Robert Sapolosky and Bruce McEwen do a lot of stress and depression work. There is a good layperson summary you can find here. http://consumer.healthday.com/encyclopedia/depression-12/depression-news-176/depression-and-stress-645983.html

I think that the relationship is going through the influence of glucocorticoids on neuroplasticity, atrophy of hippocampus and prefrontal cortex, and the effect on the neurotransmitters - serotonin and dopamine.

Melancholic depression is a type of depression that is most strongly associated with hypercortisolemia and DST nonsuppression. I think that in this type of depression cortisol is what makes the most damage, which appear in the prefrontal cortex dysfunction (apathy, lost of feelings, facial expressions) and hippocampus (poor memory).

As association is different from a cause. I believe cortisol is commonly used as a marker of anxiety. This means that it results from an anxious state of mind, not that it causes an anxious state of mind.

Changed patterns of serum cortisol have been observed in connection with abnormal ACTH levels, clinical depression, psychological stress, and physiological stressors such as hypoglycemia, illness, fever, trauma, surgery, fear, pain, physical exertion, or temperature extremes. Cortisol does not cause all these states.

relationship between cortisol and depression, means the presence of intimate relation between stress and depression.The connection between stress and depression is complex and circular. Repetitive or sustained stress can lead to major depression in susceptible people. It leads to overactivity of the body's stress-response mechanism. Indeed, it leads to elevated hormones such as cortisol, the "stress hormone," and reduced serotonin and other neurotransmitters in the brain, including dopamine, which has been linked to depression. Stress is known to affect negatively via cortisol or equivalent in animals, neurogenesis in hippocampus (anxiety) and dorsal vagal complex (anorexia nevrosa-like) (F. Chigr et al., Neurogenesis inhibition in the dorsal vagal complex by chronic immobilization stress in the adult rat.. Neuroscience. 2009 Jan 23;158(2):524-36. doi: 10.1016/j.neuroscience.2008.10.040.

See also:

-Neuropsychopharmacology. 2003 Jan;28(1):140-7.

Anxious-retarded depression: relation with plasma vasopressin and cortisol.

de Winter RF1, van Hemert AM, DeRijk RH, Zwinderman KH, Frankhuijzen-Sierevogel AC, Wiegant VM, Goekoop JG.

-Psychosom Med. 1981 Jun;43(3):235-42.

Cortisol secretion in relation to age in major depression.

Asnis GM, Sachar EJ, Halbreich U, Nathan RS, Novacenko H, Ostrow LC

Cortisol does not cause depression - there is an increase in about 50% of depressed patients  only and a correlation in some studies when depressed patients are looked at as a whole - but correlation does not imply causation. 

It is also important to emphasize that the HPA axis is activated in order to maintain homeostasis.

The question then is, what is the homeostasis disrupting event seen in depression?

On my analysis this is a single release of cytokines into the Third ventricle - that causes damage to the surrounding structures, damage that in turn produces the behavioural syndrome we call depression.

Cortisol release is secondary to this and the variations in levels seen very much depend on the timing of the measurement following the cytokine release

https://www.academia.edu/7143894/The_Third_Ventricle_Hypothesis_of_depression_A_thumbnail_sketch

Hendrie CA, Pickles AR.(2013) The failure of the antidepressant drug discovery process is systemic J Psychopharm 27: 407-416

Colin, thank you for once again reminding us of the important and basic fact that association does not imply causation.

However, in your paper you state, "Depression is not however a permanent state."

I'm not certain that I'm not reading this out of context, but that most definitely is not true as stated. While some individuals have isolated episodes of major depression that are symptomatically mild and that each disappear without treatment, the majority of patients with depression, especially with dysthymia, will suffer from their symptoms for many years, if not for their lifetime without treatment. in addition, a large number of patients with depression continue suffering at the same or a reduced level even with skillful treatment (treatment resistance).

This is the reason that psychiatry, which is evidence-based, considers it a social responsibility to offer patients psychotherapy, psychopharmacology, and other interventions that have some measure of success in managing depressive disorders. Between 80 percent and 90 percent of people with depression respond positively to treatment, and almost all patients gain some relief from their symptoms.

I think your hypotheses involving cytokines and evolutionary adaptation are interesting and worthy of research, but please make certain that you fully reflect the actual clinical picture in your statements. Depression is not just feeling bad for a few days--it kills people. Over 50 percent of all people who die by suicide suffer from major depression. If one includes alcoholics who are depressed, this figure rises to over 75 percent.

Disclaimer: I am neither a researcher in depression, nor a mental health professional. I am familiar with depression through my teaching and research in meditation.

Colin A Hendrie, what is the source of cytokines in this case?

Thanks for comments David - I have full knowledge of the clinical situation and am fully aware of the seriousness of the disorder

The point being made by saying that depression is not a permanent state is that *for most* the symptoms will lift after a period - 6 months or a year or even longer - the length of time is not however crucial - what is crucial is that *for most* the symptoms of depression are relieved over the passage of time and therapists are not needed to make this happen.

I am interested that you say that between 80 and 90% of depressives respond to treatment - do you have references for this? The more commonly accepted figures are that only 60-70% respond - if you look over the time when drug companies were interested in depression, going down to 40% now that they are not and unpublished data from clinical trials are available.

40% is, unfortunately, roughly same level of effect produced by placebo ...

See for example Figure 2 from Khan A, Faucett J, Lichtenberg P, Kirsch I, Brown WA (2012) A Systematic Review of Comparative Efficacy of Treatments and Controls for Depression. PLoS ONE 7(7): e41778

http://journals.plos.org/plosone/article?id=10.1371/journal.pone.0041778

Jacek, the source of the cytokines is not known but the suggestion is that these are released directly into the ventricle