Quantitative pedigree analysis and mitochondrial DNA sequence variants in adults with cyclic vomiting syndrome

Cyclic vomiting syndrome (CVS) is a chronic disorder characterized by episodic nausea and vomiting and is diagnosed using Rome criteria [1]. Subsets of CVS include CVS plus and catamenial CVS. CVS plus is defined by the presence of at least two neuromuscular disorders in association with CVS [2]. CVS can sometimes occur in association with the menstrual cycle and this is referred to as catamenial CVS [3]. The pathophysiology of CVS is not known but preliminary reports in children suggest that genetic factors play an important role. Extensive pedigree analysis of children with CVS has revealed a clustering of functional disorders in matrilineal relatives [4, 5]. There is a lack of similar data on matrilineal inheritance patterns in adults with CVS. Further, analysis of the mitochondrial genome in children has shown an increased frequency of two mitochondrial DNA polymorphisms 16519 T and 3010A, which together confer 17-fold higher odds of having CVS in comparison to controls [6]. The 16519 T polymorphism alone was associated with 6-fold higher odds of CVS [6]. Experts consider CVS as part of a migraine diathesis and mitochondrial single-nucleotide polymorphisms (SNP’s) have been associated with migraines as well [7].

Preliminary data by Boles et al. also suggests that CVS is genetically distinct in those with onset of symptoms < 12 years (pediatric-onset) and those > 18 years (adult–onset). A previous study showed that the association between the 16519 T and 3010A polymorphisms (the AT genotype) reported in children was absent in adults, and it was speculated that the pathophysiology of CVS may differ based on the age of onset of symptoms. This same study also found that 56% of adults with CVS had high Karolinska Scales of Personality (KSP) scores which are both sensitive and specific for mitochondrial dysfunction; findings not reconciled by the same study [8]. And so while there is some thought that CVS is genetically distinct based on onset of disease, the association with migraine, high KSP scores in adult-onset CVS which are specific and sensitive for mitochondrial dysfunction and similar response to treatment suggests otherwise and underscores the need for further analysis [9, 10].

CVS remains a clinical challenge given the absence of a biochemical marker and indeed the validity of this diagnosis is questioned by some. Thus it is imperative that we find surrogate markers for this disease; the presence of a biological marker would enable physicians to diagnose patients with a simple blood test and avoid expensive investigations to exclude various other disorders. This study is also relevant in the treatment of this disorder. CVS is currently treated with tricyclic antidepressants and mitochondrial supplements such as coenzyme Q 10 and L-carnitine. Small trials and anecdotal data showed that high doses of co-enzyme Q10, riboflavin and L-carnitine have potential efficacy in CVS [11]. Identification of mitochondrial heteroplasmies may allow us to develop therapies towards mitochondrial dysfunction to treat this and other related disorders.

We thus sought to perform a Quantitative Pedigree Analysis (QPA) using a standard questionnaire and estimate the matrilineal inheritance pattern of various functional disorders in adults with CVS. We also sought to identify the frequency of mtDNA SNP’s, 16519 T and 3010A in adults and determine if there are any significant associations with inheritance patterns and age of onset of disease.

Patients who met Rome III criteria for CVS were recruited through the CVS Clinic at the Medical College of Wisconsin. Other potential subjects were also recruited through the Cyclic Vomiting Syndrome Association (CVSA) website and message board. All subjects not seen at the Medical College of Wisconsin (MCW) were invited to contact the principal investigator’s office at MCW in person, by e-mail or postal mail in order to participate, and only those who met Rome III criteria for CVS were included. Extensive clinical data was collected using the standard questionnaire employed by the CVS clinic at MCW and saliva samples were collected in person or mailed in by the subject. The Institutional Review Boards at the Medical College of Wisconsin, University of North Carolina (UNC) and the Children’s Hospital of LA (CHLA) all approved the study. Informed consent was obtained from all subjects and a legal guardian when applicable. All subjects were > 18 years of age at the time of the study.

Saliva was collected directly by the CVS probands using a commercially available kit (Oragene, DNA Genotek Inc., Ottawa, Ontario) and then mailed to Children’s Hospital Los Angeles (CHLA) where genetic analysis was performed. The presence or absence of the 7028C polymorphism that defines haplogroup H was determined in subjects either by cyclosequencing or by RFLP following AluI digestion, as per our routine practice. Both methods were used but sequencing was used during the latter part of the study as this was more readily available and less expensive than RFLP. These methods were extensively cross-validated in RB’s laboratory. All SNP’s, including all sequence variation in the CVS subjects and any SNP’s identified were correlated with the extensive clinical data obtained.

A detailed pedigree analysis was performed using a standard scripted questionnaire (see Additional file 1). The number of functional conditions per matrilineal relative and patrilineal relative were calculated and the maternal ratio and maternal inheritance ratio were calculated using a previously published tool called Quantitative Pedigree Analysis (QPA) which is used to assesses patterns of inheritance [12]. A patient had a probable maternal inheritance (PMI) pattern if the maternal ratio was ≥1.75 and the maternal inheritance ratio (MIR) is ≥ 3 where the MIR is calculated by dividing the maternal and paternal ratios. The inheritance pattern was defined as being indeterminate (Ind) if the maternal ratio was 1.5 - <1.75 and/or the MIR was 2 - <3, and as probable non-maternal inheritance (PNMI) if either the maternal ratio was <1.5 and/or the MIR is < 2. If the patient had cognitive impairment one or both parents/caregivers were asked to provide the information on the patient’s behalf.

Quantitative Pedigree Analysis was performed in 195/216 patients with cyclic vomiting syndrome who were genotyped and 102 healthy controls. Healthy controls were recruited through mass electronic mail among the faculty, staff and students at the University of North Carolina (UNC), as well as through the UNC Center for Functional GI and Motility Disorder registry of study participants, and the International Foundation of Functional Gastrointestinal Disorders website.

Our study group was characterized based on the onset of symptoms; pediatric-onset CVS was defined by onset of stereotypical vomiting episodes at <12 years of age, adolescent-onset CVS if episodes began between 12 and 18 years of age, and adult-onset CVS if episodes began ≥18 years of age. Adolescent onset CVS patient were excluded from the analysis as they likely represent an indeterminate group given the wide variation in physical, emotional and cognitive maturity during this period.

This is the first and largest study to our knowledge in adults with CVS to determine the presence and degree of maternal inheritance in adults with CVS in a systematic manner. A detailed family history was ascertained in both first and second degree relatives of the index patient. We also sought to determine if there were any correlation between inheritance patterns and mitochondrial DNA SNP’s 16519 T and 3010A, which have been associated with increased odds of having CVS in children. The findings of our study reveal that adults with CVS patients have a higher proportion of probable maternal inheritance of various functional disorders in comparison to controls (13% vs 1%, p value <0.0001). There was no association between maternal inheritance patterns and genotype as expected. Restricting our analysis to haplogroup H (71/216, 33%) revealed no significant differences in the prevalence of 16519 T, 3010A and the AT genotype in comparison to historical controls. There was also no difference in the prevalence of these mtDNA SNP’s in adult and pediatric-onset patients. While this was not the primary intent of the study, other important findings in this study include the high percentage of patients (28%) who underwent surgery for CVS symptoms. This underscores the need for prompt diagnosis and appropriate therapy in this patient population.

Quantitative Pedigree Analysis (QPA) is a novel tool used for identifying matrilineal inheritance patterns and we obtained a detailed history of all first and second-degree relatives of adults with CVS in our cohort using a standard questionnaire. This revealed that the vast majority of adults (87%) had a non-maternal inheritance pattern of functional symptomatology. A small subset of adults (13%) had a probable maternal inheritance which was significantly different from controls (13 vs 1%, p < 0.001). However there was no correlation between the patterns of inheritance and mtDNA SNP’s 16519 T and 3010A as expected; this could be explained by the presence of yet unidentified mitochondrial polymorphisms beyond the two that were studied. While DNA methylation may be another possible explanation this would be less likely and needs to be explored in future studies.

Our genetic analysis of Haplogroup H adults with CVS (33%, 71) did not show any differences in the frequency of 16519 T, 3010A and the AT genotype between our cohort and historical controls. We were unable to comment on our non-H group as we did not have a control non-H population readily available for comparison. Age of onset was also not a determinant of the presence of the mtDNA SNP’s as reported in a previous study. The study by McCallum et al. revealed that pediatric onset CVS patients were more likely to have 16519 T and the AT genotype. One possible explanation for these differences between our study and the previous one may be due to the fact that most of the childhood onset patients in the above mentioned study were recruited through a genetic clinic which may have resulted in a referral bias. Twelve % (17) of patients in the non H group had the AT genotype but the significance of these results is unclear with this present study. Most of the studies in CVS have been performed in Haplogroup H patients as they are genetically homogenous.

A recent study by Camilleri et al. examined the frequency of irritable bowel syndrome (IBS) in both Haplogroup H and non- H patients. This interesting study revealed that the IBS –C and IBS with alternating constipation and diarrhea was more prevalent in non H individuals. In this same study, Haplogroup H individuals had higher maximal tolerated volumes (lower satiety) and 3010A in these individuals was associated with abdominal pain/dyspepsia and more rapid emptying [13]. The frequency of the AT genotype in this non H group was not readily available but future collaborative studies pooling such data would serve to make more meaningful comparisons.

The strengths of our study are that this represents a large sample of adults with CVS with extensive clinical information obtained prospectively. Pedigree analysis was conducted using a standard script, ensuring uniformity in data collection. Limitations include the possibility of recall bias when obtaining family histories and that most of the information was usually gleaned from the patients and their mothers, which may have led to missing information on the paternal side. However, every effort was made to address this and patients were called on multiple occasions if necessary to clarify the family history. However, the same limitations apply to the control group as well and are unlikely to have influenced overall results. Other limitations include the lack of a gold standard for diagnosing CVS as with other functional disorders. While this was a large heterogeneous group of CVS patients, we cannot exclude a referral bias with sicker patients being included. Arguably these are the very patients who would need more aggressive treatment, and understanding the mechanisms underlying CVS would be even more important with this population. Other considerations are that we chose to include only haplogroup H patients for genotype comparison with controls as non-H control data was not readily available. Given the changing demographics of the US population, future studies should be extended to include both groups of patients. The presence of a high degree of matrilineal inheritance of functional disorders was not associated with a specific genotype as expected and may be due to yet unidentified mtDNA sequences. Other factors beyond the mtDNA sequence, possibly including nuclear-encoded mitochondrial genes, other genes, and the environment may play a larger role in the pathophysiology of CVS in adults as opposed to children.

Our study demonstrates that CVS in a subset of adults is associated with a maternal inheritance pattern of functional GI disorders as seen in children but is not associated with 16519 T, 3010 A or the AT genotype. A matrilineal inheritance pattern was higher in pediatric onset CVS compared to adult onset CVS. However, onset of disease did not have any correlation with the frequency of the mtDNA SNP’s. While this study offers some insight into the genetic background of CVS, further studies examining the entire mtDNA, as well as the full mitochondrial exome (the full mtDNA sequence plus nuclear genes encoding mitochondrial proteins) are warranted.