There are several theories about the mechanism of action of chronotherapy in the treatment of depression.
Correction of over-arousal.
In addition to the frequent agitation and insomnia of depressive states, there is PET and fMRI evidence of increased metabolic rates in certain brain areas thought to mediate mood. Wake therapy has been found to reduce the overactivity in these brain areas. [39, 52, 53, 54, 55]
Adjustment of Circadian Rhythms.
There is much data about the presence of disturbed circadian rhythms in depression. This includes the presence of altered sleep-wake cycles, moods that change in regular patterns over daily and seasonal periods, and disturbances in the timing and magnitude of melatonin and cortisol levels and other circadian biomarkers. [44, 51, 56, 57]
Sleep deprivation, bright light therapy and sleep phase advance are all time-specific treatments that modify circadian rhythms. In Seasonal Affective Disorder, researchers have shown that phase delays of melatonin rhythms correlate with depression severity and are corrected by phase-advancing morning light exposure or afternoon/evening melatonin administration. [58, 59]
Improving serotonergic neurotransmission.
There is clinical, biological, and genetic evidence that wake therapy and bright light treatment act on, or through, the brain's serotonin system. [10, 60, 61]
Glutamate is the most prevalent, excitatory neurotransmitter in the human brain. The only other experimental method that induces rapid improvement in depressed mood (intravenous ketamine) is thought to act on the glutamatergic system. 
A recent magnetic resonance spectroscopy study demonstrated changes in brain glutamate that correlated with clinical improvement in response to wake therapy. 
10. Smeraldi, E., et al., Sustained Antidepressant Effect of Sleep Deprivation Combined with Pindolol in Bipolar Depression: A Placebo-Controlled Trial. Neuropsychopharmacology, 1999. 20(4): p. 380-385.
52. Wu, J., et al., Prediction of antidepressant effects of sleep deprivation by metabolic rates in the ventral anterior cingulate and medial prefrontal cortex.[Erratum appears in Am J Psychiatry 1999 Oct;156(10):1666]. American Journal of Psychiatry, 1999. 156(8): p. 1149-58.
53. Wu, J.C., et al., Sleep deprivation PET correlations of Hamilton symptom improvement ratings with changes in relative glucose metabolism in patients with depression. Journal of Affective Disorders, 2008. 107(1-3): p. 181-6.
55. Clark, C.P., et al., Improved anatomic delineation of the antidepressant response to partial sleep deprivation in medial frontal cortex using perfusion-weighted functional MRI. Psychiatry Research, 2006. 146(3): p. 213-22.
58. Benedetti, F., et al., Phase Advance Is an Actimetric Correlate of Antidepressant Response to Sleep Deprivation and Light Therapy in Bipolar Depression. Chronobiology International: The Journal of Biological & Medical Rhythm Research, 2007. 24(5): p. 921-937.
60. Benedetti, F., et al., Influence of a functional polymorphism within the promoter of the serotonin transporter gene on the effects of total sleep deprivation in bipolar depression.[see comment]. American Journal of Psychiatry, 1999. 156(9): p. 1450-2.
63. Benedetti, F., et al., Spectroscopic correlates of antidepressant response to sleep deprivation and light therapy: A 3.0 Tesla study of bipolar depression. Psychiatry Research: Neuroimaging, 2009. 173(3): p. 238-242.