T cell clones than BKPyV LT-Ag-specific T cells [119].

At least 39 distinct CD4+ and CD8+ T cell epitopes for the BKPyV early viral gene region have been identified and there is some homology between BKPyV-specific and JCPyV-specific T cell epitopes [120]. Furthermore there has been reported overlap between BKPyV and JCPyV immune responses where prior JCPyV infection may induce T cell responses which are able cross-protect from BKPyV-associated diseases, and vice versa, due to the homology between polyomaviruses [121–123].

Immunosuppression Management in BKVAN

BKVAN is associated with over-immunosuppression, allowing re-emergence and replication of the dormant BKPyV. There are no randomised controlled trials addressing management approaches. Comparison of available evidence is limited by differing definitions of immunological risk, immunosuppressive regimes and diagnosis of BKVAN.

The first strategy in confirmed BKVAN is to reduce immunosuppression. This can lead to resolution of BKVAN in anywhere between 25% and 95% of cases depending on the definition of ‘cure’ used [51,75,76,124–126]. With increases in the number and proportion of highly sensitised patients on national kidney transplant waiting lists, and the advent of increasing transplantation across antibody barriers, immunosuppression reduction has become a more complex consideration, and may not be appropriate in all clinical situations e.g. where concurrent donor specific antibodies are detectable. This is particularly a problem for the paediatric and young adult population who almost inevitably will require re-transplant as demonstrated by registry data [127].

Some studies have suggested the use of specific immunosuppressive medications are associated with a higher risk of BKVAN and other studies have suggested that it is the overall burden of immunosuppression that is relevant [51,76,125,126,128,129].

Induction Therapies

Bayraktar et al performed a retrospective analysis of 257 adult patients at their single centre in Turkey who underwent transplantation between 2007 and 2017 stratified by induction therapy (none, ATG alone, ATG plus basiliximab) based on immunological risk [128]. No relationship was found between the incidence of BKVAN and the induction agent used (p>0.05). All patients received standard treatment with tacrolimus, MMF and corticosteroids.

Hassig et al reported no relationship between BKVAN and the use of basiliximab compared to ATG induction (p=0.212, [129]) as did Radtke et al [130] and Jacobi et al [131]. This is in contrast to Dadhania et al who reported ATG to be associated with an increased risk of BKVAN in comparison to basiliximab [40]. Saull et al compared the use of alemtuzumab and ATG as induction agents in primary non-sensitised renal transplant patients who underwent rapid steroid withdrawal (by day 5 post-transplant) [132]. All patients received tacrolimus and MMF maintenance therapy [132]. This single centre retrospective analysis of 200 consecutive patients showed that the one year rates of BKPyV DNAaemia (defined as BKPyV DNA copies/ml >5000) were identical between the two groups (p=0.36) but BKVAN (defined as positive immunohistochemistry staining for BKPyV on protocol biopsy) was more frequent in the alemtuzumab arm (p=0.046) [132]. The two UK based single centre trials of alemtuzumab induction in low immunological risk renal transplantation did not show any difference in BKPyV DNAaemia or BKVAN rates [133,134]. Another single centre retrospective review of 666 patients where 80% received alemtuzumab did not find a significant difference in the incidence of BKPyV DNAaemia or nephropathy in those who received depleting agents [135]. It was noted that BKPyV viraemia/BKVAN was observed frequently before an episode of biopsy-proven acute rejection which the authors attributed to a reduction in maintenance immunosuppression to treat BKPyV although they did not specify the regime they used [132].

Maintenance Immunosuppression

There have been multiple case reports and cohort studies looking at reduction of maintenance immunosuppression after diagnosis of BKPyV DNAaemia/BKVAN which is a generally accepted common first step in management of BKPyV [64]. However, there is no randomised controlled trial evidence to support reduction of maintenance immunosuppression. A meta-analysis based on the observational data available was only able to confirm reduction of maintenance immunosuppression as an effective strategy [136]. The efficacy of immunosuppression reduction is believed to be in allowing BKPyV-specific T-cells to control the viral reactivation [117,137].

Some authors [138] advocate a reduction in immunosuppression as soon as BKPyV viruria is confirmed, as this is noted to precede DNAaemia [76,139], and clinically evident prior to BKVAN by some weeks; but we do not recommend this due to the poor positive predictive power of viruria. Elfadawy et al adopted a strategy of only reducing immunosuppression in viraemia with BKPyV levels >10,000 copies/ml as they demonstrated spontaneous viral clearance in 95% of patients with lower levels [125], as did Sood et al [140] and Saad et al [141]. Schaub et al commented that viral clearance was more likely to be prolonged with higher viral loads [51]. Cases of BKPyV DNAaemia and BKVAN have been seen with the use of all six of the most commonly used immunosuppressive medications (steroids, CNIs, AZA/MMF and sirolimus) [76,126,142–144] but as they are often used in combination, it has proven challenging to determine if any single agent is more implicated than others.

Two papers have reported BKVAN occurring in the absence of CNI usage [144,145]. Baek et al noted an increased risk of biopsy-proven acute rejection on multivariate analysis if MMF was entirely discontinued or if CNI levels were reduced by more than 20% from baseline [146] but this is not consistently seen in all studies [76]. The observation of BKVAN occurring after an episode of biopsy-proven acute rejection is also common [143,146–148], presumably as a consequence of intensification of immunosuppression in this context.

Reduce Antiproliferatives First?

There is no consistency as to which agents are minimised or what levels are targeted [142]. Most authors [51,75,76,124,125,129,131,137,142,143,146–154] who follow the consensus recommendations proposed by Hirsch et al [59] have reported minimisation (by at least a 50% dose reduction) or complete cessation of MMF in the first instance with secondary reduction in CNI (calcineurin inhibitor) dose (targeting trough tacrolimus levels of <6 ng/ml and ciclosporin of <150 ng/ml). Gheith et al [149] reported resolution of viruria with this approach and targeted trough tacrolimus levels of 6 ng/mL. However, they did not report if this resolved BKVAN histologically. Abdeltawab et al [151] targeted trough tacrolimus levels of 3-6ng/ml and Brennan et al [76] suggested a similar level of 3-5 ng/ml. Jacobi et al [131] used a lower trough ciclosporin level of 60-80 ng/ml. Brennan et al [76] demonstrated no difference in development of BKPyV viraemia between patients receiving MMF or azathioprine in conjunction with CNI therapy.

Tacrolimus vs ciclosporin?

Ott et al supported a change from tacrolimus to ciclosporin after cessation of MMF but did not specify their target level [150]. A number of registry analyses and studies [50,126,154–156] also reported tacrolimus to be more strongly associated with the development of BKVAN but this is not consistent in all studies [76,130,131,147]. Gard et al observed higher BKPyV viral loads at some, but not all time points, in their study of tacrolimus treated patients in comparison to those on a ciclosporin based regime [147]. They also noted lower overall viral load with an earlier decline in titre in ciclosporin treated patients [147]. Multiple authors have shown that ciclosporin has some in vitro antiviral activity by preventing BKPyV replication but this has not yet been confirmed in vivo [117,137,157]. Hirsch et al have also demonstrated that tacrolimus may actually favour BKPyV replication in human tubular epithelial cells by the action of FK-binding protein 12 [50]. Other authors demonstrated a dose-dependent effect by tacrolimus to inhibit BKPyV-specific T-cells [117,137].

There is a single prospective study of patients with biopsy proven BKVAN receiving triple immunosuppression with tacrolimus, MPA and steroids being changed to ciclosporin [75]. In this study of 24 patients over a 2-year period, target trough ciclosporin levels were 75-125 ng/ml (higher than Jacobi et al who used 60-80 ng/ml [131]) whilst maintaining the same dose of other medications [75]. No control group was included, so no direct comparison of reduced dose tacrolimus vs reduced dose ciclosporin can be made, severely limiting the interpretation of the study. Nonetheless the authors were able to demonstrate that ciclosporin at lowered trough concentration enabled all patients clear BKPyV DNAaemia at a mean of 2.7 +/- 2.0 months, although viruria persisted in the majority [75]. This is a shorter time to BKPyV DNAaemia clearance than in other studies undertaking a similar approach, which may be due to varying definitions of the threshold used to define viral clearance [51,131,158]. The mean time to clear SV40 staining on biopsy was 10.7 months (range 8.2 – 21.6 months) and this occurred in 42% of patients; with tacrolimus reduction in the intensity of the SV40 staining noted in the remaining patients [75]. Creatinine was stable in all patients with no graft losses in the 24 month follow-up period [75]. Two patients experienced rejection but both were noted to have been non-compliant with their medications as prescribed [75].

Brennan et al carried out a prospective study of 200 adult patients with reduction in immunosuppression at the point of confirmation of BKPyV DNAaemia, achieved by discontinuation of azathioprine or MMF and reduction in CNI levels after four weeks if DNAaemia persisted [76]. They used slightly higher trough CNI levels than Hirsch et al (targeting tacrolimus 3-5ng/ml and ciclosporin 100-200ng/ml) but were able to demonstrate clearance of DNAaemia in 95% of patients by one year with a mean time to clearance of 54 days (range 7-213 days) [59,76]. There was no clear clinical difference in the tacrolimus vs ciclosporin groups, though more viruria was seen in the tacrolimus group. Only 21% of patients achieved clearance of BKPyV viruria [76]. Biopsies were only performed on a ‘for cause’ basis and hence may have missed some subclinical BKVAN but the authors suggested that given the resolution of DNAaemia and stable creatinine, this was unlikely [76].

Reduce CNIs First?

Hassig et al favoured minimisation of CNI and reduction or withdrawal of corticosteroids with secondary reduction in anti-metabolite therapy by 20-25% [129], as did Baek et al [146] and Bischof et al [62]. Ginervi et al suggested a CNI reduction of 15-20% [105]. It should be noted however that Baek et al noted a higher chance of  biopsy-proven acute rejection if CNI levels were reduced by greater than 20% [146]. This did not differ if tacrolimus or ciclosporin was used. Hassig et al did not specifically report resolution rates with CNI minimisation but described no differences in the incidence of BKPyV viraemia/BKVAN between ciclosporin and tacrolimus based regimes although they did comment that high cumulative doses of corticosteroids in the ciclosporin group may have masked any difference between the two, noting that Hirsch et al and Kim et al reported an association of BKVAN with cumulative corticosteroid exposure [50,129,159].

The combination of tacrolimus and MMF was reported by Mengel et al to provide a 13 times higher odds ratio of the development of BKVAN in comparison to those regimes using only one of these agents [143]. This may be due to the observation that systemic exposure to the active metabolite of MMF, mycophenolic acid (MPA), is up to two-fold higher in tacrolimus compared to ciclosporin treated patients due to ciclosporin’s inhibition of the enterohepatic recirculation of the major metabolite, MPA glucuronide [160]. Some authors have found an association with use of MMF [125] but others have not [76].

mTOR Inhibitor (mTORi) Use

Some authors support a change from an antiproliferative agent to an mTORi (in conjunction with ciclosporin) if viral loads do not respond to immunosuppression reduction [131,153,155,161,162]. This approach is based on a registry analysis by Dharnidharka et al who found that incidence of BKVAN was substantially lower if mTORi were used [155]. This observation has been confirmed in small observational studies [131,161–164]. Hirsch et al recommended trough sirolimus levels of <6 ng/ml [59]. Jacobi et al used trough mTORi levels of 5-8 ng/ml [131]. Sánchez Fructuoso et al used either sirolimus or everolimus but did not specify the target levels used [161]. Tedesco-Silav et al reported a lower incidence of BKPyV viraemia when using higher levels of everolimus (6-12 vs. 3-8 ng/ml) [165]. Polanco et al [164] and Al-Raisi [124] found benefit of converting CNI to mTORi and continuing MMF.

mTORi are known to have antiviral properties against CMV and hepatitis C [137,166]. It has been proposed that mTORi are advantageous in BKVAN as they do not appear to limit the effectiveness of BKPyV-specific T cells in vitro [137]. An animal model showed that mTORi can increase the antiviral memory of CD8 T-cell differentiation [166]. Conversely, a case report by Hirsch et al described a patient who only cleared BKPyV viraemia after sirolimus was discontinued [167]. However, it is important to note that the patient was highly immunosuppressed with trough ciclosporin levels of 242 ng/ml and 15 mg prednisolone daily and no attempt was made in this case to minimise these first [167]. Radtke et al also demonstrated no beneficial effect of using everolimus for the risk of developing BKPyV DNAaemia [130].

Corticosteroids

High dose (cumulative intravenous methylprednisolone of ≥2 g/month) has been associated with decline in graft function when managing BKPyV DNAaemia [159]. However, there are very few studies looking directly at the specific effects of modulating the steroid component of immunosuppression. Improved outcomes have been reported by reducing other components of immunosuppression whilst maintaining prednisone at 5-10 mg/day [168]. Patients that have inflammation as a result of rejection may explain reports of improvement with steroid use.

In the setting of tubulitis escalated doses of steroids are used but the evidence is still equivocal as to their benefit [169].