sign in sign up
<<
Home , Antifungal

Molecular Psychiatry (2007) 12, 424–431 & 2007 Nature Publishing Group All rights reserved 1359-4184/07 $30.00 www.nature.com/mp FEATURE REVIEW Rationale for a trial of immunosuppressive therapy in acute schizophrenia JG Knight1, DB Menkes2, J Highton1 and DD Adams1 1Otago Medical School, University of Otago, Dunedin, New Zealand and 2Waikato Clinical School, University of Auckland, Hamilton, New Zealand

Schizophrenia is a debilitating, costly, socially disruptive, life-threatening disease in which available treatments are largely palliative and empirical, and produce significant short- and long-term side effects. Therefore, a strong case can made for exploring alternative treatments with a rational basis for use in this disease. Considerable evidence indicates that autoimmune processes may be involved in some forms of schizophrenia, including altered risk of certain autoimmune diseases in patients and their relatives, shared epidemiological features, and apparent involvement of genes known to influence the immune response repertoire. Attempts to provide direct evidence for autoimmune processes have proven elusive, possibly due to the technical difficulty inherent in accessing autoantibodies with high affinity for brain cell-surface receptors. In view of this impasse, we argue for a well-designed trial in schizophrenia of immunosuppressive therapy, which is now the mainstay of therapy for many autoimmune diseases. Analysis of disease states in which immunosuppression has been effectively used over many decades provides guidelines necessary for a meaningful trial. Molecular Psychiatry (2007) 12, 424–431. doi:10.1038/sj.mp.4001959; published online 23 January 2007 Keywords: schizophrenia; autoimmune; immunosuppressive therapy

Autoimmune processes in schizophrenia? and between schizophrenia and IDDM.5 Three possi- ble explanations for these dissociations are: Evidence pointing to autoimmune processes in schizophrenia includes: (a) used in treating one disease may protect from, or inhibit, the other; Altered risk of certain autoimmune diseases in (b) metabolic alteration associated with one disease patients with schizophrenia, or their close relatives. A may inhibit the other; recent study based on the Danish Psychiatric Register (c) genes predisposing to one disease may protect found that ‘a history of any autoimmune disease was against the other. For example, the histocompat- associated with a 45% increase in risk for schizo- ibility gene HLA-DR2 confers a 130-fold increased phrenia.’1,2 ‘Nine autoimmune disorders had higher risk of narcolepsy (possibly autoimmune), a 2.7- prevalence rates among patients with schizophrenia fold increased risk of multiple sclerosis (probably than among comparison subjects y and 12 auto- autoimmune), and a 13-fold increased risk of immune diseases had higher prevalence rates among Goodpasture’s disease (autoimmune). The same parents of schizophrenia patients than among parents gene confers a 10-fold decreased risk of IDDM, of comparison subjects.’ Wright et al.3 found schizo- and a 25-fold decreased risk in Japanese of phrenic patients with an affected first-degree relative pemphigus vulgaris,6 both autoimmune diseases. were significantly more likely to have a parent or sibling with an autoimmune disease. A significant Schizophrenia shares several features with known excess of insulin-dependent diabetes mellitus (IDDM) or suspected autoimmune diseases7,8 such as Graves’ was found in the parents and siblings of patients with disease, IDDM, pernicious anaemia, rheumatoid 3 schizophrenia. These findings are consistent with arthritis and multiple sclerosis, including: shared genetic predisposition for these various dis- eases, and suggest a commonality of mechanism. (a) tendency to run in families; Earlier reports indicated a strong dissociation (b) approximately 50% discordance in monozygotic between schizophrenia and rheumatoid arthritis,4 twins; (c) predisposition caused by a combination of genes rather than by a single gene; Correspondence: Dr JG Knight, Otago Medical School, University (d) ‘juvenile grace gap’ and random age of onset; of Otago, PO Box 56, Dunedin 9001, New Zealand. (e) remitting and relapsing course; E-mail: [email protected] Received 2 October 2006; revised 15 November 2006; accepted 4 (f) involvement of infectious triggers and other December 2006; published online 23 January 2007 environmental precipitants including (in schizo- Immunosuppressive therapy in acute schizophrenia JG Knight et al 425 phrenia) prenatal ,9,10 influenza infec- 25 genome scans for schizophrenia have now been tion,11 head injury12 and abuse.13 published, with no evidence for a gene of major effect.’ Carter31 points out that ‘over 130 genes have Reported immunological abnormalities in been associated with schizophrenia in genetic stu- schizophrenia include dies. None of these has reached a sufficient level of confidence to be accepted as a universal suscept- (a) oligoclonal antibodies in cerebrospinal fluid ibility gene and problems of replicability suggest that (CSF), suggesting antibody synthesis within the many may be false positives.’ To date, linkage regions central (CNS).14 This is reminis- identifying promising candidate genes include:30 cent of findings in multiple sclerosis of ‘a B-cell 22q12–q13 (or 22q11–1232), 8p22–p21, 6p24–p22, response to a specific antigen occurring in the 13q14–q32, 6q21–q22 and 1q21–22.32 The candidate central nervous system, whereas corresponding genes that researchers have focussed on are those clones are absent from the peripheral circula- having some obvious connection with brain path- tion’;15 ways and mechanisms. However, of particular inter- (b) elevated soluble interleukin-2 (IL-2) receptors in est to the thesis advanced here is that several of these serum, comparable to elevations in rheumatoid linkage regions are closely associated with loci arthritis, systemic lupus erythematosus, and mul- related to the . For example, the tiple sclerosis – all known or suspected to be 6p24-p22 is adjacent to the MHC (6p21), the ‘most autoimmune. Increased levels of IL-2 in CSF,16 important region in the vertebrate genome with correlating with recurrence of symptoms in respect to infection and autoimmunity’;33 the 1q21- schizophrenia;17 22 region includes two immunoglobulin Fc receptor (c) increased interleukin-6 (IL-6)18 indicative of im- loci, both at 1q21; the 22q11–13 region includes the mune system activation; IGL (immunoglobulin lambda) genes (at 22q11.2), (d) increased numbers of CD5 þ B lymphocytes,19 including the lambda light chain variable region similar to autoimmune diseases including rheu- genes, which provide a key component of the immune matoid arthritis, progressive systemic sclerosis, response repertoire. These immune system-related Graves’ disease, Sjogren’s syndrome and IDDM.20 genes may have been overlooked as candidates by Many genetic studies in schizophrenia have impli- researchers focussing on genes obviously connected cated genes known to determine predisposition to with neural, rather than immune, processes. The autoimmune diseases especially genes associated tentative hints that have so far come from genetic with the major histocompatibility complex (MHC).21 studies in schizophrenia appear compatible with To some extent this could be due to the fact that these multiple genes influencing immune response reper- are highly polymorphic loci that researchers have toire involvement in this disease, as in many known found it convenient to study. There is some consensus autoimmune diseases.34 that HLA A9 (or its A24 subspecificity) may confer an increased risk of schizophrenia.21 Wright et al.21 The uniquely effective ‘atypical’ concluded that ‘there is some evidence that clozapine1 is more potent than would be predicted DQB1*0602 and/or DRB1*04 alleles (or alleles at from its pharmacological profile as a dopamine D2 linked loci) may protect against schizophrenia.’ They receptor antagonist. This has led to speculation that rightly point out that ‘taken together, however, HLA its mode of action involves serotonin or other receptor association investigations provide only weak evi- systems. A well recognized and problematic adverse dence for the existence of either resistance or suscep- effect of clozapine is granulocytopenia and, more tibility loci for schizophrenia close to the HLA region rarely, agranulocytosis, leading to the standard prac- at the 6p21.3 band.’21 tice of monitoring white cell counts during A seven fold relative risk of schizophrenia has been treatment. However, it is clear that clozapine has reported in subjects with the complement C4BQO effects on hematopoietic cells other than granulo- allele;22 this finding is particularly interesting cytes, leading to the possibility that one of its effects, because C4A and C4B null alleles have been reported or even its main effect, may be to diminish antibody to be associated with several autoimmune diseases synthesis. Indeed McAllister et al.35 note that ‘several including systemic lupus erythematosus (SLE), patients treated with clozapine had a significant IDDM, subacute sclerosing panencephalitis and auto- reduction in B lymphocytes (i.e. antibody-secreting immune chronic active hepatitis.23 These C4 null cells) despite an insignificant change in the total associations may be due to linkage with other MHC white blood cell count.’ However, these authors were genes, the C4 deficiency itself, or both. Indications are seeking an increase in B-cell numbers in support of that, in diseases such as IDDM and multiple sclerosis, the hypothesis that the agranulocytosis was auto- susceptibility is determined by a combination of: antibody-mediated so the observation received no HLA Class II (D locus), MHC Class III (complement) further comment. Others have observed that cloza- genes, antibody structural genes and T lymphocyte pine increases soluble IL-2 receptor levels, and receptor genes.24–29 These are all genes expected concluded that this may be one mechanism by which to influence the immune response repertoire. Accord- clozapine causes immunosuppression.36 The extent ing to Riley and Kendler’s recent review30 ‘a total of to which antipsychotic drugs other than clozapine

Molecular Psychiatry Immunosuppressive therapy in acute schizophrenia JG Knight et al 426 share its immune effects is uncertain, although whether adrenal insufficiency contributed to schizo- Leykin et al.37 found that haloperidol also can be phrenia.41 Twenty-one patients with chronic schizo- immunosuppressive. phrenia who had been ill for more than 6 years were given intensive cortisone therapy, and six of the 21 Anti-brain antibodies. Although it is tempting to were reportedly cured and discharged from hospital; cite reports of anti-brain autoantibodies as supporting paranoid patients were the most responsive to treat- evidence for an autoimmune basis for schizophrenia, ment.41 Unfortunately, these spectacular results could these have proven controversial and lacking in not be confirmed in subsequent studies;42,43 the reproducibility.38 Antibodies with high affinity for reason for this discrepancy is not clear. It could be specific brain antigens, such as cell-surface recep- significant that the study which reported a beneficial tors,8 would likely become bound soon after synthesis effect used a relatively low dose (typically 100 mg and therefore unlikely to occur in measurable con- cortisone daily) for one to two weeks, followed by an centrations in CSF – let alone in blood. This high- even lower dose (50 mg daily) for a sustained period lights the difficulty of providing convincing direct of 6–8 weeks or more.41 In contrast, the subsequent evidence for or against an autoimmune hypothesis. studies42,43 used 400–500 mg/day of cortisone, which would be regarded as massive today. As a result of the severity of the side effects of these huge doses, therapy treatment lasted only 3 weeks in one study, and 3–10 In view of the above accumulated data indicating weeks in another. In this latter study,43 three patients possible autoimmune mechanisms in schizophrenia, showed slight and transient psychiatric improvement an adequate trial of immunosuppressive therapy is whereas five were made worse. Interestingly, one called for – namely synthetic , perhaps female patient showed definite improvement, but coupled with azathioprine or 6-mercaptopurine (both when the dose was dropped to below 100 mg per day thiopurines). The reason for past reluctance to use showed exacerbation of symptoms. When the dose corticosteroids may stem from a belief expressed in was increased to 200 mg per day the psychotic the psychiatric literature and in text- manifestations abated, but reappeared when the books that a history of mental illness contraindicates dose was lowered again.43 As detailed below, corti- their use. This does not appear well-founded, and is sone tends to be ineffective in treating exophthalmos based mainly on anecdotal case reports rather than on associated with Graves’ disease, in contrast to the controlled studies. In contrast, the Boston Collabora- synthetic corticosteroids prednisone and predniso- tive Drug Surveillance Program39 analysed the hospi- lone, which have a higher therapeutic index and are tal records of 718 patients treated with prednisone strikingly effective in treating this and other auto- for various diseases. Twenty-one patients (3%) expe- immune diseases.44 rienced acute psychiatric symptoms; eight exhibited inappropriate while 13 became psychotic Other immune-related therapies tried in with manifestations including hallucinations, delu- schizophrenia sions, violent behaviour, mania or severe depression. The euphoric patients recovered following reduction Efforts to test the efficacy of plasma exchange in in dosage, whereas the psychotic patients required schizophrenia proved negative, and this was inter- additional brief antipsychotic therapy. This study preted as counting against an autoimmune hypo- found a statistically significant dose–response effect, thesis.45 However, it is doubtful whether such a with psychiatric symptoms recorded in: 1.3% (6/463) procedure would succeed in removing autoantibodies patients receiving 40 mg/day or less; 4.6% (8/175) of from the brain side of the blood–brain barrier, those receiving 41–80 mg/day; and in 18.4% (7/38) of particularly if autoantibody synthesis were localized those receiving 80 mg/day or more.39 in the brain. Recent proposals – together with Pre-existing psychiatric illness was not found preliminary supporting data – have advocated use to increase the likelihood of an adverse reaction to of adoptive with in vitro-activated corticosteroids beyond that in the general popula- immune cells in schizophrenia and other mental tion,39,40 despite the contrary view in some earlier disorders.46,47 The proposed mechanism of action is textbooks. ‘some kind of restoration of distorted cytokine levels important for neuronal function’.47 We feel that it is Trials of cortisone in schizophrenia premature to comment on this proposed therapeutic strategy until further work supports or refutes it. Non- In the 1950s natural , in particular cortisone, steroidal anti-inflammatory drugs have been investi- were found ineffective in ameliorating the symp- gated for their potential to attenuate or prevent toms of schizophrenia, despite an earlier optimistic cognitive impairment in schizophrenia. A prospective report.41 It may be unfortunate that the water became double-blind study of supplementary treatment with muddied by early attempts when the only compounds celecoxib, a selective cyclooxygenase-2 (COX-2) in- available were of low therapeutic index, thereby hibitor, reported a beneficial effect during the acute necessitating the use of large doses causing major phase compared with risperidone therapy alone.48 adverse effects. The first of these studies set out to test However, a subsequent international multi-centre

Molecular Psychiatry Immunosuppressive therapy in acute schizophrenia JG Knight et al 427 trial could not demonstrate any benefit of celecoxib Corticosteroids are not stored in the adrenal, but are treatment.49 Riedel et al.49 argue ‘duration of illness synthesized at the rate required for normal secretion. may have been the reason for an apparent non- Corticosteroids control protein synthesis by binding response to celecoxib treatment.’ The hypothesized to receptors in sensitive cells and regulating the therapeutic effect of celecoxib on the psychopathol- of mRNA. Corticosteroids profoundly ogy of schizophrenic patients was thought to be based alter the immune responses of lymphocytes and are of on inhibition of COX activity.48 However, COX-2 immense value in treating diseases that result from expression has been found to not be upregulated in undesirable immune reactions.40 Sub-populations of the hippocampus of schizophrenia cases, in contrast lymphocytes are differentially affected, and the to what is observed in Alzheimer’s disease.50 clinical effects of therapy are often greater The equivocal nature of findings using these diverse than can be explained simply in terms of lymphocyte alternative immune-related therapies in our view numbers. In part, their action concerns alterations in strengthens the case for using mainstream immuno- distribution between different lymphocyte pools. suppressive therapy which is tried and true in a whole Multiple other mechanisms are involved in the range of medical disorders, as discussed below. suppression of immune inflammation by gluco- corticoids, including inhibition of the production of Mode of action of immunosuppressive drugs pro-inflammatory cytokines such as IL-1 and TNF, and the production of critical enzymes such as COX-2 There are several major categories of immunosup- and iNOS.52 pressive agents, including X-irradiation (used mainly Side effects from corticosteroids are correspondingly for its cytocidal effects on certain types of neoplastic profound, necessitating use of the smallest dose cells), anti-lymphocyte globulin, cytotoxic agents, possible and attention to amelioration of predictable corticosteroids, calcineurin inhibitors such as cyclos- consequences. According to Goodman & Gilman’s porine and related agents, more targeted inhibitors of Pharmacological Basis of Therapeutics:40 ‘Because purine and pyrimidine such as mycophe- they exert effects on almost every organ system, the nolate mofetil and leflunomide and newer ‘biological’ clinical use of and withdrawal from corticosteroids are agents aimed at more specific cells, receptors and complicated by a number of serious side effects, some cytokines involved in immune interactions.40 The of which are life threatening. Therefore, the decision to semi-selective action of many of these agents is based institute therapy with corticosteroids always requires on the rapid cell division that distinguishes cells of careful consideration of the relative risks and benefits the immune system from other tissues, particularly in each patient.’ Among the potential side effects from when participating in a current immune response. It prolonged therapy are: cataracts, fluid and electrolyte follows that cells which are part of an active immune imbalances, redistribution, hypertension, hypergly- response at the time of treatment will be more affected caemia, increased susceptibility to infection, myopa- than cells which represent either immunological thy, osteoporosis, osteonecrosis, and psychosis. ‘memory’, or that part of the immune response Despite these dose-related side effects, the usefulness repertoire which is potential rather than realized. of corticosteroids in effectively treating a wide variety Commonly used cytotoxic drugs include cyclopho- of inflammatory and autoimmune diseases makes them sphamide (an alkylating agent), azathioprine (a purine among the most frequently prescribed classes of analogue ), and methotrexate (a folic acid drugs.40 Potential interactions with other drugs53–55 antagonist). These agents have been used mainly in are shown in Table 1. These require careful considera- diseases in which therapy was found to be tion, but none of the known interactions would partially effective. Combining cytotoxic agents with preclude combining corticosteroids and/or azathiopr- steroids results in greater efficacy, especially in more ine with anti-psychotic , with the possible severe disease such as lupus nephritis. Combined use of exception of clozapine. steroids and cytotoxic agents also incorporates a ‘steroid sparing’ strategy based on the concept that administering Immunosuppressive therapy of selected two drugs with distinct modes of action often enables use autoimmune diseases of lower doses, thus minimizing side effects associated with high doses of corticosteroids. Alkylating agents Corticosteroids and other immunosuppressive drugs such as have severe side effects have been widely used in treating autoimmune including infection, ovarian failure and carcinogenesis, conditions, including those mediated by autoantibo- which restricts use to life threatening disease. Thus, the dies, for example myasthenia gravis, Graves’ disease, current trend in treatment of lupus nephritis, for thrombocytopenia, warm type haemolytic anaemia, example, is to use a short induction course of cyclopho- pemphigus vulgaris. sphamide followed by maintenance with a safer treat- ment such as azathioprine or mycophenolate mofetil.40,51 Myasthenia gravis The adrenal cortex synthesizes two classes of There is some controversy about the best form of steroids from (the corticosteroids with 21 therapy to use in treating myasthenia gravis – a carbon atoms and the adrenal androgens with 19 disease caused by autoantibodies against the acetyl- carbons) by a series of enzymatic conversion steps. choline receptor on muscle cells.56 Thymectomy,

Molecular Psychiatry Immunosuppressive therapy in acute schizophrenia JG Knight et al 428 Table 1 Known drug interactions between corticosteroids, azathioprine, anti-psychotic drugs used in treating schizophrenia, and other drugs which patients could be taking for unrelated reasons53–55

A Drugs reported to possibly decrease therapeutic effects of corticosteroids:54,55 ; barbiturates; ; ephedrine; hydantoins (e.g. ); rifamycins (e.g. rifabutin, – also known as rifampin, rifapentine). B Drugs reported to possibly potentiate therapeutic and toxic effects of corticosteroids:54,55 Aprepitant; agents (e.g. , , , , ); diltiazem; ; grapefruit juice; (e.g. clarithromycin, , ); . C Drugs reported to have adverse interactions with corticosteroids:54 Potassium-depleting (severe depletion may occur if used together); beta blockers (an isolated report of hyperkalemia). (Also, epilepsy may be exacerbated by use of corticosteroids; great care needed in glaucoma, peptic ulceration, and acute .) D Reported interactions between and corticosteroids:53 (steroid) is listed as a medicine-type that has a low degree of inhibition of antipsychotics of the type that are metabolized by the CYP3A subsystem (i.e. , clozapine, haloperidol, mesoridazine, pimozide, quetiapine, sertindole, ziprasidone). E Drugs reported to potentiate therapeutic effects and toxicity of azathioprine:54 Allopurinol; balsalazide; mesalazine; olsalazine; sulfasalazine; Hematological toxicity (potentially life-threatening) by co-trimoxazole or trimethoprin; increased by doxorubicin. F Drugs contraindicated in patients treated with clozapine:54 Because clozapine can cause blood dyscrasias and potentially fatal agranulocytosis, it should not be given with other drugs that can cause agranulocytosis including: Cytotoxic drugs (e.g. azathioprine); carbamazepine; co-trimoxazole; ; penicillamine; phenylbutazone; pyrazolone ; sulphonamides.

corticosteroids, and other forms of immunosuppres- tissue in the orbit and the shins, respectively. The sion are all useful in treatment.57 core pathology of Graves’ disease is hyperthyroidism, In a study of 116 myasthenic patients treated caused by thyroid-stimulating autoantibodies. This with corticosteroids,58 a satisfactory outcome (marked can be effectively controlled by surgical sub-total improvement or remission) was reported in 80.2%. thyroidectomy, partial ablation of the thyroid with The outcome was unsatisfactory in the remaining radioiodine, or by use of anti-thyroid drugs, so there 19.8%, with 5.2% being unimproved and 14.7% has been no reason to investigate immune suppres- showing only moderate improvement. The dose used sion as a means of controlling this disease. However, was 60–80 mg daily until improvement was noted recently, a patient with extreme thyrotoxicosis was on each of 3 successive days, then an equivalent found to be resistant to the anti-thyroid drugs alternate-day regime was instituted. It is noteworthy carbimazole and propylthiouracil, causing him to be that 56 patients (48%) showed an initial exacerbation, treated effectively with prednisone, 40 mg daily.44 which in 8.6% was severe; occurrence of exacerbation Cortisone was used for treating exophthalmos in was not related to degree of ultimate improvement.58 the 1950s, there being controversy regarding its This could be relevant to the finding in the cortisone effectiveness.59 However, substitution of prednisone, trials in schizophrenia of exacerbation in some ‘resulted in effective and remarkable degrees of patients, resulting in early termination of the treat- control’, leading to retention of vision in patients ment.42,43 Of the 93 myasthenic patients who re- who would otherwise have lost it.59 Pretibial sponded satisfactorily, the duration of therapy to the myxoedema is cured by topical application of cortico- time of maximal improvement ranged from 2 weeks to steroids.40 6 years, with the median time to maximal improve- ment being 5.5 months. The period required to achieve a state of marked improvement ranged from 1.5 weeks to 18 months, mean 3.1 months.58 Of Hemolytic anaemia relevance to the present proposal is that the frequency Corticosteroids and azathioprine are regarded as the of ‘psychological disturbance’ was 3 of 129 patients or treatment of choice for warm-type autoimmune 2.3%, closely matching the figure reported in the haemolytic anaemia.40 Cyclophosphamide is also Boston Collaborative Study.39 used, but the multi-system toxicity and oncogenic potential of this and other alkylating agents indicate Exophthalmos, pretibial myxoedema and that they should only be used when corticosteroids hyperthyroidism of Graves’ disease and azathioprine fail. Approximately 50% of patients The complications of Graves’ disease are presumed to who do not respond to corticosteroid treatment will be caused by variants of thyroid-stimulating auto- respond to combined therapy of azathioprine plus antibodies that have a high affinity for receptors on low-dose corticosteroids.40

Molecular Psychiatry Immunosuppressive therapy in acute schizophrenia JG Knight et al 429 Goodpasture’s syndrome selective immunomodulating drug for the treatment Standard treatment consists of plasma exchange, to of multiple sclerosis.66 Treatment options are most remove the anti-basement membrane antibodies, effective during the relapsing-remitting phase, and coupled with immunosuppressive therapy.40 largely ineffective during the later progressive neuro- degenerative phase.15 Pemphigus vulgaris Patients with pemphigus vulgaris produce IgG anti- bodies directed against squamous epithelial cell Relevant considerations for an effective trial surfaces.40,57 This once-fatal disease is now regarded of immunosuppression in schizophrenia as controllable with corticosteroids, or combined therapy with corticosteroids plus cytotoxic drugs (1) Dexamethasone is more likely than prednisone or such as azathioprine, methotrexate or cyclophospha- to achieve potentially therapeutic mide.40 concentrations within the brain because of its lower tendency to bind to plasma proteins. (2) A combination of dexamethasone plus azathio- Use of immunosuppressive agents in treating brain prine is likely to be better tolerated than either diseases drug alone, because the dose of each can be kept Autoimmune diseases affecting the brain require lower. This combination of corticosteroid plus special consideration. The ‘sanctuary effect’ created cytotoxic drug is a thoroughly tried and tested by the blood–brain barrier prevents many drugs from form of therapy in autoimmune diseases. Azathio- achieving cytotoxic concentrations within the CNS.60 prine also has the advantage that measurement Corticosteroids have been used to treat brain tumours of thiopurine methyl transferase (TPMT) before and acute lymphoblastic leukaemia involving the treatment may limit the occurrence of neutro- meninges, and differences have been noted in the penia, and levels of 6-thioguanine nucleotide, efficacy of dexamethasone and prednisone for these the principal therapeutic metabolite, can be uses.60,61 After intravenous bolus injection into rhesus measured.67 monkeys, dexamethasone and prednisolone were (3) A realistic aim is not to ‘’, but to reduce found to reach equally potent peak concentrations psychotic and other symptoms, and thereby help in the CSF. However, prednisolone had a shorter half- to preserve or restore function. Clearly immuno- life in the CSF than did dexamethasone, attributable suppressive therapy cannot terminate an auto- to its greater tendency to bind to plasma proteins. It immune process without turning off the entire was estimated that CSF levels of prednisolone immune system. following standard oral doses would be 5- to 10-fold (4) In many autoimmune diseases it takes several lower than those achieved with an equipotent weeks, sometimes months, for the full beneficial systemic dose of dexamethasone.61 Patients with effect of combined treatment with corticosteroids acute lymphoblastic leukaemia had a significantly and azathioprine to be seen. A trial of immuno- lower rate of CNS relapse when treated with dexa- suppression in schizophrenia would need to methasone than with prednisone.62 continue several (3–6) months in order to be meaningful. The possibility of spontaneous remis- Systemic lupus erythematosus sion would need to be controlled for statistically. Systemic lupus erythematosus (SLE) is an autoim- (5) In view of the likelihood that ‘schizophrenia’ mune disease in which CNS involvement is observed encompasses multiple diseases of differing etiol- in approximately 14% of cases.63 The standard ogies, it is unlikely that all patients would approach to treatment of severe cases involves the respond to one type of treatment. Thus, schizo- use of high dose prednisone and cyclophosphamide phrenia sub-type would need to be carefully and other strategies using prednisone and azathio- recorded at trial entry. Furthermore, in autoim- prine administered systemically, similar to treatment mune diseases that respond to immunosuppres- of lupus renal disease.64 It is thought that abnormal- sion, there is considerable individual variation in ities of the blood–brain barrier exist in SLE, allowing responsiveness to particular therapies. Accord- access of systemic treatments to the CNS.65 However, ingly, to be meaningful and ethical, a trial would one innovative approach reported the use of combi- have to be appropriately powered to detect an nation therapy with methotrexate and dexamethasone effect in some patients but not others. intrathecally.63 (6) Patient selection is likely to be important, given the fact that the usual course of schizophrenia has Multiple sclerosis both progressive and relapsing-remitting features. Various immunosuppressive therapies have been Logically, first episode illness would be most tried in multiple sclerosis,40 including: cortico- likely to coincide with active autoimmune pro- steroids and cyclophosphamide, azathioprine, cyclo- cesses and thus may offer the best chance of sporin A, interferon-beta, and anti-lymphocyte effective immunosuppression affecting the dis- globulin.15 A recent addition is natalizumab, a ease. On the other hand, it may be ethically easier recombinant monoclonal antibody, which is the first to justify a novel treatment after conventional

Molecular Psychiatry Immunosuppressive therapy in acute schizophrenia JG Knight et al 430 treatment has failed, or proven only partially could provide valuable insights into aetiology, as well effective.1 as providing principles for effective therapy. (7) Thus, a trial of immunosuppression could be considered in patients, including those having their first episode, with inadequate response References to conventional therapy after a specified time interval (e.g. 1 month) according to one of the 1 Jones P, Barnes T, Davies L, Dunn G, Lloyd H, Hayhurst K et al. following designs: Randomized controlled trial of the effect on quality of life of (i) Continued on anti-psychotic drug (APD) treat- second- vs first-generation antipsychotic drugs in schizophrenia. Arch Gen Psychiatry 2006; 63: 1079–1087. ment and randomized to adjunctive treatment 2 Eaton W, Byrne M, Ewald H, Mors O, Chen C-Y, Agerbo E et al. (immunosuppression or placebo); Association of schizophrenia and autoimmune diseases: linkage of (ii) Withdrawn from initial APD, given alternative Danish national registers. Am J Psychiatry 2006; 163: 521–528. APD plus randomized adjunctive treatment 3 Wright P, Sham P, Gilvarry C, Jones P, Cannon M, Sharma T et al. Autoimmune diseases in the pedigrees of schizophrenic and (immunosuppression or placebo); control subjects. Schizophr Res 1996; 20: 261–267. (iii) Withdrawn from initial APD and randomized 4 Mellsop G, Koadlow L, Syme J, Whittingham S. Absence of to alternative APD or immunosuppression. rheumatoid arthritis in schizophrenia. Aust NZ J Med 1974; 4: 247–252. Option (iii) is the most difficult to justify on ethical 5 Finney G. Juvenile onset diabetes and schizophrenia? Lancet 1989; ii: 1214–1215. grounds, followed by (i), then (ii). In terms of 6 Tiwari J, Terasaki P. HLA and Disease Associations. Springer- statistical power to detect an effect of immunosup- Verlag: New York, 1985. pression, the order is (i) > (iii) > (ii). Accordingly, (i) 7 Knight J, Knight A, Pert C. Is schizophrenia a virally triggered may provide the best trade-off between ethics and antireceptor autoimmune disease?. In: Helmchen H and Henn F (eds). Biological Perspectives of Schizophrenia. Wiley: Chichester, power, whereas (ii) would be easier to justify in the 1987, pp 107–127. present ethical climate. Option (i) has the further 8 Knight J. Dopamine-stimulating autoantibodies: a possible cause advantage of allowing for dose optimization of initial of schizophrenia. Lancet 1982; ii: 1073–1076. APD and monitoring of delayed responses. 9 Mednick S, Machon R, Huttunen M, Bonett D. Adult schizo- phrenia following prenatal exposure to an influenza epidemic. Arch Gen Psychiatry 1988; 45: 189–192. (8) As a result of its documented effects on the 10 O’Callaghan E, Sham P, Takei N, Glover G, Murray R. Schizo- immune system, clozapine should probably be phrenia after prenatal exposure to 1957 A2 influenza epidemic. avoided. Apart from clozapine, APDs could be Lancet 1991; 1: 1248–1250. whatever the treating physician was accustomed 11 Menninger K. Influenza and schizophrenia. An analysis of post- influenzal praecox as of 1918 and five years later. Am J to prescribing, and the nature of the drug (first Psychiatry 1926; 82: 469–529. generation, second generation) used as a grouping 12 Davison K. Organic and toxic concomitants of schizophrenia. variable in the analysis. Other psychotropics (e.g. In: Helmchen H and Henn F (eds). Biological Perspectives of and antiparkinson agents) would need Schizophrenia. Wiley: Chichester, 1987, pp 139–160. to be carefully monitored and included as covari- 13 Turner W, Tsuang M. Impact of substance abuse on the course and outcome of schizophrenia. Schizophr Bull 1990; 16: 87–95. ates in the analysis. 14 Kirch D, Kaufmann C, Papadopoulos N, Martin B, Weinberger D. (9) Outcome measures would need to include vali- Abnormal cerebrospinal fluid protein indices in schizophrenia. dated, sensitive scales for the various symptom Biol Psychiatry 1985; 20: 1039–1046. clusters of schizophrenia (positive, negative, dis- 15 Frohman E, Racke M, Raine C. Multiple sclerosis – the plaque and its pathogenesis. N Engl J Med 2006; 354: 942–955. organized) as these may be differentially affected 16 Licinio J, Seibyl J, Altemus M, Charney D, Krystal J. Elevated CSF by treatment. They would also need to include levels of interleukin-2 in neuroleptic-free schizophrenic patients. measures of side effects and quality of life. A Am J Psychiatry 1993; 150: 1408–1410. possible set of outcome measures selected from 17 McAllister C, van Kammen D, Rehn T, Miller A, Gurklis J, Kelley those used by Jones et al.1 is: Positive and M et al. Increases in CSF levels of interleukin-2 in schizophrenia: effects of recurrence of psychosis and medication status. Am J Negative Syndrome Scale (PANSS); Simpson- Psychiatry 1995; 152: 1291–1297. Angus EPS rating scale; Barnes Akathisia Rating 18 Ganguli R, Yang Z, Shurin G, Chengappa R, Brar J, Gubbi A et al. Scale; Antipsychotic Non-Neurological Side-Ef- Serum interleukin-6 concentration in schizophrenia: elevation fects Rating Scale (especially important for second associated with duration of illness. Psychiatry Res 1994; 51: 1–10. 19 McAllister C, Rapaport M, Pickar D, Podruchny T, Christison G, generation anti-psychotics); Quality of Life Scale Alphs L et al. Increased numbers of CD5 þ B lymphocytes in (QLS); Participant Satisfaction Scale. schizophrenic patients. Arch Gen Psychiatry 1989; 46: 890–894. 20 Schatz D, Lang F, Cantor A, Riley W, Maclaren N, Sleasman J et al. CD5 þ B lymphocytes in high-risk islet cell antibody-positive and Conclusion newly diagnosed IDDM subjects. Diabetes 1991; 40: 1314–1318. 21 Wright P, Nimgaonkar V, Donaldson P, Murray R. Schizophrenia Several decades of research effort have failed to and HLA: a review. Schizophr Res 2001; 47: 1–12. provide clear indications of causative mechanisms 22 Rudduck C, Beckman L, Franzen G, Jacobsson L, Lindstrom L. in schizophrenia. We review the many sources of Complement factor C4 in schizophrenia. Hum Hered 1985; 35: evidence pointing to autoimmune processes. Widely 223–226. 23 Lokki M-L, Circolo A, Ahokas P, Rupert K, Yu C, Colten H. accepted immunosuppressive therapy developed and Deficiency of Human Complement Protein C4 Due to Identical refined for treating many autoimmune diseases Frameshift Mutations in the C4A and C4B Genes. J Immunol 1999; should be trialled for efficacy in schizophrenia. This 162: 3687–3693.

Molecular Psychiatry Immunosuppressive therapy in acute schizophrenia JG Knight et al 431 24 Thomsen M, Molvig J, Zerbib A, de Preval C, Abbal M, Dugoujon J hypothesis to combine drug treatment and adoptive immunother- et al. The susceptibility to insulin-dependent diabetes mellitus is apy. Med Hypotheses 2005; 65: 243–252. associated with C4 allotypes independently of the association 47 Wank R. Schizophrenia and other mental disorders require with HLA-DQ alleles in HLA-DR3,4 heterozygotes. Immunoge- long-term adoptive immunotherapy. Med Hypotheses 2002; 59: netics 1988; 28: 320–327. 154–158. 25 Field L, Dizier M, Anderson C, Spence M, Rotter J. HLA- 48 Muller N, Riedel M, Scheppach C, Brandstatter B, Sokullu S, dependent Gm effects in insulin-dependent diabetes: evidence Krampe K et al. Beneficial antipsychotic effects of celecoxib add- from pairs of affected siblings. Am J Hum Genet 1986; 39: 640–647. on therapy compared to risperidone alone in schizophrenia. Am J 26 Millward B, Welsh K, Leslie R, Pyke D, Demaine A. T cell receptor Psychiatry 2002; 159: 1029–1034. beta chain polymorphisms are associated with insulin-dependent 49 Riedel M, Strassnig M, Schwarz M, Muller N. COX-2 inhibitors as diabetes. Clin Exp Immunol 1987; 70: 152–157. adjunctive therapy in schizophrenia: rationale for use and 27 Salier J, Sesboue R, Martin-Mondiere C, Daveau M, Cesaro P, evidence to date. CNS Drugs 2005; 19: 805–819. Cavelier B et al. Combined influences of Gm and HLA phenotypes 50 Yokota O, Terada S, Ishihara T, Nakashima H, Kugo A, Ujike H et upon multiple sclerosis susceptibility and severity. J Clin Invest al. Neuronal expression of cyclooxygenase-2, a pro-inflammatory 1986; 78: 533–538. protein, in the hippocampus of patients with schizophrenia. Prog 28 Seboun E, Robinson M, Doolittle T, Ciulla T, Kindt T, Hauser S. A Neuropsychopharmacol Biol Psychiatry 2004; 28: 715–721. susceptibility locus for multiple sclerosis is linked to the T-cell 51 Balow J, Austin H. Maintenance therapy for lupus nephritis receptor beta chain complex. Cell 1989; 57: 1095–1100. – something old, something new. NEnglJMed2004; 350: 29 Zhang W, Cobain T, Dawkins R, Grimsley G, Mackay I. Comple- 1044–1046. ment allotyping explains MHC associations in multiple sclerosis. 52 Pelaia G, Vatrella A, Cuda G, Maselli R. Molecular mechanisms Ann NY Acad Sci 1988; 540: 372–373. of corticosteroid actions in chronic inflammatory airway disease. 30 Riley B, Kendler K. Molecular genetic studies of schizophrenia. Life Sci 2003; 72: 1549–1561. Eur J Hum Genet 2006; 14: 669–680. 53 Welner M. Antipsychotic drugs and interactions: implications for 31 Carter C. Schizophrenia susceptibility genes converge on inter- criminal and civil litigation. In: Mozayani A and Raymon L (eds). linked pathways related to glutamatergic transmission and long- Handbook of Drug Interactions. Humana Press: Totowa, NJ, 2004. term potentiation, oxidative stress and oligodendrocyte viability. 54 Baxter K, Stockley I. Stockley’s Drug Interactions: A Source Book Schizophr Res 2006; 86: 1–14. of Interactions, their Mechanisms, Clinical Importance and 32 O’Donovan M, Williams N, Owen M. Recent advances in the Management. Pharmaceutical Press: London, 2006. genetics of schizophrenia. Hum Mol Genet 2003; 12: R125–R133. 55 Tatro D. Facts. Wolters Kluwer Health Inc: 33 Horton R, Wilming L, Rand V, Lovering R, Bruford E, Khodiyar V St Louis, Missouri, 2006. et al. Gene map of the extended human MHC. Nat Rev Genet 2004; 56 Drachman D. Myasthenia gravis and other diseases of the 5: 889–899. neromuscular junction. In: Kasper D (ed). Harrison’s Principles 34 Knight J, Knight A, Ungvari G. Can autoimmune mechanisms of Internal Medicine. McGraw Hill: London, 2005, pp 2518–2523. account for the genetic predisposition to schizophrenia? Br J 57 Sweetman S. Martindale: The Complete Drug Reference. Pharma- Psychiatry 1992; 160: 537–540. ceutical Press: London, 2005. 35 McAllister C, Rapaport M, Pickar D, Paul S. Effects of short-term 58 Johns T. Long term corticosteroid treatment of myasthenia gravis. administration of antipsychotic drugs on lymphocyte subsets in Ann NY Acad Sci 1987; 505: 568–583. schizophrenic patients. Arch Gen Psychiatry 1989; 46: 956–957. 59 Werner S, Ingbar S. The Thyroid. Harper and Row: New York, 36 Maes M, Melzer H, Bosmans E. Immune-inflammatory markers in 1975. schizophrenia: comparison to normal controls and effects of 60 Berg S, Poplack D. Treatment of meningeal . Oncologist clozapine. Acta Psychiatr Scand 1994; 89: 346–351. 1996; 1: 56–61. 37 Leykin I, Mayer R, Shinitzky M. Short and long term immuno- 61 Balis F, Lester C, Chrousos G, Heideman R, Poplack D. Differences suppressive effects of clozapine and haloperidol. Immunophar- in cerebrospinal fluid penetration of corticosteroids: possible macology 1997; 37: 75–86. relationship to the prevention of meningeal leukemia. J Clin Oncol 38 Knight J, Knight A, Menkes D, Mullen P. Autoantibodies against 1987; 5: 202–207. brain septal region antigens specific to unmedicated schizophre- 62 Jones B, Freeman A, Shuster J, Jacquillat C, Pochedly C, Sinks L et nia? Biol Psychiatry 1990; 28: 467–474. al. Lower incidence of meningeal leukemia when prednisone is 39 Ling M, Perry P, Tsuang M. Side effects of corticosteroid therapy. replaced by dexamethasone in the treatment of acute lymphocytic Arch Gen Psychiatry 1981; 38: 471–477. leukemia. Med Pediatr Oncol 1991; 19: 269–275. 40 Brunton L, Lazo J, Parker K. Goodman & Gilman’s The Pharma- 63 Valesini G, Priori R, Francia A, Balestrieri G, Tincani A, Airo P et cological Basis of Therapeutics. McGraw-Hill: New York, 2006. al. Central nervous system involvement in systemic lupus 41 Cohn J, Steckler G, Rubinstein J. Effect of cortisone in treatment of erythematosus: a new therapeutic approach with intrathecal psychotic illnesses. J Clin Exp Psychopathol 1953; xiv: 161–175. dexamethasone and methotrexate. Springer Semin Immunopathol 42 Rees L, King G. Intensive cortisone therapy in schizophrenia. 1994; 16: 313–321. J Ment Sci 1956; 102: 155–159. 64 Sanna G, Bertolaccini M, Mathieu A. Central nervous system 43 Polatin P, Lesse S, Harris M. Use of large doses of cortisone in lupus: a clinical approach to therapy. Lupus 2003; 12: 935–942. schizophrenia. AMA Arch Neurol Psychiatry 1955; 73: 485–495. 65 Abbott N, Mendoca L, Dolman D. The blood-brain barrier in 44 Adams D, Knight J, Manning P, Smith G. An informative case of systemic lupus erythematosus. Lupus 2003; 12: 908–915. Graves’ disease with implications for schizophrenia. J Clin Lab 66 Ropper A. Selective treatment of multiple sclerosis. N Engl J Med Immunol 2005; 53: 13–25. 2006; 354: 965–967. 45 Schulz S, van Kammen D, Waters R, Klein H, Balow J, Bunney W. 67 Al Hadithy A, de Boer N, Derijks L, Escher J, Mulder C, Brouwers J. Double-blind evaluation of plasmapharesis in schizophrenic Thiopurines in inflammatory bowel disease: pharmacogenetics, patients: a pilot study. Artif Organs 1983; 7: 317–321. therapeutic drug monitoring and clinical recommendations. Dig 46 Fellerhoff B, Laumbacher B, Wank R. High risk of schizophrenia Dis 2005; 37: 228–297. and other mental disorders associated with chlamydial infections:

Molecular Psychiatry