SEARCH

SEARCH BY CITATION

Keywords:

  • 3-hydroxykynurenine;
  • depression;
  • interferon-α;
  • kynurenine;
  • kynurenic acid;
  • neurotoxicity;
  • quinolinic acid

Abstract

  1. Top of page
  2. Abstract
  3. METHODS
  4. RESULTS
  5. DISCUSSION
  6. ACKNOWLEDGMENTS
  7. REFERENCES

Aims:  Immunotherapy with interferon-α (IFN-α) is associated with psychiatric side-effects, including depression. One of the putative pathways underlying these psychiatric side-effects involves tryptophan (TRP) metabolism. Cytokines including IFN-α induce the enzyme indoleamine 2,3-dioxygenase (IDO), which converts TRP to kynurenine (KYN), leading to a shortage of serotonin (5-HT). In addition, the production of neurotoxic metabolites of KYN such as 3-hydroxykynurenine and quinolinic acid (QA) might increase and contribute to IFN-α-induced psychopathology. In contrast, other catabolites of KYN, such as kynurenic acid (KA), are thought to have neuroprotective properties.

Methods:  In a group of 24 patients treated with standard IFN-α for metastatic renal cell carcinoma (RCC), combined psychiatric and laboratory assessments were performed at baseline, 4 and 8 weeks, and at 6 months.

Results:  No psychopathology was observed, despite an increase in neurotoxic challenge as reflected in indices for the balance between neurotoxic and neuroprotective metabolites of KYN.

Conclusions:  The present hypothesis that a shift in the balance between neurotoxic and neuroprotective metabolites of KYN underlies the neuropsychiatric side-effects of IFN-α-based immunotherapy, is neither supported nor rejected.

TREATMENT WITH THE cytokine interferon-α (IFN-α) is associated with the development of psychiatric side-effects, most notably depression. Unraveling the pathophysiological mechanisms underlying IFN-α-induced mood disorder could provide insight into the pathogenesis of depression in general.1

Many pathways and mechanisms have been proposed by which pro-inflammatory cytokines such as IFN-α could reach and affect the central nervous system (CNS) and give rise to mood disturbance.2 One of these putative mechanisms is the influence of cytokines on tryptophan (TRP) catabolism. According to this hypothesis, pro-inflammatory cytokines induce the enzyme indoleamine 2,3-dioxygenase (IDO), which converts TRP to kynurenine (KYN). As a result, plasma levels of TRP decrease, with a consequent lowering of influx of TRP into the CNS. Because TRP is the precursor of serotonin (5-HT), 5-HT production in the neurons decreases, which may result in the emergence of depressive symptoms.3,4

IFN-α may also induce the enzyme kynurenine hydroxylase, and therefore treatment with this cytokine could lead to an increase of the production of neurotoxic metabolites of KYN such as 3-hydroxykynurenine (3-OH-KYN) and quinolinic acid (QA; Fig. 1).5–8 These metabolites exert their action by inducing neuronal apoptosis (3-OH-KYN) and via activation of N-methyl-d-aspartic acid (NMDA) receptors (QA).9 In contrast, kynurenic acid (KA), another metabolite of KYN, which blocks the NMDA receptor, is hypothesized to counteract the effects of neurotoxic metabolites such as QA.8 Moreover, 3-OH-KYN and 3-hydroxyanthranilic acid (3-HAA) are neurotoxic due to oxidative stress.10–13 According to the neurodegeneration hypothesis of major depression, an imbalance between neurotoxic (such as QA) and neuroprotective (such as KA) metabolites of KYN contributes to depression.14 Finally, KA also blocks the α7-nicotinic acetylcholine receptor, which is thought to play a role in the pathophysiology of several psychiatric disorders, especially schizophrenia.15

image

Figure 1. Flow diagram of the kynurenine pathway. Bold, kynurenine and its metabolites; italics, enzymes.

Download figure to PowerPoint

Peripherally produced KYN can be transported across the blood–brain barrier and 60% of the CNS KYN is thought to have been produced outside the brain.16 In addition, catabolism of TRP to KYN and production of neurotoxic metabolites probably also takes place in astrocytes, providing another source of neurotoxic KYN metabolites.8

To our knowledge, few clinical data exist on the importance of neurotoxic metabolites of KYN for psychiatric disturbance. One study found no increase in cerebrospinal fluid levels of QA in patients with depression.17 Another study in depressed inpatients found a decrease in the ratio of TRP to the large neutral amino acids (TRP/LNAA ratio) – an index for TRP availability to the CNS – compared to controls, an increase in TRP catabolism and a decrease in the ratio of KA (the putative neuroprotective metabolite) to KYN (KA/KYN ratio), an index for the balance between neurotoxic influences and neuroprotection.8 In hepatitis patients, treatment with IFN-α resulted in an increase in the KYN/KA ratio, which reflects the neurotoxic challenge, and this increase was associated with higher depression scores.18 In oncology patients treated with IFN-α a more marked increase in KYN was seen in patients who developed depression at one time point, but not at the other time points during treatment.19 In that study concentrations of KA were not reported.

Previously, we investigated a group of 43 patients with high-risk malignant melanoma and renal cell carcinoma (RCC) in their first half year of treatment with pegylated and standard IFN-α, respectively.20,21 Little clinically relevant psychopathology was seen, and in the group as a whole few changes were seen on psychiatric rating scales, mostly in so-called somatic items such as appetite. This was observed, despite a clear and lasting decrease of concentration of TRP, an increase in concentration of KYN, an increase in the ratio of KYN to TRP (KYN/TRP ratio, an index of TRP breakdown) and an increase in the KYN/LNAA ratio, an index of influx in the brain of peripherally produced KYN. Furthermore, no consistent associations were observed during treatment with IFN-α between changes in any of the psychometric or biochemical variables compared to baseline.

The aim of the present study was therefore to investigate more precisely the concentrations and the aforementioned balance between neurotoxic and neuroprotective metabolites of KYN, in a subgroup of patients with metastatic RCC from the aforementioned study.20,21 This was done by measuring plasma levels of the neuroprotective substance KA and of the immediate precursor of QA, 3-HAA, at several time points in the first 6 months of treatment with standard IFN-α and by correlating these measures with psychiatric parameters.

METHODS

  1. Top of page
  2. Abstract
  3. METHODS
  4. RESULTS
  5. DISCUSSION
  6. ACKNOWLEDGMENTS
  7. REFERENCES

Subjects

The study group consisted of 24 patients (16 men, eight women, median age 60.5 years, range 47– 72 years) who consecutively enrolled in the second half of the aforementioned larger study (from participant number 20 onwards). Patients were treated for metastatic RCC with conventional short-acting IFN-α, initially at a dose of 3 MU, three times a week s.c. and escalating within 4 weeks to 9 MU s.c., three times a week. Excluded were patients concurrently using antidepressants, antipsychotics, mood stabilizers or corticosteroids, patients with major depression according to DSM-IV criteria or who were suffering from severe neuropsychiatric disorders, as well as patients with known brain metastases. In nine patients significant comorbid disease was present, mostly cardiovascular disorder and diabetes mellitus. Psychiatric assessment and blood sampling took place before the start of treatment, at 4 and 8 weeks and at 6 months. No assessments took place after IFN-α was stopped or when patients started using prohibited medication such as antidepressants and anti-epileptic drugs. The presence of a major depressive episode was ascertained with the relevant module of the Mini-International Neuropsychiatric Interview (MINI).22 In addition, for the present analysis the sum score and the scores on the subscales for anxiety, depression, hostility of the Symptom Check List-90 (SCL-90) were used. The SCL-90 is a well-validated, multidimensional self-report symptom inventory, designed to assess various dimensions of psychopathology.23,24

Procedure

Blood was collected in vacutainer tubes containing ethylenediamine tetra-acetic acid as anti-coagulant and obtained by venipuncture at the same time as the psychiatric assessments were performed. For practical reasons it was not possible to obtain blood samples at fixed times or under fasting conditions. After immediate centrifugation (10 min at 1700 g) plasma was separated and frozen at −80°C. Concentrations of KA, KYN and 3-HAA were measured as described elsewhere.8 Amino acids were determined as previously described.25 Because 3-HAA is the immediate catabolite of 3-OH-KYN and the direct precursor of QA, concentrations of 3-HAA were considered as an index for neurotoxic burden of metabolites of KYN. The ratio of KA to 3-HAA was considered as an index for the balance between neuroprotection and neurotoxicity. In addition, the KA/KYN ratio (plasma concentration of KA in nmol/L × 1000 divided by the plasma concentration of KYN in mmol/L) was computed because it was used as an index for the aforementioned balance in other studies.8 The KYN/LNAA ratio, which may be considered as an indicator for the supply of peripherally produced KYN to the CNS, was calculated by dividing the plasma concentration of KYN (multiplied by 100) by the sum of the LNAA, that is, tyrosine, valine, phenylalanine, leucine, isoleucine and TRP.

Statistical analysis

Data were stored and analyzed using SPSS version 10.0 (SPSS, Chicago, IL, USA). Outcomes at 4 and 8 weeks and at 6 months were compared to those at baseline using the Wilcoxon matched-pairs signed-ranks test. The Kolmogorov–Smirnov test was used for comparisons between groups of laboratory values. The Spearman rank correlation test was used to evaluate changes in laboratory parameters during follow up compared to baseline. The Spearman rank correlation coefficient (ρ) and the corresponding P were calculated for all pairs of changes. All reported P are two-sided, and P < 0.05 was considered statistically significant.

RESULTS

  1. Top of page
  2. Abstract
  3. METHODS
  4. RESULTS
  5. DISCUSSION
  6. ACKNOWLEDGMENTS
  7. REFERENCES

After a gradual build up, median weekly dose of IFN-α used was 27 MU at weeks 4 and 8, and at 6 months after the start of treatment. In the course of time, the number of patients in the study decreased: from the 24 patients, psychiatric and laboratory data were obtained for 22 patients at 4 weeks, for 20 at 8 weeks and for 10 patients at 6 months, due to cessation of treatment because of progression of disease (eight patients), cessation of treatment because of severe side-effects (n = 4), start of prohibited medication (n = 1) and administrative failure (n = 1). As shown in Table 1, the plasma concentrations of KYN increased at 4 and 8 weeks, the KYN/LNAA ratio increased at all three time points, concentrations of KA were decreased at 4 and 8 weeks, concentrations of 3-HAA were increased at 8 weeks, and both the KA/3-HAA ratio and the KA/KYN ratio were decreased at all points in time.

Table 1.  Baseline and follow-up concentration
 No. samplesBaseline4 weeks8 weeks6 months
24222010
  • Differences compared to baseline using the Wilcoxon matched-pairs signed-rank test.

  • 3-HAA, 3-hydroxyanthranilic acid; KA, kynurenic acid; KYN, kynurenine; LNAA, large neutral amino acids.

KYN (10−6 mol/L)Median4.466.105.744.94
Mean ± SD4.60 ± 1.215.71 ± 1.545.79 ± 1.125.25 ± 1.38
Range3.03–7.632.55–8.384.08–8.983.19–7.81
P <0.01<0.010.07
KYN/LNAA ratioMedian0.690.950.991.00
Mean ± SD0.77 ± 0.201.00 ± 0.270.99 ± 0.301.01 ± 0.31
Range0.50–1.290.48–1.550.53–1.660.45–1.51
P <0.01<0.010.04
KA (10−9 mol/L)Median55.853.249.853.0
Mean ± SD64.9 ± 26.558.9 ± 20.656.5 ± 19.656.5 ± 21.6
Range32.2–135.632.3–105.730.9–109.833.3–104.2
P 0.01<0.010.14
3-HAA (10−6 mol/L)Median18.519.021.422.0
Mean ± SD18.3 ± 3.520.0 ± 6.121.010 ± 5.522.4 ± 6.6
Range11.6–24.29.6–31.211.7–31.812.9–32.6
P 0.280.050.07
KA/3-HAA ratioMedian3.242.962.572.72
Mean ± SD3.54 ± 1.193.29 ± 1.672.82 ± 1.012.69 ± 1.08
Range1.93–6.031.36–7.180.97–4.981.17–4.53
P 0.04<0.010.01
KA/KYN ratioMedian12.89.28.410.4
Mean ± SD14.2 ± 5.010.9 ± 4.39.8 ± 2.910.9 ± 3.3
Range8.4–29.35.5–19.26.6–14.96.1–16.2
P <0.01<0.010.01

The scores on the SCL-90 before the start of therapy were average or below average compared to Dutch reference groups (median of the total score, 107.5 ± 15.5; range, 90–139; anxiety subscale median, 12.0 ± 1.97; range, 10–16; depression subscale median, 21 ± 4.4; range, 16–33; hostility subscale median, 6.0 ± 0.9; range, 6–9) and did not change statistically significantly during follow up with the exception of a decrease in anxiety at 4 weeks (P = 0.023). No depressive episodes were diagnosed in this subgroup.

No relationships were observed between changes compared to baseline for the psychiatric measures, and changes compared to baseline in laboratory parameters (KA, 3-HAA, KA/3-HAA ratio, KA/KYN ratio) at 4 weeks, 8 weeks and 6 months, with the exception of a correlation at 4 weeks between change in hostility and change in KA/3-HAA ratio (Spearman ρ = 0.46; P = 0.033), a correlation at 4 weeks between change in hostility and change in KA/KYN ratio (Spearman ρ = 0.58; P = 0.005) and a correlation at 8 weeks between change in anxiety and change in concentration of KA (Spearman ρ = −0.45; P = 0.045). In addition, we compared the changes in laboratory values compared to baseline on a given time point, between patients in whom an increase on the subscales or the sum score of the SCL-90 was seen, to patients in whom no changes were seen or who had a decrease. Because too few patients had an actual increase in anxiety or hostility and the number of patients was too small for testing at 6 months, this was possible only for increase versus no change or decrease on SCL-90 sum score and SCL-90 depression at 4 and 8 weeks. No differences in changes in laboratory values were seen between the groups with an increase of psychic complaints versus those with a decrease or no change (data not shown).

DISCUSSION

  1. Top of page
  2. Abstract
  3. METHODS
  4. RESULTS
  5. DISCUSSION
  6. ACKNOWLEDGMENTS
  7. REFERENCES

During treatment with IFN-α in oncology patients, concentration of KYN increases, and in view of the rise of the KYN/LNAA ratio, the supply of peripherally produced KYN to the brain is likely to be increased as well. Furthermore, the balance between neurotoxic and neuroprotective metabolites of KYN appears to change because both the KA/3-HAA ratio and the KA/KYN ratio were decreased at all points in time. This shift in balance, however, was not accompanied by clinically manifest depression in this subgroup and no consistent correlations could be documented between the laboratory parameters representing the neurotoxic balance and the psychiatric parameters. Partly, this observation is in contrast with the findings in patients with hepatitis C treated with IFN-α for whom a relationship was documented between neurotoxic challenge (defined as the ratio from KYN to KA) and depressive symptomatology.18

In the present study the concentrations of QA and 3-OH-KYN were not measured, but the concentrations of 3-HAA (the immediate catabolite of 3-OH-KYN and the immediate precursor of QA) did not increase, with the exception of one time point. This suggests that the production of QA and 3-OH-KYN is not strongly increased and the resulting neurotoxic burden on brain cells is limited.

It should be kept in mind that the present study had several limitations. These included small sample size, considerable attrition and the sampling of blood at non-fixed times. A further important limitation was the reliance (for obvious reasons) on peripheral measures. Although it has been stated that 60% of the brain KYN is derived from outside the brain and therefore peripheral concentrations of KYN are likely to reliably represent the brain concentrations of KYN, one cannot be sure if the same applies for 3-HAA and KA. Finally, the present findings must be interpreted with some caution because most studies in patients treated with IFN-α have reported considerably higher rates of psychopathology than we observed. Other factors, such as the gradual increase of the dose of IFN-α in the present study, could have counterbalanced effects resulting from changes in KYN metabolism and could have provided the present patients with a resilience to the psychotropic effects of IFN-α.

In the present study changes in the laboratory parameters measured did not appear to translate into mental changes. Both the absence of psychopathology during treatment with IFN-α and the other aforementioned study limitations preclude the drawing of a firm conclusion regarding the hypothesis that a shift in the balance between neurotoxic and neuroprotective metabolites of KYN, defined as a decrease in the KA/3-HAA ratio or the KA/KYN ratio, underlies IFN-α-induced mood disturbance.

ACKNOWLEDGMENTS

  1. Top of page
  2. Abstract
  3. METHODS
  4. RESULTS
  5. DISCUSSION
  6. ACKNOWLEDGMENTS
  7. REFERENCES

The authors thank Ms E. Bogaerts-Taal, Ms S.A. van der Heide-Mulder, Ms A.C.C. Voskuilen-Kooijman, Ms M. Dros, Ms A. van der Eng-Schipper, Ms H. van der Eng, Ms M. Mojka, Ms C.H.C. van Noort, Ms T.J.P. Pronk and Mr H. van der Meulen for their skilled technical assistance. This research was financially supported by a grant of the Foundation Nuts OHRA.

REFERENCES

  1. Top of page
  2. Abstract
  3. METHODS
  4. RESULTS
  5. DISCUSSION
  6. ACKNOWLEDGMENTS
  7. REFERENCES