Pediatric Nephrology

, Volume 25, Issue 5, pp 965–969

Tacrolimus nephrotoxicity: beware of the association of diarrhea, drug interaction and pharmacogenetics

Authors

  • Sandrine Leroy
    • Department of Pediatric NephrologyArmand-Trousseau Hospital, AP-HP and University Paris VI
  • Arnaud Isapof
    • Department of Pediatric NephrologyArmand-Trousseau Hospital, AP-HP and University Paris VI
  • Sonia Fargue
    • Department of BiologyUniversity College London
    • Department of Pediatric Nephrology—Reference Centre for Rare Renal DiseasesFemme Mère Enfant Hospital, University of Lyon
  • May Fakhoury
    • Department of Pediatric Pharmacology—Pharmacogenetics and Clinical Investigation Center—CIC9202Robert-Debré Hospital, AP-HP
  • Albert Bensman
    • Department of Pediatric NephrologyArmand-Trousseau Hospital, AP-HP and University Paris VI
  • Georges Deschênes
    • Department of Pediatric NephrologyRobert-Debré Hospital, AP-HP
  • Evelyne Jacqz-Aigrain
    • Department of Pediatric Pharmacology—Pharmacogenetics and Clinical Investigation Center—CIC9202Robert-Debré Hospital, AP-HP
    • Department of Pediatric NephrologyArmand-Trousseau Hospital, AP-HP and University Paris VI
    • Inserm UMR S938Université Pierre et Marie Curie
Brief Report

DOI: 10.1007/s00467-009-1402-8

Cite this article as:
Leroy, S., Isapof, A., Fargue, S. et al. Pediatr Nephrol (2010) 25: 965. doi:10.1007/s00467-009-1402-8
  • 464 Views

Abstract

Tacrolimus is known to potentially lead to adverse events in recipients with diarrhoea and/or calcium channel blocker (CCB) co-administration. We report a renal transplant recipient who suffered from severe nephrotoxicity related to a toxic tacrolimus trough concentration in both conditions, diarrhoea and CCB co-administration, and with genotyped CYP3A system and P-glycoprotein (P-gp) polymorphisms. To our knowledge, this is the first case to be investigated for such polymorphisms. Clinicians should be reminded of the possibility of highly increased levels of tacrolimus in situations of diarrhoea and/or co-administration of CCBs. It also highlights the key role in tacrolimus pharmacokinetics of the CYP3A system and P-gp polymorphisms, and their influence in high-risk situations when enzyme activity is already affected by enterocyte damage due to diarrhoea and CCB competition.

Keywords

Calcium channel blockerCytochrome P450DiarrhoeaP-glycoproteinPolymorphismsRenal transplantTacrolimus

Abbreviations

CCB

Calcium channel blocker

CYP

Cytochrome P450

MDR1

Multi-drug resistance 1

P-gp

P-glycoprotein

SNP

Single nucleotide polymorphism

Introduction

Tacrolimus is a calcineurin inhibitor increasingly used for immunosuppression in renal transplantation [1]. Following oral administration, tacrolimus is preferentially absorbed in the duodenum and jejunum with highly variable pharmacokinetics [2]. Tacrolimus is a substrate for cytochrome P450 (CYP) 3A4 and 3A5 isoforms, and for P-glycoprotein (P-gp), a multi-drug efflux pump, all present at high levels in enterocytes [2]. Intestinal inflammation and infection influence tacrolimus bioavailability by damaging enterocytes and reducing metabolism [3]. Calcium channel blockers (CCBs), which are frequently administered in renal transplantation, also influence tacrolimus bioavailability as they are competitive substrates of the CYP3A system and P-gp [4]. Finally, single-nucleotide polymorphisms (SNPs) of genes encoding the CYP3A system and P-gp contribute to the highly inter-individual variable expression of the CYP3A system and P-gp, and the inconsistent pharmacokinetics of tacrolimus [5]. We report here, for the first time to the best of our knowledge, the combined effect of acute diarrhoea and co-administration of CCBs in a situation of P-gp and CYP3A system SNPs leading to a toxic tacrolimus trough blood level in a renal transplant recipient, which resulted in acute renal failure.

Case report

A 5-year-old Caucasian boy underwent a renal transplantation for end stage renal disease (ESRD) because of reflux nephropathy. For 9 years following transplantation, the patient had normal renal function (creatininaemia ∼100 µmol/L), with low-dose steroids, azathioprine (2 mg/kg/day), and tacrolimus (0.1 mg/kg/day) with trough levels within the desired therapeutic range (3–6 ng/mL). At the age of 14 years, he then presented with acute viral gastroenteritis without dehydration, and was hospitalized 1 week later with acute renal failure (creatininaemia: 264 µmol/L) without fluid overload. Tacrolimus trough level was 27.2 ng/mL; renal biopsy only showed histological signs of chronic graft rejection combined with signs of chronic calcineurin inhibitor toxicity, without tubular necrosis signs. Duplex ultrasound of the kidney transplant was normal. Tacrolimus dose was reduced by one third, and amlodipine was rapidly increased from 10 mg/day to 30 mg/day in order to control hypertension (160/100 mmHg). Renal function carried on decreasing (creatininaemia: 478 µmol/L), and the patient became oliguric, necessitating haemodialysis, while tacrolimus trough level finally reached a toxic level of 38.8 ng/mL. No other explanation could be found for the nephrotoxicity except the high tacrolimus concentrations. The diarrhoea was resolved after 7 days. Tacrolimus was then discontinued, its trough level having returned to within a non-toxic range within 7 days. Amlodipine was interrupted after 10 days, and haemodialysis was discontinued after 20 days with complete recovery of previous renal function. The patient was found to have a CYP3A5 deficiency with a non-expressor CYP3A5*3/*3 genotype, apparently normal CYP3A4 activity as he had CYP3A4*1/1 genotype, and was heterozygous for the C3435T allele in multi-drug resistance 1 (MDR1) gene, whose influence on P-gp transport activity is still debated. The time course of the patient’s response to treatment is shown in Fig. 1.
https://static-content.springer.com/image/art%3A10.1007%2Fs00467-009-1402-8/MediaObjects/467_2009_1402_Fig1_HTML.gif
Fig. 1

Time course of tacrolimus trough blood level, renal function after diarrhoea and their relation to treatments. On admission, 7 days after the onset of viral diarrhoea, creatininaemia was 264 µmol/L and tacrolimus trough blood level was 27.2 ng/mL. Amlodipine was administered between day 7 and 17, starting at 10 mg/kg per day and rapidly increased to 30 mg/kg per day in order to control hypertension (160/100 mmHg) in the absence of fluid overload. Despite reducing tacrolimus by 30%, the tacrolimus trough level increased to 38.8 ng/mL on day 14, at which time it was stopped and dialysis started (creatininaemia was 478 µmol/L) for a duration of 20 days

Discussion

To our knowledge, this is the first report describing a renal transplant recipient who suffered from severe nephrotoxicity related to a raised trough blood level of tacrolimus in a situation of combined factors influencing tacrolimus pharmacokinetics: acute gastroenteritis, co-administration of a CCB, and demonstrated the CYP3A4, CYP3A5 and MDR1 SNPs, which had never been investigated previously.

We can speculate as to the pathogenesis of the increased trough tacrolimus level in our patient. We first believed that the increased tacrolimus trough at admission was only the result of the well-known combined effect of diarrhoea and tacrolimus administration. As non-dihydropiridine CCBs are mainly known to affect the metabolism of tacrolimus, we introduced amlodipine to control the patient’s hypertension, unfortunately underestimating the drug interaction in this particular context of enzymatic deficiency of the CYP3A system and decreased P-gp transport activity in this patient. To our knowledge, there are only a few reports of raised tacrolimus trough levels related to either CCB co-administration [6] or diarrhoea [7]. None of these reported patients presented with both exogenous factors at the same time, and more interestingly, none was investigated for CYP3A system and MDR1 SNPs. We may hypothesise that the tacrolimus trough level reached such a high and toxic value because of the conjunction of these three factors: diarrhoea, co-administration of CCB and the particular pharmacogenetic background.

Diarrhoea, intestinal inflammation and infection could have decreased the activity of the CYP3A system and P-gp by damaging enterocytes, and then tacrolimus bioavailability might have been modified [3]. Intestinal epithelial cells may be destroyed during enterocolitis, particularly in the case of rotavirus replication [8], reducing the global enzymatic activity of the CYP3A system and P-gp in enterocytes, thereby increasing the levels of tacrolimus [9]. The delay before normalisation of the tacrolimus blood level may thus reflect the time required for enterocyte regeneration and recovery of CYP3A enzymatic activity and P-gp function after gastrointestinal injury [10]. The shortened intestinal transit time during diarrhoea may also contribute to the elevated tacrolimus levels by increasing absorption [11]. The CYP3A system is highly expressed in the duodenum and then progressively declines to the colon [911]. If higher concentrations of tacrolimus reach the colon, in which CYP3A system activity is lower, this may theoretically contribute to elevated blood levels [8].

Calcium channel blockers are frequently administered in renal transplantation in order to control hypertension, partly attributed to the widespread use of calcineurin inhibitors [12]. CCBs are thus theoretically well suited to their dominant dilatory effect on the afferent glomerular arteriole, where the vasoconstriction of calcineurin inhibitors is most prominent [13]. However, CCBs have a potential interaction with calcineurin inhibitors, such as tacrolimus, through their common metabolism by the CYP3A system and P-gp [4], and decrease the clearance of tacrolimus by this partial competitive inhibition of the metabolic pathway, leading to a significantly elevated blood level and to the related toxicity. This is demonstrated for the interaction of diltiazem and cyclosporine with several reports of decreased clearance of cyclosporine in transplant recipients who received cyclosporine as part of their immunosuppressive regimen [14], even if pharmacodynamic interactions are difficult to assess in vivo. Interestingly, to our knowledge, there have been only three other reports of tacrolimus toxicity induced by drug interaction between tacrolimus and CCBs [6, 15, 16]. All observations involved non-dihydropiridine CCBs (diltiazem [6] and mibefradil [15, 16]). The three reported patients presented with clinically symptomatic adverse events, such as acute renal failure in two [15, 16], and/or neurotoxicity in two [6, 15].

Tacrolimus nephrotoxicity, more frequently observed in patients with CCBs than in those with diarrhoea, might be related to a direct mechanism on renal tubular epithelial cells [17]. CYP3A5 and P-gp are found at high levels in epithelial cells in the kidney [18, 19], and cyclosporine was found to be responsible for endoplasmic reticulum stress in in vitro cultures of human tubular cells, leading to epithelial phenotypic changes, which partly explains cyclosporine nephrotoxicity [17]. Therefore, if CYP3A5 and P-gp are partially inhibited by CCBs, we could hypothesise that non-metabolised tacrolimus might accumulate in renal epithelial cells with a direct toxic effect on cells, as may occur for cyclosporine.

The intestinal CYP3A system and P-gp play key roles in the bioavailability and pharmacokinetics of tacrolimus [2, 4], and the corresponding SNPs participate in the large inter-individual variability in tacrolimus pharmacokinetics [5]. CYP3A5 represents a considerable proportion of total CYP3A and significantly contributes to its catalytic activity [19]. Our patient presented with the most common CYP3A5 SNP (∼80% in Caucasians) [4]: a splice site mutation 16986G in intron 3, encoding for an aberrantly spliced mRNA, translating into an unstable protein, and thus leading to an absence of CYP3A5 enzymatic activity [19]. C3435T allele in the MDR1 gene is located on exon 26 and discrepancies abound on its effect [20, 21]. Tacrolimus might thus not be fully metabolised in the reported patient’s enterocytes and only partially transported back into the intestinal lumen because of all these SNPs, both mechanisms favouring an increased blood level and contributing to the presence of diarrhoea and drug interaction to reach a toxic tacrolimus trough blood level. Interestingly, CYP3A or MDR1 SNPs have never been genotyped in any observation of tacrolimus adverse events; they have only been fully investigated in cohort studies designed for the analysis of their influence on tacrolimus dose-adjusted trough levels and the daily tacrolimus doses in recipients [4]. It would be pertinent to conduct similar pharmacogenetic analyses in patients with toxic tacrolimus levels, to understand whether a particular CYP3A system and P-gp genetic background might explain the variability observed in the clinical adverse events of tacrolimus.

Considering the evidence on tacrolimus pharmacokinetics and pharmacogenetics, the potential impact of diarrhoea and the competitive inhibition by amlodipine, we can elaborate on our understanding of the clinical course of our patient as follows. The patient might have had a slightly decreased CYP3A system and P-gp transport activity. Indeed, he was a non-expressor of the CYP3A5*3/3 allele, as frequently seen in Caucasian people [4], he was apparently homozygous wild-type for CYP3A4, and heterozygous for the C3435T allele in the MDR1 gene (which leads to decreased P-gp transport activity). Diarrhoea might have carried on decreasing these activities by destroying intestinal epithelial cells, leading to a high tacrolimus trough level at the patient’s arrival. Because of uncontrolled hypertension, amlodipine was then added and progressively increased up to three-fold. This might have again decreased tacrolimus metabolism by competitive inhibition of the activity of CYP3A4 and P-gp enzymes, leading to a high and quick increase in tacrolimus trough level from 27.2 ng/mL to 38.8 ng/mL. This might be related to direct toxicity on tubular cells [17] as CYP3A5 and P-gp are found at high levels in epithelial cells in the kidney [18, 19]. Therefore, as our patient is a CYP3A5 non-expressor and heterozygous for the C3435T allele in the MDR1 gene, we could hypothesise that non-metabolised tacrolimus was accumulating in the renal epithelial cells with a direct toxic effect on cells, as might occur for cyclosporine. All these mechanisms are offered as theoretical explanations of the events reported in this patient.

Pharmacogenetic analysis would have been very interesting and would have highlighted that knowledge of the patient’s genetic background would not only help clinicians to choose the right day-to-day tacrolimus dose, but also guide them in adapting it in situations of high risk. Searches for CYP3A or MDR1 SNPs have been reported in cohort studies designed for the analysis of their influence on tacrolimus dose-adjusted trough levels and the daily tacrolimus doses chosen [4, 20, 2225] or for the analysis of their relation to signs of chronic calcinueurin inhibitors on biopsy [25]. However, CYP3A or MDR1 SNP genotyping has never been reported in any observations of clinical events related to an acute toxic tacrolimus trough blood level, and our case report, to the best of our knowledge, is unique in this respect.

In conclusion, whatever the pathogenic mechanisms involved and even though only partly understood, clinicians should remember the possibility of highly increased levels of tacrolimus in situations of diarrhoea and/or co-administration of CCBs, and should either avoid CCB co-administration or carefully monitor tacrolimus concentrations. The pharmacogenetics analysis performed in our patient may highlight the real clinical potential interest of searching for SNPs in order to not only closely adapt tacrolimus doses, but also to detect patients at high risk of complications in specific situations, such as diarrhoea and/or CCB co-administration. Further studies, such as comparison of SNP frequencies in patients and tacrolimus trough levels in high-risk situations, would be pertinent to assess the impact of SNPs in these particular situations and the possible benefits of using pharmacogenetic analysis in more routine practice.

Copyright information

© IPNA 2010