This study evaluated the hypothesis that traumatic stress can increase risk of bulimia nervosa (BN) in individuals who are genetically disposed towards lower modulation of physiological stress reactions. We explored the extent to which childhood abuse (physical or sexual), variants of a main glucocorticoid receptor (GR) polymorphism (Bcl1), or their interaction, differentiated women with and without BN. Women seeking treatment for BN (N = 129) and non-eating-disordered comparison women (N = 98) provided blood samples for assays of the Bcl1 polymorphism, and completed structured interviews assessing eating symptoms, psychiatric symptoms and childhood abuse. Compared to normal-eaters, bulimic women were significantly more likely to carry the low-function Bcl1 C allele (CC or CG genotypes), to report a history of childhood abuse and, more importantly, to be positive for both factors. We interpret our findings as indicating that traumatic stress, when impacting individuals disposed to lower GR modulation, can be etiological for BN.
Current etiological theories postulate that bulimia nervosa (BN) often implicates the activation of hereditary susceptibilities by environmental forces (e.g., Steiger, 2004; Bulik, 2005; Steiger and Bruce, 2007). In this study, we evaluated the hypothesis that traumatic stress can increase risk of BN in individuals who are genetically disposed towards lower inhibition of hypothalamic–pituitary–adrenal (HPA) axis stress responses.
Modal figures from clinical samples suggest that about 30% of adults with a bulimic syndrome report unwanted sexual experiences during childhood, whereas over half report physical maltreatment in childhood (see Wonderlich et al., 1997 and Steiger and Bruce, 2008). Effects of childhood abuse have been thought to be enacted, in part, via stress-induced alterations in the functioning of the brain's main stress–response system, the HPA axis—exposure to traumatic stress having been linked to anomalous HPA-axis activity in eating- (Steiger et al., 2001 and Díaz-Marsá et al., 2007) and non-eating-disordered populations (Stein et al., 1997, Wüst et al., 2004 and Yehuda et al., 2002). HPA-axis findings in BN are variable (see Brambilla and Monteleone, 2003), but a modal result seems to be “hypercortisolemia” (i.e., unusually high baseline cortisol levels)—which could reflect relatively low inhibition of cortisol release by HPA-axis glucocorticoid receptors (GRs).
The NR3C1 gene (located on chromosome 5q31) affects GR sensitivity and, in turn, strength of inhibitory feedback within the HPA axis (DeRijk et al., 2002). The most widely studied of polymorphisms associated with GR expression is the Bcl1 restriction fragment length polymorphism, a C/G single nucleotide polymorphism in intron B, 646 nucleotides downstream of exon 2. Suggesting that Bcl1 mediates inhibitory feedback within the HPA axis, Bcl1 high-function (G) allele carriers have been observed to display reduced cortisol responses following psychosocial stressors (Wüst et al., 2004 and Kumsta et al., 2007) and greater suppression of cortisol after dexamethasone (van Rossum et al., 2003). Such findings lead to the conjecture that traumatic experiences might have more pronounced effects in individuals whose genetic propensities code for lesser modulatory feedback within the HPA axis. Consistent with the preceding, changes in methylation that would cause reduced hippocampal NR3C1 expression have been observed in suicide completers with a history of childhood abuse (McGowan et al., 2009). Likewise, corticotropin-releasing hormone receptor genes have been reported to modulate proneness to depression in previously traumatized adults (Bradley et al., 2008 and Tyrka et al., 2009).
The Present Study: We tested the hypothesis that bulimic individuals would, more frequently than normal-eaters, carry the low-function allele of the Bcl1 polymorphism (associated with relatively low GR feedback) and report severe childhood stressors (i.e., sexual or physical abuse). As other syndromes that co-occur frequently with BN (notably major depressive disorder, some anxiety disorders and post-traumatic stress disorder) also implicate altered HPA-axis activity ( Stein et al., 1997, Wüst et al., 2004, Ströhle and Holsboer, 2003, Belmaker and Agam, 2008, DeRijk et al., 2008 and Yehuda et al., 2002), we took steps to explore and control potential confounding effects of these other psychiatric entities.
