2.2. Survey
This study was approved by the Institutional Review Board (IRB) of the Greenville Health System (GHS) (ID: Pro00061122). The survey utilized in this study was designed by our research group based in part on previous informal survey studies performed by the Autism Research Institute (ARI). These questions were further adapted and expanded according to an additional literature search of relevant clinical symptomology for our topics of interest. (See Supplementary File 1 for full survey.)
The survey was built on and hosted by the website, SurveyGizmo.com, and was advertised via the ARI newsletter; the ASD forum, Wrong Planet; and a variety of FaceBook ASD- and EDS-specific webcommunities, such as the Autism Women’s Network (AWN), the Autism Spectrum Women’s Group, AutismTalk, the Ehlers-Danlos Support Group, and Ehlers-Danlos Worldwide. The administrative teams of all participating web communities were informed that the survey was IRB-approved and were given access to the survey and, when requested, a copy of the IRB protocol prior to approval. Following administrator approval, either ELC posted the survey announcement or administrators posted it themselves. The survey weblink (
www.autismwomensstudy.com) led potential respondents to a description of the purpose and expectations of the study, potential risks and benefits, investigator contact information, and a waiver of consent. The survey was open and participants were actively recruited for approximately three months.
Survey questions focused on topics concerning ASD and EDS/GJH diagnoses; symptoms involving the immune and endocrine systems; chronic pain; GI dysfunction such as IBS; seizures; and limited aspects of medication history (hormone treatment, antiseizure medications). Additional topics were covered but were not used for the current study.
Questions on immune symptomology included items concerning hospitalization history; respiratory disorders like asthma, allergies, sinusitis/rhinitis, and reactions to medications or environmental chemicals; and autoimmunity [19,20,21]. Hormone-mediated symptoms included items such as chronic irregularities in menstruation and associated pain syndromes; PCOS; and other clinical symptoms indicative of the metabolic syndrome, such as type 2 diabetes/insulin resistance, hypertension, and high cholesterol [11,22]. The data of respondents who agreed to participate but failed to complete the survey were discarded.
2.3. Statistical Analyses
When assessing group differences for quantitative variables (e.g., age, immune- and endocrine-mediated symptoms), Welch two-sample t-tests were conducted unless the distributions were heavily skewed, in which case the Wilcoxon rank sum test with continuity correction was used. Two sample tests of proportions were used to compare groups for binary categorical data (e.g., presence/absence of diabetes, infertility, etc.). Fisher’s Exact Test was used in cases of small sample sizes. Where appropriate, a false discovery rate adjustment was used to account for multiple comparisons. A significance level of 0.05 was used for all analyses.
3. Results
3.1. Immune-Mediated Disorders
Although women with ASD and ASD/GJH did not differ in the presence of one or more immune-mediated symptoms (χ2 = 1.162, p = 0.281), ASD/GJH women were, however, more likely to report multiple symptoms (t = −3.860, df = 30.981, p = 0.001), an effect that differed by age (W = 534.5–563, p = 0.009–0.017) (Figure 2A). Women with ASD and ASD/GJH also reported similar proportions of specific immune exophenotypes (χ2 = 0.788–4.744, p = 0.137–0.375), with an overall trend towards higher proportions in the ASD/GJH group. However, one exception concerned autoimmune disorders: while 13% of the ASD group had an autoimmune disorder, 45% of women with ASD/GJH reported the same (χ2 = 8.813, p = 0.027) (Table 2).
Behavsci 08 00035 g002 550
Figure 2. (A) Number of immune-mediated symptoms across ASD and ASD/GJH groups. ‘Any immune’ = 1 or more immune symptoms. (B) Network of immune-mediated symptoms. (C) Number of endocrine-mediates symptoms across ASD and ASD/GJH groups. ‘Any endocr.’ = 1 or more endocrine symptoms. (D) Network of endocrine-mediated symptoms.
Table 2. Reported rates of various immune-related symptoms according to group, as well as estimated general prevalence rates.
Table
The presence of most immune exophenotypes exhibited a significant association with one another, suggesting that a similar etiological background underlies many of these symptoms. Allergies, rhinitis, sinusitis, asthma, ear infections, reaction to medications, and reaction to environmental chemicals all seemed to share a strong interrelationship (p < 0.001–0.020). Meanwhile, autoimmunity was significantly associated with ear infections (CI = 1.509–18.898, p = 0.014) and showed a trend towards significance with asthma (CI = 1.003–9.582, p = 0.059) (Figure 2B).
Interestingly, the proportions of IBS/GI dysfunction did not differ significantly between groups, though the ASD/GJH group reported modestly higher rates (χ2 = 0.648, p = 0.946). In spite of IBS’ links with immunity, it did not share a significant relationship with immune exophenotypes in general (CI = 0.282–7.60, p = 0.717), although our study may have been too underpowered to glean an effect [31]. Yet in spite of modest numbers, IBS/GI dysfunction was significantly linked with hormonal exophenotypes: individuals with IBS/GI dysfunction had more hormone-mediated symptoms on average than those without (W = 811, p = 0.002).
In spite of the previously reported relationship between hormones and seizure propensity, the presence of complex hormonal exophenotypes was not associated with epilepsy in our cohort (W = 274, unadj. p = 0.3742) [32]. However, despite the small number of women reporting epilepsy (N = 7), epilepsy shared a modest positive relationship with the number of immune-mediated disorders irrespective of the group (W = 166, unadj. p = 0.022). While the average number of immune-mediated disorders reported across the entire cohort was approximately 3.6 (SD = 2.45), women with epilepsy averaged approximately 5.7 (SD = 2.43).
Finally, joint pain was reported in all cases of ASD/GJH compared to 29% in the ASD group (χ2 = 30.122, p < 0.001). Meanwhile, differences in joint pain were not accounted for by age (W = 559.5, p = 0.101) or obesity (χ2 = 0, p = 1.000). Other types of chronic pain were also reported more often in ASD/GJH (75% vs. 31%) (χ2 = 11.072, p < 0.001), including conditions such as fibromyalgia [33].
3.2. Hormone Disorders
Though the ASD/GJH and ASD groups did not differ in the presence of one or more hormone-mediated disorders (χ2 = 0.728, p = 0.394), ASD/GJH women reported significantly more symptoms than their non-GJH counterparts (W = 434, p = 0.001) (Figure 2C). On a symptom-by-symptom basis, ASD/GJH women reported higher rates of endometriosis (χ2 = 9.265, p = 0.018), dysmenorrhea (χ2 = 19.599, p < 0.001), and severe teen acne (χ2 = 7.817, p = 0.026) (Table 3). Dysmenorrhea, in particular, was reported three times more often (85% vs. 28%) in ASD/GJH compared to ASD (χ2 = 19.60, p < 0.001), a frequency similar to that reported in previous EDS research (see Table 3). In its extreme form, dysmenorrhea is typically associated with endometriosis and both share links with immune dysfunction in the general population [34,35].
Table 3. Reported rates of various endocrine-related symptoms according to group, as well as estimated general prevalence rates.
Table
We found no significant interaction between group and birth control/hormone treatment in relation to the average number of hormone-mediated symptoms, indicating that such treatment is an unlikely group confound in this study. There was, however, a significant relationship between the number of hormone-mediated symptoms reported and whether an individual was receiving some form of hormonal treatment (t(101) = 2.75, p = 0.004). While we cannot rule out a potential confound, instead, we conclude that this is likely a reflection of the severity of endocrine disorders in our cohort and their prescribed treatments [51].
Like immune symptoms, individual hormone-mediated symptoms were often associated with one another (Figure 2D). PCOS shared links with other symptoms, including diabetes, adult acne, irregular menses, and hirsutism, all of which are either diagnostic of or commonly reported in PCOS (p = 0.005–0.045). In contrast, infertility, overweight/obesity, amenorrhea, hypertension, and high low-density lipoprotein (LDL) cholesterol did not associate with PCOS in our groups (p = 0.073–0.809) [52]. There was, however, a trend towards significance between PCOS and infertility, suggesting our data may have been underpowered, requiring a larger pool of respondents in the future (OR 95% CI = 1.20–37.800, p = 0.073). Endometriosis and dysmenorrhea were also associated with one another (CI = 2.229–785.323, p = 0.013).
3.3. The Relationship between Immune- & Hormone-Mediated Symptoms
There was no significant association between the general presence of immune- and endocrine-mediated disorders across our cohort (OR 95% CI = 0.538, 21.119, p = 0.098). However, the number of immune-mediated symptoms per individual greatly predicted the number of hormone-mediated symptoms (Spearman’s rho = 0.35, p < 0.001). This suggests that the complexity and severity of immune- and endocrine-mediated disorders share a strong positive relationship with one another in autism and potentially within the general population, e.g., [53].
4. Discussion
The present study attempts to address phenotypic differences between ASD women with and without GJH. This research supports a growing body of literature indicating that immune-mediated disorders are a common comorbid feature in hEDS and GJH. In addition, we have also shown that this dysfunction may be paired with endocrine dysregulation, leading to complex immune and hormonal exophenotypes, such as autoimmune disorders, allergic rhinitis, asthma, endometriosis, and dysmenorrhea. While we have not addressed autism and GJH comorbidity rates in this study, their co-occurrence in the adult ASD female population suggests links between the dysfunction of connective tissue and the immune and endocrine systems in this subpopulation.
As discussed, the immune system has been a popular area of investigation in autism research. However, reports of clinical manifestations in the child population seem to vary [54,55,56,57]. Some clinical manifestations arise during or progress in severity with the advent of puberty, highlighting the role the endocrine system plays in immune function, e.g., [58]. In addition, women are more frequent targets of such dysfunction, suggesting that studying immune dysregulation in prepubertal individuals with autism, while also ignoring gender confounds, dramatically underrepresents the frequency of clinical symptoms in the autism population [19]. For these reasons, we limited our study population to women aged 25 years or older on the autism spectrum.
4.1. Immune-Mediated Disorders in Association with Connective Tissue Disorders
Loeys-Dietz Syndrome (LDS) is a connective tissue disorder caused by mutations directly targeting the TGF-β pathway and is characterized primarily by enlargement of the aorta. People with LDS have high rates of immune-mediated disorders such as respiratory and food allergies and occasionally present with Hyper-IgE Syndrome, a type of primary immunodeficiency [59]. In addition, they also share many of the same dysmorphic features as those seen in the connective tissue disorder, Marfan Syndrome (MFS) [60].
Although MFS is associated with mutations in the Fibrillin-1 (FBN1) gene whose protein product is a component of the extracellular matrix (ECM), FBN1 mutations lead to marked TGF-β dysregulation [61,62,63]. Fibrillin appears to control the activity of TGF-β by acting as a structural platform for the Latent TGF-β Binding Protein (LTBP) that sequesters and inactivates TGF-β, acting as a reserve pool for rapid injury response [64]. Given its role as a foundational morphogen, it is believed this overlap in TGF-β pathway dysregulation leads to the overlapping features of MFS and LDS [65].
Like LDS, some individuals with hEDS present with a Marfanoid (Marfan-like) habitus [66]. However, unlike MFS that results from dysfunctional fibrillin, EDS is typically linked with dysfunction of the ECM protein, collagen. Marfan and Marfanoid features in all three of these disorders suggest considerable overlap and interaction between the ECM and the TGF-β pathway. In addition, TGF-β serves as a link between the ECM and immune system disruption as it is a key immunomodulator, implicated not only within the joints in these connective tissue disorders, but also in other organ systems such as the lungs [65]. Interestingly, several studies have consistently found lower TGF-β1 levels in autism, which according to Ashwood et al. [67], may help explain some of the immune dysregulation in the condition [67,68]. For these reasons, the TGF-β pathway and upstream networks may be prime areas of study for future work into the overlapping etiologies of both connective tissue disorders and autism.
4.2. The Effects of Estrogen on Collagen Production & the Immune System
Similar to certain immune disorders like autoimmunity, GJH and hEDS preferentially target women for reasons not well understood [69]. One possibility may stem from sex differences in muscle mass, in which stronger muscles help to counteract joint laxity and ensuant pain [16]. For this reason, one of the foci of physical therapy in the treatment of GJH/hEDS centers around improved muscle strength surrounding problem joints [70]. However, female-specific effects may result not only from low testosterone levels, but also estrogen metabolites that either suppress collagen production directly, particularly within the skin, or result in a more rapid turnover of collagen within tendons and ligaments [71,72,73].
Estrogen is also a major immunomodulator. It is capable of driving activation of the Th2 branch of the immune system, boosting humoral immunity and the ability of the body to target parasites and other extracellular infections. Estrogen also stimulates mast cell degranulation, prompting a release of chemicals such as histamines, TNF-α, various amines, chymase, and tryptase [74,75]. Mast cell activation, in turn, may drive both Th1/Th2 immune responses depending on the invading pathogen, the target tissue, and other variable factors [76].
Interestingly, estrogen also increases the synthesis of TGF-β within numerous cell types, the latter of which is itself a key morphogen and immunodulator. In addition, estrogen further interacts with the TGF-β pathway by forming a complex with Smad 3/4, redirecting TGF-β target genes. Finally, TGF-β and estrogen are able to interact at the level of various Ras complexes, by which TGF-β enhances estrogenic action [77]. All of these data together suggest significant interaction of estrogen with various networks implicated in connective tissue disorders and their secondary symptoms.