Dopamine transporter binding in the brain is linked to irritable bowel syndrome in Parkinson's disease

Abstract Background Gastrointestinal symptoms are common in Parkinson's disease (PD), but their neurophysiological correlates are not well understood. We recently reported that functional gastrointestinal symptoms were not associated with asymmetry per se but might be associated with lower left striatal dopamine transporter (DAT) binding. The purpose of this study was to further investigate if specific gastrointestinal symptoms associate with monoamine transporter changes in specific striatal or extrastriatal areas. Methods Ninety PD patients, who underwent DAT ¹2 3I‐FP‐CIT SPECT imaging, were assessed using the MDS‐Unified Parkinson's Disease Rating Scale part III, Rome III, and Wexner constipation score. DAT binding was calculated from striatal subregions using region‐to‐occipital cortex ratio. Voxel‐wise analysis was used to assess the relationship between gastrointestinal symptoms and striatal DAT and extrastriatal serotonin transporter (SERT) binding. Results Irritable bowel syndrome (IBS) criteria were fulfilled in 17 patients and were linked to higher ¹2 3I‐FP‐CIT binding in the right posterior putamen and adjacent areas as compared to patients without IBS. No other significant associations between gastrointestinal symptoms and DAT or SERT binding were found. Conclusions These findings suggest that PD patients with IBS may have higher DAT binding in the right hemisphere. This finding implicates alterations of brain neurotransmitter physiology in the gastrointestinal symptoms of PD patients.


INTRODUCTION
Parkinson's disease (PD) is a very common progressive neurodegenerative disease characterized by a variety of motor and nonmotor manifestations. The cardinal motor symptoms bradykinesia and rigidity are linked to the progressive degeneration of nigral dopaminergic neurons (Pirker, 2003). These neurons project their axons to the striatum enabling imaging of presynaptic striatal dopamine transporter (DAT) binding.
Nonmotor symptoms (NMS) may precede motor symptoms by several years (Chaudhuri et al., 2005). These include REM sleep behavior disorder, orthostatic hypotension, depression, pain, fatigue, and a variety of gastrointestinal symptoms (Abbott et al., 2001;Maass & Reichmann, 2013;Pfeiffer, 2018;Poirier et al., 2016). It has been observed that a significant number of PD patients also fulfill the symptomatic criteria of irritable bowel syndrome (IBS) (Mertsalmi et al., 2017). IBS is defined as recurrent abdominal pain or discomfortexperienced at least 3 days a month in the last 3 months before filling the questionnaire (for women, the pain does not occur only during menstrual bleeding)-that is associated with two or more of the following features: (i) improvement with defecation, and onset associated with a change in (ii) frequency or (iii) form of stool (Longstreth et al., 2006). IBS is more complex than just constipation and they are mutually exclusive disorders (Mertsalmi et al., 2017).
The pathological mechanisms behind the NMS are not well understood, although autopsy studies have shown α-synuclein pathology throughout the spinal cord and peripheral autonomic nervous system in PD patients that could play a role (Beach & Adler et al., 2010). Both genetics and environmental triggers, as well as modifiers for PD such as pesticides, toxins, and altered gut microbiota, have been linked to the pathogenesis of NMS (Chen et al., 2022). A recently introduced α-synuclein origin site and connectome (SOC) model suggests that, in a subgroup of patients, the neuropathology initiates in the peripheral autonomic nervous system, most probably in the gut (body-first subtype). This could lead to a more symmetrical central α-synuclein pathology and a higher nonmotor symptom burden when compared to the brain-first pathology (Borghammer, 2021).
DAT imaging with single-photon emission computerized tomography (SPECT) is being used as an in vivo biomarker of nigrostriatal neuron loss in PD, and the overall striatal DAT binding correlates well with the Hoehn and Yahr stage (Pirker, 2003). On the other hand, the relation between NMS, and especially gastrointestinal symptoms (GIS), and DAT binding has been investigated in only few studies with incon-clusive results (Hinkle et al., 2018;Kim & Jun, 2019;Murtomäki et al., 2022;van Deursen et al., 2020).
The connection between GIS and asymmetrical nigrostriatal dopaminergic degeneration (dopaminergic degeneration starting in one hemisphere and remaining more pronounced in that hemisphere) is poorly understood. van Deursen et al. (2020) found that autonomic symptoms, mainly gastrointestinal (GI) and cardiovascular symptoms, were related to lower DAT binding ratios in the right caudate nucleus.
In a recent study, we investigated the connection between the degree of asymmetry of striatal dopaminergic defect and GIS in PD . While we found no significant differences in GIS between symmetric and asymmetric cases, lower left-versus rightsided putaminal DAT uptake, as measured using asymmetry indices (mean putamen asymmetry index calculated as AI = (putamen highestputamen lowest )/(putamen highest + putamen lowest )), was associated with increased reported GIS and constipation. Multiple studies investigated the association of IBS and PD, and in a systematic review and meta-analysis by Lu et al. (2022), the overall risk for PD in IBS patients was significantly higher as compared to the general population (odds ratio: 1.5). Therefore, intestinal dysfunction may be a prodromal PD symptom (Lu et al., 2022).
In addition to high affinity to the presynaptic DAT, the ź 23 I-FP-CIT ligand has a modest affinity for the presynaptic serotonin transporter (SERT) (Abi-Dargham et al., 1996;Joling et al., 2019). Therefore, it is possible to simultaneously use ź 23 I-FP-CIT as a proxy for the integrity of both the striatal dopaminergic and the extrastriatal serotonergic system in vivo (Ziebell et al., 2010).
In this analysis, we aimed to get further insight into the relationship between asymmetric DAT binding and GI symptoms and extended the analysis to extrastriatal regions. To this end, we here investigated striatal and extrastriatal ź 23 I-FP-CIT binding in relation to GIS in the same PD population using voxel-wise analysis .

METHODS
This cross-sectional clinical and imaging study consisted of patients imaged with 123 I-FP-CIT SPECT because of parkinsonism or tremor referred to imaging by their neurologist

Clinical features
All patients were clinically examined before imaging. The clinical examinations included an interview, MDS-Unified Parkinson's Disease Rating Scale part III (MDS-UPDRS part III) (Goetz et al., 2008) in ON state, and MMSE (Folstein et al., 1975). All investigators had completed the MDS-UPDRS Training Program and Exercise. For the parkinsonian motor features, the MDS-UPDRS part III total score was calculated.
A permission to use MDS-UPDRS part III was asked and granted by Movement Disorder Society.
GIS were evaluated with Wexner constipation score (Agachan et al., 1996) and Rome III questionnaire (Drossman & Dumitrascu, 2006). If there were more than 20% of answers missing, the value was labeled as missing. If there were less than 20% of answers missing, the values were corrected for the missing ones (value × questions/(questionsunanswered questions)) (Joutsa et al., 2012) for Rome III questionnaire and Wexner score as in our previous article   (Albert et al., 2016). Specific ź 23 I-FP-CIT binding was calculated from striatal subregions (caudate, anterior putamen, posterior putamen) using region-to-occipital cortex ratio, as in our earlier studies (Kaasinen et al., 2014). The image quality was inspected visually by the investigators. To mitigate the potential confounding effects of different scanners, the SPECT device was calibrated using instructions for camera corrections from the ENC-DAT database (Varrone et al., 2013).
The classification of patients to normal and abnormal ź 23 I-FP-CIT binding groups was based on the automated semiquantitative BRASS-analysis (Hermes Medical Solutions AB, Stockholm, Sweden).
The limit for abnormal area was more than 2 standard deviations below the reference mean in any of the six analyzed regions.
Images with borderline results were re-evaluated by a movement disorder specialist with extensive experience in brain dopamine imaging (V.K.). CIT binding between DAT and SERT, and each GIS voxel-by-voxel with age, sex, and UPDRS part III total score as covariates. These covariates were chosen because they have been shown to be associated with ź 23 I-FP-CIT binding (Honkanen et al., 2021;Pirker, 2003). The ratio images were smoothed using 8 mm at full-width-half-maximum Gaussian kernel to improve the signal-to-noise ratio. An analysis mask including the frontal lobes, cingulate cortex, basal ganglia, thalamus, media temporal lobe, and midbrain was created using WFU pickatlas (version 3.05) . Cluster-level family-wise error (FWE) correction was used to test for significance (with p < .005 uncorrected and after .05 FWE cluster-level corrected p-values). For GIS showing significant association with ź 23 I-FP-CIT binding in the striatum, we created unthresholded striatal maps to further visualize the exact regions within the striatum using unthresholded t-maps.

GIS in PD patients
Basic cohort characteristics are summarized in Tables 1 and 2

SPECT imaging and GIS
When comparing ź 23 I-FP-CIT binding with voxel-based SPM between IBS+ and IBS-patients, we found a cluster of higher binding in IBS patients mostly in the right posterior putamen, extending to the globus pallidus, internal capsule, claustrum, insula, and adjacent white matter ( Figure 1). In addition, Wexner constipation scores were significantly associated with higher ź 23 I-FP-CIT binding in a widespread area bilaterally in the dorsal frontal cortex (cluster size: 15,637 voxels, peak at Montreal Neurological Institute coordinates: −6, −22, 74, p FWE < .001); however, the average binding in this cluster did not exceed the uptake in the reference regions, indicating this finding was not driven by true differences in specific monoamine transporter binding. There were no significant associations for other symptom scores.
We also analyzed the relationship between IBS and SBRs using ROIs in the right and left posterior putamen, right and left anterior putamen, and right and left caudate nucleus. There were no clinically significant associations in any of the ROIs (

DISCUSSION
This study was prompted by our previous finding that lower left-ver- There was a trend association between higher geriatric depression scores and lower pallidal DAT binding. Significant associations between autonomic scale GI subscore (SCOPA-AUT) and lower uptake in mean putamen and caudate nucleus were also found. This was in line with the results from a previous study by Hinkle et al. (2018). Also, SCOPA-AUT urinary subscore was negatively correlated with striatal uptake as in a previous study by Kim and Jun (2019). However, they did not report any lateralization or higher binding in contrast to our results. In a 18 F-Dopa PET imaging study, Ito and colleagues showed reductions in both striatal and nigral brain dopaminergic functions in early PD (Ito et al., 1999). In another study with 18 F-FP-CIT PET, there was significantly decreased 18 F-FP-CIT binding in the contralateral putamen in early-stage PD patients compared to controls (Ma et al., 2002). To our knowledge, PET imaging with voxel-wise analyses has not been used to study GIS and PD.
Previously, the lateralization of dopaminergic degeneration and GIS in PD has been studied in only few studies. Deursen and colleagues observed that autonomic symptoms, mainly cardiovascular and SCOPA-AUT GI subscores, were associated with lower ź 23 I-FP-CIT binding ratios in the right caudate nucleus (van Deursen et al., 2020).
A negative association was shown between DAT binding and SCOPA-AUT GI subscores in the right caudate nucleus. This is in contrast to our result of higher binding in the right posterior putamen linked to IBS. In Deursen's study, the seven-question SCOPA-AUT GI subdomain, with questions on swallowing, drooling, and early fullness, was used and only three questions focused on bowel symptoms (Visser et al., 2004), whereas we used the Rome III questionnaire to evaluate different FGIDs. The use of these different questionnaires might explain some of the differences between the results. In a Swedish study by Liu et al., IBS diagnosis was associated with a 44% higher risk of PD (Liu et al., 2021). In another article by Mertsalmi et al., the association between IBS and PD was confirmed but may have been explained by reverse causation and detection bias (Mertsalmi et al., 2021). Mishima et al. found the prevalence of IBS in PD to be 17.0%, which is a higher prevalence compared with the general population (Mertsalmi et al., 2017;Mishima et al., 2017). Also, in a study by Lai et al. (2014), the adjusted hazard ratio of PD associated with IBS was 1.48, compared with the non-IBS group. So, IBS seems to be more common in PD patients. The high prevalence of IBS symptoms in PD and the suggestion that IBS may be a risk factor for PD warrant further investigation of the underlying mechanisms (Mertsalmi et al., 2017;. It is unclear whether the increased DAT signal in the right posterior putamen reflects a lesser degree of neurodegeneration or upregulation of DAT in PD patients reporting IBS symptoms. To our knowledge, no previous studies have used DAT imaging in IBS patients. In MRI studies of IBS patients, the most consistent structural and functional alterations are found in the somatosensory network, including the globus pallidus, putamen, and caudate nucleus (Hillestad et al., 2022).
A diffusion tensor imaging (DTI) study on IBS patients found white matter damage in the splenium of the corpus callosum, the right retrolenticular area of the internal capsule, and the right superior corona F I G U R E 1 A. Brain regions showing significantly higher 123 I-FP-CIT binding in patients with IBS compared to patients without IBS (cluster size 2327 voxels, peak at MNI-coordinate 24 −10 22, P uncorr <0.005, PFWE=0.008). B. Unthresholded striatal t map of the contrast patients with IBS > patients without IBS demonstrating higher binding in the right sensorimotor striatum. Note that panel B is intended to visualize the group differences in the whole striatum and not for statistical testing of any hypotheses. C-D. PD patients with IBS (C) and without IBS (D) compared to healthy controls to illustrate the overall 123 I-FP-CIT binding abnormality in these patients (PFWE<0.05 for all clusters). Left side in the figure = right side of the brain.
radiata (Fang et al., 2017). In another DTI study, decreased fractional anisotropy in the right dorsal cingulum bundle in adolescent female patients with IBS was found when compared to controls (Hubbard et al., 2018). IBS patients with alexithymia showed also altered brain responses to rectal distension in the right insula (Kano et al., 2020).
Further, increased SERT binding in the midbrain and thalamus has been associated with functional dyspepsia (Tominaga et al., 2015). These results suggest that IBS is associated with central nervous system changes, and the basal ganglia and lateralization might have an effect on IBS and other FGIDs. On the other hand, elevated SERT has been also found in PD patients with depression and it was speculated to be caused by upregulation, so upregulation might be one explanation (Boileau et al., 2008). Our finding has not been reported before in studies on PD and NMS and warrants more investigation on that topic.
A strength of our study is that the diagnosis of PD was based on ź 23 I-FP-CIT SPECT and the opinion of a movement disorder specialist. The patients were being followed up to 4 years to diminish the possibility of false PD diagnosis, and the mean follow-up time at the time of our analyses was 30 months (2-57 months). The patients were sent to imaging for diagnostic purposes and most of them had earlystage symptoms without long use of dopaminergic medication. On the other hand, selection bias cannot be excluded, because patients were referred to ź 23 I-FP-CIT SPECT imaging, which may reflect some initial diagnostic uncertainty.
While we used multiple GIS questionnaires, the lack of objective measurements of GI functions is a slight limitation. However, it has been shown that PD patients and non-PD patients with functional constipation have severe functional dysmotility of the colon and rectum based on objective GI investigations , supporting the use of the Rome questionnaire.
It should also be noted that, unlike the voxel-wise analyses, our ROI analyses failed to identify significant associations between IBS and ź 23 I-FP-CIT binding. However, there was a trend-level finding in the right posterior putamen, supporting our voxel-wise findings. The ROI analysis averages voxels from an anatomical region with the assumption that the entire anatomical region has the same function, which often is not a valid assumption. In fact, the cluster associated with IBS covered only the ventral part of the posterior putamen extending to other anatomical regions, suggesting that there may be specific subregions in the putamen and other brain regions driving this effect.
It should be noted that since SPECT generally is less sensitive and has poorer spatial resolution compared to PET, investigating IBS in PD using PET imaging could possibly be useful to further improve the spatial specificity of our findings.
Patient-reported GIS reflect the patients' subjective experience of symptoms, which is clinically most relevant and likely results from a combination of underlying mechanisms. The etiologies behind the GIS are very complex, and altered bacterial flora (Mertsalmi et al., 2017), small intestinal bacterial overgrowth, increased intestinal permeability as well as abnormal gut motility might be etiologic factors (Mulak & Bonaz, 2015). Furthermore, there is evidence for connections between mental health and gut health. A study on anxiety, depression, and GIS in PD suggested a possible cyclical relationship, where GIS increase an individual's risk for worsening mood symptoms, which in turn increase their risk for worsening GIS (Jones et al., 2021).
Taken together, the patient's symptom experience is a combination of multiple mechanisms.
PD patients tend to have a larger number of different NMS compared with normal controls, and NMS tend to be more frequent and more severe (Kim et al., 2013;Krishnan et al., 2011). The NMS have a large effect on the quality of life for the patient and caregiver and also on institutionalization rates (Chaudhuri et al., 2006;Pfeiffer, 2016). It is very important to investigate the NMS' etiologies to develop novel therapeutic strategies. Our study suggests a link between the lateralization of DAT availability and GIS in PD and elucidates the role of the dopaminergic system in the gut-brain axis in PD patients. funding acquisition (lead); investigation (supporting); methodology (supporting); project administration (supporting); supervision (equal); validation (equal); writing-review and editing (lead).

ACKNOWLEDGMENTS
We thank Antti Loimaala and Sorjo Mätzke as well as the teams of the nuclear medicine departments in Meilahti and Jorvi hospitals for all support. We thank all the participants in the study. We also thank research nurses Kari Lindholm and Tanja Nojonen for their work with the study.

DATA AVAILABILITY STATEMENT
Data available on request due to privacy/ethical restrictions.