Repeat polymorphisms in the Homo sapiens heme oxygenase-1 gene in diabetic and idiopathic gastroparesis

Abstract

Background: Idiopathic and diabetic gastroparesis in Homo sapiens cause significant morbidity. Etiology or risk factors have not been clearly identified. Failure to sustain elevated heme oxygenase-1 (HO1) expression is associated with delayed gastric emptying in diabetic mice and polymorphisms in the HO1 gene (HMOX1, NCBI Gene ID:3162) are associated with worse outcomes in other diseases.

Aim: Our hypothesis was that longer polyGT alleles are more common in the HMOX1 genes of individuals with gastroparesis than in controls without upper gastrointestinal motility disorders.

Methods: Repeat length was determined in genomic DNA. Controls with diabetes (84 type 1, 84 type 2) and without diabetes (n = 170) were compared to diabetic gastroparetics (99 type 1, 72 type 2) and idiopathic gastroparetics (n = 234). Correlations of repeat lengths with clinical symptom sub-scores on the gastroparesis cardinal symptom index (GCSI) were done. Statistical analyses of short (<29), medium and long (>32) repeat alleles and differences in allele length were used to test for associations with gastroparesis.

Results: The distribution of allele lengths was different between groups (P = 0.016). Allele lengths were longest in type 2 diabetics with gastroparesis (29.18±0.35, mean ± SEM) and longer in gastroparetics compared to non-diabetic controls (28.50±0.14 vs 27.64±0.20 GT repeats/allele, P = 0.0008). Type 2 diabetic controls had longer alleles than non-diabetic controls. In all gastroparetic groups, allele lengths were longer in African Americans compared to other racial groups, differences in the proportion of African Americans in the groups accounted for the differences between gastroparetics and controls. Diabetic gastroparetics with 1 or 2 long alleles had worse GCSI nausea sub-scores (3.30±0.23) as compared to those with 0 long alleles (2.66±0.12), P = 0.022.

Conclusions: Longer poly-GT repeats in the HMOX1 gene are more common in African Americans with gastroparesis. Nausea symptoms are worse in subjects with longer alleles.

Fig 1. Frequency distribution of the lengths of the polyGT repeat allele in the HMOX1 gene.

a) Frequency distribution of the lengths of the polyGT repeat allele in the HMOX1 gene from control subjects with no diabetes and no symptoms of gastroparesis. Dotted lines indicate division chosen between short, medium and long alleles. b) Fractional frequency distribution of the lengths of polyGT repeat in HMOX1 from control, non-diabetic subjects (n = 170) compared to all gastroparetic subjects (n = 576). The distribution of the allele lengths was significantly different between the groups (P = 0.004, Mann Whitney test). c) Distribution of allele lengths in subjects with type 2 diabetes and gastroparesis, d) type 1 diabetes and gastroparesis, e) idiopathic gastroparesis. Allele length was determined by PCR amplification of genomic DNA from blood using the ABI 3730 platform and capillary electrophoresis and using primers flanking the GT repeat region.

Fig 2. Size distribution of the polyGT repeat lengths.

a) The size distribution of the polyGT repeat length in type 2 diabetic controls is significantly longer than the allele length in non-diabetic controls. The fractional frequency distribution of the lengths of polyGT repeat in HMOX1 from non-diabetic control subjects with no GI motility disorders (n = 170) compared to subjects with type 2 diabetes and no GI motility disorders (n = 84) are shown (P < 0.05 Dunn’s test). b) Fractional frequency distribution of the lengths of polyGT repeat in HMOX1 from subjects with type 2 diabetes and no GI motility disorders (n = 84) compared to subjects with type 2 diabetes and gastroparesis (n = 72, P = NS, Dunn’s test). c) Fractional frequency distribution of the lengths of polyGT repeat in HMOX1 from subjects with type 1 diabetes and no GI motility disorders (n = 84) compared to subjects with type 1 diabetes and gastroparesis (n = 99, P = NS, Dunn’s test). Allele length was determined by PCR amplification of genomic DNA from blood using the ABI 3730 platform and capillary electrophoresis and using primers flanking the GT repeat region.

Fig 3. Allele distribution in gastroparetic and control subjects.

Allele distribution of long (L), medium (M) and short (S) alleles a) for subjects with gastroparesis is significantly different from the distribution for control subjects (Chi² P value = 0.019) and b) for subjects with idiopathic gastroparesis compared to non-diabetic controls (Chi² P value = 0.049). Short alleles were defined as shorter than 29 polyGT repeats and long alleles were defined as longer than 32 repeats (see Fig 1a for definition).

Fig 4. Relationship between polyGT allele length and nausea sub-score.

In patients with diabetic gastroparesis, the relationship between polyGT allele length and nausea sub-score on the GCSI questionnaire fit to a slope that was significantly different from zero (P < 0.006 by linear regression, dotted lines are 95% confidence intervals) and b) subjects with one or two long alleles (open circles in Fig 4a) had significantly higher nausea sub-scores than subjects with zero long alleles (closed circles in Fig 4a, P = 0.022, Mann Whitney test). Whiskers are the medians with the interquartile ranges for the nausea sub-score.