GI symptoms such as constipation and bloating are the most common non-motor symptoms in PD patients 
. It is generally believed that these symptoms are a consequence of PD and are the result of intestinal motility disorders and the associated intestinal bacterial overgrowth. However, over the last decade there has been mounting evidence that supports a role for the GI tract and the enteric nervous system (ENS) in the pathogenesis of PD 
. Two early studies found the first GI Lewy bodies (LBs) in the esophagus and colon in PD patients 
. Later studies by a third group substantiated these early findings of LB pathology in both the submucosal plexis (SMP) and myenteric plexis (MP) in the ENS of PD patients 
. Recently, Lewy bodies were found in colonic biopsies from 4 out of 5 
and 21 out of 29 
PD patients, but in none of the controls. It is notable that all of these prior studies used later stage patients. The similarity of our findings in early untreated patients suggests that this pathology is unrelated to drug treatment and supports the concept for GI-related parameters to serve as a biomarker for the disease.
None of our subjects complained of constipation and the frequency of bowel movement in all subjects were within normal range. These data are now included in . In fact, we designed the study as such to minimize the potential confounding effects of constipation due to longstanding Parkinson's disease and included only newly diagnosed PD patients. Thus, it is unlikely that constipation played a role in any of our findings. Although we did not directly study small bowel bacterial overgrowth, none of the patients complained of bloating and there is no known physiological reason for small bowel bacterial overgrowth to cause disruption of colonic permeability. Of course changes in colonic microbiota could play a critical role in changes in colonic permeability in PD patients because there is compelling evidence of cross talk occurring between the colonic microbiota and intestinal epithelial cells. Indeed, part of our broad hypothesis is that changes in colonic microbiota could play a mechanistic role in our observed intestinal hyperpermeability and further studies are needed to interrogate intestinal microbiota in patients with PD.
While compelling, these observations do not necessarily suggest a causal role for the intestines in the pathogenesis of PD and could simply indicate widespread neuronal damage resulting from PD. However, these findings by others, as well as by Braak and colleagues, led to the proposal 
that the ENS may be a route by which a toxin or pathogen initiates the progression of PD over a period of many years. In support, a pathway of α-synuclein expressing neurons from the gut to the CNS was recently shown 
. A recent publication from our group has shown α-synuclein staining positive sigmoid mucosal neuronal tissue in all 9 of the early untreated PD patients examined in this present study
, while none of the control samples exhibited substantial staining indicating these processes occur early on in the pathogenesis of PD. None the less, these observations still do not prove that PD pathology begins in the ENS and they may simply represent progression of PD pathology that could explain the high frequency of GI symptoms in PD patients 
. Further longitudinal studies are needed to determine whether α-synuclein staining in the intestinal mucosa is present in PD patients years prior to onset of clinical neurological evidence of PD.
Regardless of whether intestinal α-synuclein aggregates are the primary or secondary events in PD, in order to better understand its role in PD, we need to determine the mechanism of their development. It is now widely believed that α-synuclein aggregates are the consequence of oxidative injury to neurons 
. It is also well established that inflammation and pro-inflammatory cytokines like IL-1β, IL-6 and TNF-α are one major source of tissue oxidative stress 
. Indeed, neuroinflammation is present within the brain of early PD patients 
and considerable evidence supports a role for neuroinflammation in PD pathogenesis 
Thus, it is a reasonable and plausible hypothesis that ENS α-synuclein aggregates are the consequence of a local pro-inflammatory milieu. Braak and colleagues propose that toxin and/or pathogen presence in the intestinal lumen may trigger neuroinflammation. However, it should be noted that the presence of toxins and/or pathogens might not necessarily supply the required trigger because even healthy subjects have a large amount of luminal proinflammatory products resulting from normal intestinal microbiota including endotoxins. It is more compelling to consider that increased intestinal permeability (leaky gut) and translocation of these substances is the critical factor for intestinal neuronal oxidative injury. Several lines of evidence strongly suggest that endotoxins are plausible neuroinflammatory triggers in PD. First, endotoxin (LPS) has been suggested as one environmental trigger of PD 
. Second, LPS administration either systemically or directly into the CNS is a widely used animal model of PD. For example, mice given a single dose of LPS systemically 
or prenatally 
develop CNS inflammation including selective loss of DA neurons in the substantia nigra (SN) 
. In the present study we showed evidence of excessive endotoxin exposure in PD patients because serum LPS binding proteins were significantly lower in PD patients compared to controls. Generally, lower levels of LBP have been shown to support increased proinflammatory signaling and cytokine production by endotoxin 
. To the best of our knowledge, our finding is the first to show increased bacterial/endotoxin exposures in PD. It should be noted that we did not find significant increase in serum endotoxin in PD patients. We believe that our findings of significant changes in LBP and positive staining of E.coli
antigen in the colonic biopsies strongly suggest that our observed increased intestinal permeability in PD has significant biological consequence resulting in exposure of intestinal mucosa and the mucosal immune system to the luminal bacteria products. However, the translocation of luminal bacteria products could be limited to the colonic submucosa leading to oxidative injury to the colonic mucosa neurons and aggregation of α-synuclein. Another possibility is that leaked endotoxin could reach the portal systemic circulation but is cleared by the liver and thus a high level of endotoxin was not observed in our patients. Regardless, even if leaked endotoxin cannot reach the circulation, it still can play a central role in production of α-synuclein aggregates in the colon and pathogenesis of PD. Endotoxins (or gram negative bacteria such as E. coli
) can bind to and thus be trapped in the intestinal mucosa and not reach the systemic circulation. Positive E. coli
staining in the sigmoid mucosa of our PD patients not only further supports the conclusion of increased permeation of endotoxins across intestinal epithelial cells, it also supports the possibility that endotoxins become trapped in the intestinal mucosa. It should also be noted that endotoxins might be directly involved in initiation and/or progression of neuroinflammatory injury in both the ENS and CNS in PD without systemic endotoxemia 
. Further studies are needed to evaluate whether there is cell loss in the enteric nervous system associated with α-synuclein aggregates in our PD subjects
How does endotoxin promote neuroinflammation? One possibility is that endotoxins stimulate the enteric immune response either directly or via glial cells to promote local oxidative stress leading to α-synuclein misfolding, aggregation, and subsequent neuronal damage in the ENS in individuals genetically susceptible for PD. Indeed, we have shown an increased nitrotyrosine staining indicating increased oxidative stress in the sigmoid mucosa of PD patients and this increase correlated with plasma LBP and intensity of intestinal E. coli
staining. Our finding that serum sCD14 levels in PD patients was not significantly increased suggests that macrophages are not involved in endotoxin-induced neuroinflammation because activation of macrophages by endotoxin through ligation with TLR4 results in shedding of sCD14 and increased levels of sCD14 in the blood 
. However, it is possible that shedding of sCD14 in the intestinal mucosa is not reflected in the serum measurement. It should also be noted that endotoxins can trigger immune-inflammatory pathways via TLR4 ligation without requirement of CD14 in non-macrophage immune cells 
If indeed endotoxins cause neuroinflammation then why does this occur only in individuals susceptible for PD? First, neuronal tissues in those susceptible to PD may be more sensitive to the endotoxin effects. There are several experimental findings to support this possibility. Rodent nigral neurons have been shown to be more vulnerable than either hippocampal or cortical neurons to LPS-induced degeneration in vivo
and in vitro
, and this was shown to be due to the higher number of activated microglia per unit area in the SNpc compared with the other brain regions 
. Inherent processing of α-synuclein may also play a role. Transgenic mice overexpressing human A53T mutant α-synuclein were given LPS systemically (i.e., i.p. injection). Only the transgenic mice, and not controls, developed sustained neuroinflammation including Lewy bodies staining of nigral neurons 
. This supports a synergistic interaction between defective α-synuclein processing and endotoxin exposure in promoting PD-related nerve pathology.
Second, endotoxin-mediated neuronal damage might not be limited to PD. For example, normal human rectal tissue biopsies incubated with LPS exhibit increased enteric glial cell oxidative stress burden (i.e., iNOS production) 
, a possible source for the observed nitrotyrosine staining in our subjects. Furthermore, co-culture of rat enteric neurons and enteroglial cells stimulated with LPS results in increased glial cell production of proinflammatory cytokine IL-1β 
. It is therefore not surprising that endotoxemia has been reported in other neurological disorders like autism, ALS, and depression 
,  
Another question that needs to be answered is what is the cause of increased endotoxin exposure in PD? One obvious mechanism is increased intestinal permeation to endotoxin (i.e., gut leakiness). We show here that PD patients have a “leaky gut”. PD patients not only had marked and significant increased urinary sucralose, they also had increased intensity of E. coli
staining in the deep zone of the sigmoid mucosa. Thus, PD patients appear to have a hyperpermeable intestinal epithelium and this may have the important biological consequence of a significantly increased exposure to endotoxins, oxidative stress, and neuronal injury to the ENS. The increase in urinary sucralose with no significant change in urinary lactulose or L/M ratio in PD patients strongly suggests that the site of enhanced leakiness is the colon. To our knowledge, intestinal permeability has previously been evaluated in PD subjects in only one other study 
. In that study they found PD patients exhibited decreased absorption of mannitol resulting in an increased L/M ratio. However sucralose absorption was not tested as we have shown in our study and very significantly the median age of those patients was 73.9 y/o (vs. 57 y/o in our study) and all had responded to L-dopa therapy (vs. no therapy in our study). Also, no intestinal biopsy staining for α-synuclein or markers of endotoxin was carried out. Thus we are the first to demonstrate increased intestinal permeability and its significant association with increased intestinal exposure to endotoxin, intestinal oxidative stress, and intestinal α-synuclein aggregation in early, untreated PD patients.
In summary, our study presents data for a new model for PD pathogenesis involving increased intestinal permeability to proinflammatory bacteria and bacterial products such as endotoxins (i.e., LPS). This model provides for an easily obtained, inexpensive biomarker that may be useful in identifying those individuals with premotor PD that will get classic PD signs and symptoms later in life. Our data demonstrate that increased intestinal permeability strongly correlates with markers of increased exposure to endotoxin and with a marker indicating increased oxidative stress burden in the intestine, which together may be responsible for the abnormal accumulation of α-synuclein in enteric neurons we also observe as have other studies. Such increased bacterial endotoxin may therefore initiate a cascade of proinflammatory events promoting PD in genetically susceptible individuals. We acknowledge that our cross sectional observational study cannot establish a causal relationship between increased permeability and PD pathophysiology and that further interventional and animal studies are needed to determine whether gut leakiness and α-synuclein deposition in the colon have a causal role in the pathogenesis of PD. These studies will be needed to determine whether increased intestinal permeability plays a causative role in PD initiation or is a consequence of PD pathogenesis. At the least, such a mechanism for increased exposure of the ENS to intestinal endotoxin may fuel the progression of PD in susceptible subjects.