If cytokines are involved in the pathophysiology of depression, their production should be increased in depressed patients. The relationship between depression and immunity is certainly the topic that has been studied the most intensively in clinically oriented psychoneuroimmunology. At the beginning of this research, the expectations were quite high. In concordance with the belief that the mind influences the body, mental depression was viewed as being necessarily associated with immunosuppression. The first results were quite encouraging since assays of mitogen-induced lymphocyte proliferation and natural killer (NK) cell cytotoxic activity carried out on white blood cells confirmed that depressed patients have decreased cell mediated immunity. This decrease was more marked in patients with melancholia than in other categories of patients. Could this be one of the physiological markers of melancholia? These alterations were viewed as the consequence of the hypercortisolemia that characterizes depressed patients and, as expected, the disappearance of hypercortisolemia in response to antidepressant treatment was accompanied by normalization of immune functions. However, this first wave of optimism did not last for very long. After one decade of intense investigation of the impact of depression on immunity, it became quite clear that the results were not consistent, and that the immunosuppression, when it was present, remained within the limits of normal variation. Furthermore, at the immunological level, the concept that mental depression is necessarily associated with immune depression was challenged by findings of an increase in certain markers of immune activation.
In his review of the immune changes that are observed in depressed patients, Irwin (this volume) points out that it is premature to throw the baby with the bath water. From the data available on lymphocyte functions, it is clear that depressed patients have a decreased number of lymphocytes, reduced mitogen-induced lymphocyte proliferation, and decreased NK cell activity. Some of these changes could be due to sleep disorders rather than to depression per se, and they might be associated with an increased sensitivity to certain viruses. Alternatively, a sleep disorder and/or a viral infection may be integral parts of a depressive state. For example, sleep may be affected by immune response to a viral infection (Krueger and Majde, 1994). Molecular and cellular changes may underlie neurobehavioral changes resulting in a specific depressive state. A depressive state may be an integral sum of discernable and quantifiable behaviors "out of tune" with each other. This disharmony may be better understood at one of the more basic levels of organization than the behavioral level. Concerning immune activation, there are less studies that are available, and the results are still, and not surprisingly, heterogeneous, even if it is possible to observe associations between depression and increased number of white blood cells and neutrophils, and enhanced levels of haptoglobin, a positive acute phase protein. The evidence in favor of an enhanced activation of the monocytic arm of the immune response in depressed patients has been previously summarized by Maes and his colleagues (Maes, Smith, & Scharpe, 1995) and is detailed further by Maes; Seidel, Rothermund, & Rink; and Sluzewska in their respective chapters.
There are many methodological problems with this type of research. The most important one is the health status of patients. Normally, patients who are suffering from a concurrent organic disease should not be included in studies on the relationship between depression and immunity. However, the levels of C-reactive proteins that are displayed by some of the depressed patients characterized as having signs of immune activation would be sufficient to consider them as ill even if they do not show any clinical symptom of infection or inflammation. Another important issue is the nature of the assay that is used to assess production of cytokines. Since cytokines are not hormones, with the exception perhaps of IL-6, plasma levels of cytokines are meaningless. The capacity of white blood cells to produce cytokines in response to LPS and interferon-y is a better option, which explains why whole blood assays have been developed for the purpose of measuring cytokine production in clinical settings. However, the validation and standardization of these assays are usually rudimentary, which allows to understand why the favor is still given to more global indexes of inflammation, such as the ratio of positive to negative acute phase proteins. Whatever the case, it still remains that even if production of cytokines can be measured with adequate techniques, this is a peripheral event that does not necessarily correlate with the ability of brain cells to produce cytokines. Overactivity of the brain IL-1 system, induced by icv administration of IL-1 in rats, has been shown to be associated with increased peripheral levels of IL-6 (DeSimoni, Seroni, DeLuigi, Manfridi, Mantovani, & Ghezzi, 1990), but the way this response generalizes to other cytokines and is reflected in production of cytokines by white blood cells has not been investigated. There is clearly a need for further investigation of the relationships between the central and peripheral cytokine compartments.
Independently of the site of production of cytokines, the question of specificity of the alterations in cytokine production that is observed in depressed patients still needs to be addressed properly. Elevated levels of proinflammatory cytokines (mainly IL-6 and its soluble receptor) and biochemical signs of an acute phase reaction have been evidenced in patients diagnosed with schizophrenia, mania, and post-traumatic stress disorder. However, such overlapping results are not uncommon in biological psychiatry, and they make one question the usefulness of correlating biological variables with disease entities that are defined exclusively on the basis of clinical criteria. The challenge is certainly to transcend this arbitrary categorization by identifying the biologically relevant transnosographic dimensions that are obliterated in the DSM-IV classification.
Was this article helpful?