Old drug, new tricks: haloperidol inhibits secretion of proinflammatory cytokines
- aRheumatology Research Group, University Hospital Aintree, Longmoor Lane, Liverpool L9 7AL, bDepartment of Rheumatology, Birmingham University, Birmingham, cDepartment of Immunology, University of Liverpool, Liverpool
- Dr R J Moots.
- Accepted 24 March 1999
OBJECTIVES It was noted that treatment of a patient with acute mania by haloperidol was associated with marked improvement in activity of rheumatoid arthritis. The objective of this study was to examine the effects of haloperidol on inflammatory cytokine release in vitro, as a potential mechanism to explain the in vivo anti-inflammatory effects of haloperidol.
METHODS The effect of haloperidol on the production of inflammatory cytokines interleukin 1β (IL1β) and tumour necrosis factor α (TNFα) was measured in bacterial lipopolysaccharide stimulated whole blood cultures and on the promonocyte cell line THP-1, using commercial and in house enzyme linked immunosorbent assays to measure cytokine concentrations.
RESULTS Haloperidol inhibited lipopolysaccharide stimulated production of both IL1β and TNFα in vitro in a dose dependent manner and over a prolonged time period. Marked inhibition was seen over a range of concentrations of haloperidol from 0.5 μg/ml to 50 μg/ml, including those predicted to occur in the patient’s blood.
CONCLUSIONS Haloperidol treatment seemed to alleviate inflammation in rheumatoid arthritis. In vitro experiments would suggest that the mechanism is by direct inhibition of proinflammatory cytokine release. This phenomenon requires further investigation and may potentially lead to the development of novel treatment.
Proinflammatory cytokines such as tumour necrosis factor α (TNFα) and interleukin 1β (IL1β) are intimately involved in the pathological process of rheumatoid arthritis (RA)1 and may be useful targets for treatment.2 We observed a dramatic improvement of RA in a patient taking the major tranquiliser haloperidol. On in vitro examination, this drug was able to reproducibly suppress both TNFα and IL1β production from a variety of cells using commercial or in house enzyme linked immunosorbent assays (ELISAs). This highlights the intimate relation between neuroendocrine and immune systems and may provide the basis for novel drug design in RA.
A 32 year old woman with stable seropositive RA controlled by non-steroidal anti-inflammatory drugs (NSAIDs) was admitted with acute mania. She was coincidentally noted to have extensive active synovitis involving the small joints of the hands and feet, together with knees and wrists, and early morning stiffness lasting until the evening. Markers of inflammation in the blood were raised with an erythrocyte sedimentation rate (ESR) of 65 mm 1st h and C reactive protein (CRP) 90 mg/dl. These clinical and laboratory features remained after two days of observation, and continued administration of NSAID. Oral haloperidol at a dose of 5 mg twice a day was then started to control the mania. Over the next two days not only did her mood normalise, but also the synovitis markedly resolved in her hands, wrists and knees. The CRP dropped in parallel to 20 mg/ml. During this time, the patient noticed that her early morning stiffness was limited to only approximately 90 minutes.
Haloperidol was gradually stopped over the next week. The patient remained fully ambulant and active throughout, and there was no other change in treatment. Two days after stopping haloperidol she suffered a marked flare of synovitis affecting the same joints as before, and complained of fatigue, with a parallel rise in CRP to 80 mg/ml and early morning stiffness of approximately six hours. Her mood remained normal and she was no more active around the ward than before. The flare was managed by intra-articular corticosteroid injections. She was discharged home prescibed sulfasalazine and has remained well.
The potential effect of haloperidol (stock solution 5 mg/ml for injection: Janssen Pharmaceuticals Ltd) on proinflammatory cytokines IL1β and TNFα production by whole blood on stimulation with bacterial lipopolysaccharide (LPS) was investigated using commercial ELISAs.
The method was adapted from that described by Van Wauweet al.3 Briefly, peripheral blood was collected from a healthy male donor into heparin tubes (12.5 u heparin/ml). It was diluted in RPMI 1640 medium and aliquoted into 250 μ1 fractions in 24 well Nunclon multidish plates containing bacterial LPS (E coli0111: B4, Difco) at a final concentration of 80 μg/ml, together with haloperidol at final concentrations of 0.5 μg/ml, 5 μg/ml, or 50 μg/ml. At various time points (6 hours, 24 hours, 48 hours, and 72 hours) cell free supernatants were collected by two times centrifugation of whole blood cultures at 2000g for five minutes. These were stored at −20°C for up to three days before assay for TNFα and IL1β.
The Genzyme human TNFα duo set and R+D systems IL1β matched pair ELISA kits were used to measure relevant cytokines. Both kits involve antigen capture with monoclonal antibody, addition of biotinylated second layer antibody and visualisation with horseradish peroxidase/peroxidase substrate in an optical plate reader. Each ELISA measurement was performed in triplicate wells. An in house ELISA4 was also used in parallel to estimate IL1β production from an unstimulated promonocyte cell line THP-1. In this case, an additional drug with effects on the central nervous system, diazepam, was studied in parallel as control. Measurements were again performed in triplicate wells.
Haloperidol suppressed production of both IL1β and TNFα from bacterial LPS stimulated whole blood over a wide range of concentrations, including those predicted pharmacologically (fig 1). These observation were consistent over prolonged time periods (fig 1) and with other concentrations of haloperidol over the same time periods (data not shown). In addition, haloperidol and not the control diazepam similarly suppressed unstimulated IL1β production from THP-1 cells using the in house IL1β ELISA (fig 2). The potential effects of haloperidol on cell viability, assessed by trypan blue exclusion, and general morphology by light microscopy, were also studied. Both of these features were unaffected by addition of haloperidol over the same range of concentrations. A 3H-leucine uptake assay, to measure total protein synthesis of proliferating cells, was performed in the presence of haloperidol at a concentration of 0.5 mg/ml (data not shown). Again, there was no effect on total protein synthesis.
Haloperidol has been reported to chronically suppress inflammatory arthritis. In a small trial using technetium scans as a marker of joint inflammation, it was observed that there was a reduction of isotope uptake of technetium in patients taking similar doses to that described in this report. The mechanism was unknown.5 Similar, less dramatic observations have been associated with chlorpromazine—of different chemical structure and class to haloperidol. Chlorpromazine is known to inhibit secretion of some cytokines.6-8 Our data indicate that haloperidol inhibits release of both proinflammatory IL1β and TNFα, potentially explaining the striking clinical improvement observed in our patient. The short time interval (up to six hours) before effect on TNFα suggests that haloperidol may have an inhibitory effect on release. The later effects (about 48 hours) on IL1β production implies that inhibition of this cytokine may be modulated at the gene level. In a recent study of psychotic patients taking major anti-psychotic drugs it was observed that withdrawal of haloperidol was associated with an increase in cerebrospinal fluid IL1β concentration, but without a clinical change.9 New generations of agents that directly interfere with cytokine secretion, perhaps binding to the same receptors on lymphocytes as haloperidol, may play a key part in the management of inflammatory joint disease.