The primary finding of the present research was that monocytes in healthy male PTSD subjects showed a predominant pattern of decreased gene expression. These results are consistent with the finding of diminished gene expression of transcription factors reported by Segman et al (Segman et al., 2005
). Notable categories of decreased expression pertain to cytokine/chemokine signaling, platelet function, and histone activity. Thus, there was no evidence for chronic inflammation in our male PTSD subjects. In contrast, preliminary data from our pilot sample of women subjects with PTSD showed a relatively balanced pattern of increased and decreased expression of genes with ontology suggestive of increased activity of pathways related to immune activation. We did not replicate the published finding of decreased expression of FKBP5 in either our male or female samples. Overall, our results indicate altered monocyte gene expression profiles in males and females with PTSD, and are suggestive of gender dimorphism in biologic pathways activated in PTSD, particularly in the domain of immune function.
Diminished gene activation in immune cells in PTSD has been reported in both published studies (Segman et al., 2005
; Yehuda et al., 2009
) and a third unpublished study (Kerry Ressler, personal communication). In the Segman el al study (Segman et al., 2005
), PTSD subjects at 4 months post trauma showed a similar pattern of diminished transcription in PBMCs relative to controls. In our secondary analysis of the publically available data from this study, we found that with a cut-off of 50% difference in expression levels and using p
values less than 0.05 (without corrections for multiple testing), 386 transcripts were downregulated and only 56 were upregulated. Using Ingenuity Pathways Analysis software suite (Ingenuity Systems) to annotate gene function in this data set, we found 28 known transcription regulators were under-expressed versus 2 that were over-expressed. Yehuda and colleagues found that PTSD subjects differed from controls in 17 transcription factors, 15 of which were under-expressed (Yehuda et al., 2009
) in whole blood. In addition, Miller et al found in a sample of chronically stressed caregivers, that monocyte genes that were differentially expressed by at least 50%, were predominantly in the direction of under-expression (79% vs. 21 %) (Miller et al., 2008
). Recently, Segman and colleagues found in a study using microarrays of PBMCs sampled from postpartum women, that women with post partum depression had marked diminished gene expression in a broad range of markers (Segman et al., 2010
). It is possible that PTSD, chronic stress, and postpartum depression each have broad general effects on overall gene transcription in circulating immune cells, which by function are supposed to exhibit a high degree of plasticity. It suggests that these disorders may involve adaptations that reduce cellular reactivity to homeostatic challenges, and if proven, may account for the high rates of comorbid medical illness associated with these conditions (Cohen et al., 2007
; Evans et al., 2005
; Schnurr et al., 2000
Our results in the male PTSD subjects did not show a pro-inflammatory pattern as was previously reported by Miller and colleagues in chronically stressed caregivers (Miller et al., 2008
). This difference may be related to two issues. The first is that biological aspects of chronic PTSD maybe fundamentally different from those associated with chronic stress exposure as has been emphasized by several investigators (e.g. (Yehuda and McFarlane, 1995
)). Second, the sample of chronic caregivers in the Miller study was predominantly women. Our results suggest that gender maybe a crucial factor for examining the role of inflammation in the disease burden associated with PTSD.
In contrast with findings in our male sample with PTSD, we found that our pilot sample of women with PTSD had gene expression patterns indicative of greater immune activation compared with women without PTSD. Previous research has shown differences in stress-related inflammatory responses between males and females, but the pattern of findings has been mixed and apparently related to the menstrual phase and menopausal status of the female participants (Prather et al., 2009
; Rohleder et al., 2001
). Sex hormones may mediate these effects because immune cells bear receptors for the hormones estrogen and progesterone, and levels of these hormones have been found to modulate inflammatory activity and immune cell gene expression (Burger and Dayer, 2002
; Kramer and Wray, 2002
). Female gender is a well-established risk factor for both PTSD (Kessler et al., 1995
) and autoimmune disorders (Fairweather et al., 2008
). Further research, dedicated to understanding the transcriptional control pathways mediating observed patterns of gene expression in women and men with and without PTSD, will be necessary to begin examining potential mechanisms of the sex differences in gene expression observed in the present study.
This study has a number of limitations including a limited sample size, cross-sectional design, and genetically heterogeneous population. Further, our small female PTSD− control sample included a mix of trauma exposed and trauma unexposed women meaning that our findings of menstrual phase gender differences must be taken with caution until replicated. Further, we did not control for in our female sample. Another limitation relates to the presence of comorbid major depressive disorder (MDD) in approximately a quarter of our male and female subjects. PTSD and MDD have common overlapping criterion symptoms, and hence are by definition are intricately intertwined. Covarying for the presence of MDD is not a satisfactory statistical approach to this problem because presence of MDD increases with PTSD symptom severity. PTSD symptoms residualized for presence of MDD represents variance attributable to milder severity of PTSD. In the absence of a robust biomarker which distinguishes PTSD from MDD, comborbid MDD remains a limitation of this study. Significant strengths of this study include examination of subjects on the day of blood draw, rigorous screening for chronic illness, leukocytosis screening and the selection of subjects who were afebrile and medication free. Moreover, our selection of CD14+ monocytes for analysis as a homogenous population of immune cells limits sources of variability not related to group assignment and thereby extends previous research that has focused on heterogeneous populations of cells. It is possible that biological pathways relevant to PTSD that are co-expressed in the brain and circulating immune cells may be discernible only in specific cell types. In order to discern these cell-specific patterns, it is necessary to analyze homogenous cell populations as we have done.
Differential gene expression in male subjects with PTSD was disproportionately in the direction of under-expression of genes across functional classes (within the Gene Ontology). There was no evidence for chronic inflammation in male PTSD subjects. In contrast, preliminary data from our pilot sample of female subjects with PTSD showed a different profile of gene expression, particularly in pathways related to immune activation. The results are suggestive of gender dimorphism in biologic pathways altered in PTSD particularly in the domain of immune function. In sum, our data indicate that monocyte gene expression patterns are altered in PTSD, supporting the idea that changes in monocyte gene expression could contribute to the increased medical morbidity observed in PTSD.
Male subjects with PTSD had an overall pattern of under-expression of genes on monocytes (47 under-expressed versus 4 over-expressed genes).
There was no transcriptional evidence of chronic inflammation in male PTSD+ subjects.
In contrast, preliminary data from our pilot sample of female PTSD+ subjects showed a relatively balanced pattern of increased and decreased expression of genes and an increase in activity of pathways related to immune activation.
The results indicate differential patterns of monocyte gene expression in PTSD, and are suggestive of gender dimorphism in biologic pathways activated in PTSD.