Brain Glucose Hypometabolism and Iron Accumulation in Different Brain Regions in Alzheimer’s and Parkinson’s Diseases

The aim of this study was to examine the relationship between the presence of glucose hypometabolism (GHM) and brain iron accumulation (BIA), two potential pathological mechanisms in neurodegenerative disease, in different regions of the brain in people with late-onset Alzheimer’s disease (AD) or Parkinson’s disease (PD). Studies that conducted fluorodeoxyglucose positron emission tomography (FDG-PET) to map GHM or quantitative susceptibility mapping—magnetic resonance imaging (QSM–MRI) to map BIA in the brains of patients with AD or PD were reviewed. Regions of the brain where GHM or BIA were reported in each disease were compared. In AD, both GHM and BIA were reported in the hippocampus, temporal, and parietal lobes. GHM alone was reported in the cingulate gyrus, precuneus and occipital lobe. BIA alone was reported in the caudate nucleus, putamen and globus pallidus. In PD, both GHM and BIA were reported in thalamus, globus pallidus, putamen, hippocampus, and temporal and frontal lobes. GHM alone was reported in cingulate gyrus, caudate nucleus, cerebellum, and parietal and occipital lobes. BIA alone was reported in the substantia nigra and red nucleus. GHM and BIA are observed independent of one another in various brain regions in both AD and PD. This suggests that GHM is not always necessary or sufficient to cause BIA and vice versa. Hypothesis-driven FDG-PET and QSM–MRI imaging studies, where both are conducted on individuals with AD or PD, are needed to confirm or disprove the observations presented here about the potential relationship or lack thereof between GHM and BIA in AD and PD.


Introduction
In neurodegenerative diseases such as AD and PD, biomarkers have been recently employed in clinical practice and research to help confirm the disease process, evaluate progression, and indicate potential targets for therapeutic interventions. In AD, amyloid plaques and neurofibrillary tangles characterize the disease from a neuropathological standpoint, although alternative theories exist regarding whether these factors or others, such as insulin signaling abnormalities, account for disease progression. PD is pathologically characterized as an alpha-synucleinopathy, which starts within the olfactory bulbs or myenteric plexus, eventually spreading to the brainstem and subcortical regions over time. Biomarkers may be disease specific such as cerebrospinal beta amyloid and phosphorylated tau in AD or may be non-specific, offering evidence for neurodegeneration in the form of decreased cerebral glucose uptake and metabolism, and the abnormal accumulation of iron. This review focuses on two neurodegenerative biomarkers-glucose hypometabolism (GHM) and excessive brain iron accumulation (BIA) in AD and PD. relationship between GHM and BIA. Studies that assessed brain iron in individuals with AD or PD using quantitative susceptibility mapping-magnetic resonance imaging (QSM-MRI) or GHM using FDG-PET are reviewed herein. The examination of the relationship of the presence of GHM and BIA in brain regions of individuals with AD and PD could reveal important aspects in their association, which may be beneficial in developing not only a better understanding of the pathophysiology of AD and PD, but may also inform future treatments addressing GHM and BIA in these neurodegenerative diseases.

Methods
To describe the anatomical relationship between the distribution of GHM and BIA in different brain regions, published literature on GHM and BIA in AD and PD was identified through PubMed using the search terms glucose hypometabolism, brain iron accumulation, Alzheimer's disease, Parkinson's disease, FDG-PET, and QSM-MRI. Studies that conducted FDG-PET scans to locate regions with elevated GHM in the brains of patients with AD or PD as compared to healthy controls were examined. In addition, studies that conducted QSM-MRI on the brains of patients with AD or PD as compared to healthy control brains were selected and regions where these studies reported elevated BIA were examined.
Studies were removed from consideration for several reasons. Studies that recruited subjects with young/early onset AD or PD were not included since these cases likely involve different causative factors, such as gene mutations [29,30]. Additionally, papers that examined only two or fewer brain regions were not included due to the number of regions of interest being insufficient for a whole brain analysis. Papers published prior to the year 2000 were excluded due to changes in technology and methodology of brain imaging and data analysis. In the case of AD, studies that only compared patients with mild cognitive impairment (MCI) to healthy controls were excluded. In PD, only studies that examined patients with idiopathic Parkinson's disease were reviewed, while studies on individuals with other Parkinsonian diseases, dementia with Lewy Bodies, etc., were not included. Postmortem studies were excluded.
In summarizing the presence of GHM and BIA in brain regions, there were several guiding principles. If two or more papers reported finding either of the two biomarkers in a certain region, the region was recorded and included in the summary. When there was a difference in specificity of brain regions where a particular biomarker was seen, both whole lobes and their more specific parts were included in the table. Because the hippocampus is part of the temporal lobe, it is also included in the temporal lobe column of the tables, even if the cited paper only mentioned the hippocampus. When only one paper mentioned a very specific region of a lobe, for example, the parahippocampal gyrus, the temporal lobe was marked with a "+", but the parahippocampal gyrus was not added as a region to the table.

Results
In AD, 10 studies were reviewed for GHM and 8 studies for BIA. However, not all studies examined all regions. GHM was examined and reported present in cingulate gyrus, temporal lobe, parietal lobe, precuneus, hippocampus, frontal lobe, and occipital lobe (Table 1a). BIA was examined and reported present in the hippocampus, caudate nucleus, putamen, globus pallidus, temporal, and parietal lobes (Table 1b). All 10 of the studies that examined the temporal lobe reported GHM present, while all 5 studies that examined this region for BIA reported it present. All nine studies that examined the cingulate gyrus and the parietal lobe for GHM reported it present, while only one study examined and reported BIA present in cingulate gyrus and two studies examined and reported BIA present in parietal lobe. Six of the seven studies that examined BIA in the caudate nucleus reported it present, while only one of the two studies that examined GHM there reported it present. All six of the studies that examined BIA in the putamen reported it present, while the one study that examined GHM in the putamen reported it absent. Only GHM was reported present in the precuneus and occipital lobe, while only BIA was reported present in the putamen and globus pallidus. GHM and BIA were both reported present in the temporal and parietal lobes. Both biomarkers were also observed in the hippocampus. All three of the studies that examined GHM and all five of the studies that examined BIA reported it present in the hippocampus. Table 1. Glucose Hypometabolism and Brain Iron Accumulation in Alzheimer's Disease. a. Glucose hypometabolism (GHM) in AD: present (+), absent (−), not reported (NR). Papers reporting parieto-temporal location are shown by merging the columns for the parietal and temporal lobes. Under each brain region is the number of studies reporting GHM present over the number of studies that examined this region. In PD, 15 studies were reviewed for GHM, and 15 studies were reviewed for BIA. GHM was reported present in parietal, occipital, frontal, and temporal lobes, the cerebellum, cingulate gyrus, thalamus, caudate nucleus, globus pallidus, putamen, and hippocampus (Table 2a). BIA was reported present in the substantia nigra, red nucleus, globus pallidus, putamen, thalamus, hippocampus, and frontal and temporal lobes ( Table 2b). All of the reviewed papers reported GHM present in the parietal lobe. GHM was also reported present in 12 of the 14 studies that examined the occipital lobe, 11 of the 14 studies that examined the frontal lobe, and 10 of the 11 studies that examined the temporal lobe. Only one study examined and reported BIA present in the parietal lobe, and only two examined and reported it present in the frontal and temporal lobes. The one study that examined the occipital lobe reported it absent. Thirteen of the fifteen studies that examined BIA in the substantia nigra reported it present, and six of the fourteen studies that examined it in the red nucleus reported it present. None of the papers examined or reported GHM in these two regions.

Study
Comparing the relationship of the presence of GHM and BIA between AD and PD reveals that GHM is present in the parietal and temporal lobes in more than 90% of the reviewed studies in both diseases (Table 3). In AD, while all of the studies that examined the cingulate gyrus and precuneus reported the presence of GHM in these regions, only one of the studies examined and reported BIA present in the cingulate gyrus and none examined BIA in the precuneus. In PD, while BIA was reported present in the substantia nigra in 87% of the PD studies, no study examined GHM in this region. The anatomical distribution of GHM and BIA in AD and PD is graphically displayed in Figure 1.   Table 3. Summary of the number of studies reporting GHM or BIA as a fraction of the number of studies examining and reporting on this brain region in AD and PD.

Discussion
This is the first study to explicitly describe the spatial relationship in the brain between GHM and BIA, two biomarkers that are elevated in comparison to healthy controls in both AD and PD, by comparing data from brain FDG-PET and QSM-MRI studies. In AD, GHM and BIA were both reported in the hippocampus, and temporal and parietal lobes. It is likely that regions of the brain where both GHM and BIA occur will experience more damage than those showing only one of the two. For example, the hippocampus, which shows both GHM and BIA, is widely accepted as an important memory region whose functioning is greatly reduced in AD. Further, a 2017 study found that the medial temporal lobe, which includes the hippocampus, shows significant atrophy in AD, adding to the evidence [76]. Neurofibrillary tangles appear earlier in the disease process [77] and have also been observed with tau-PET in this region [78]. Since beta amyloid binds both

Discussion
This is the first study to explicitly describe the spatial relationship in the brain between GHM and BIA, two biomarkers that are elevated in comparison to healthy controls in both AD and PD, by comparing data from brain FDG-PET and QSM-MRI studies. In AD, GHM and BIA were both reported in the hippocampus, and temporal and parietal lobes. It is likely that regions of the brain where both GHM and BIA occur will experience more damage than those showing only one of the two. For example, the hippocampus, which shows both GHM and BIA, is widely accepted as an important memory region whose functioning is greatly reduced in AD. Further, a 2017 study found that the medial temporal lobe, which includes the hippocampus, shows significant atrophy in AD, adding to the evidence [76]. Neurofibrillary tangles appear earlier in the disease process [77] and have also been observed with tau-PET in this region [78]. Since beta amyloid binds both iron [79] and free heme [11], it is not surprising that many of the regions typically associated with AD are those that show both beta amyloid, and consequently, BIA [11,80].
In PD however, one of the most affected motor regions, the substantia nigra, was only reported to show BIA. However other regions such as the globus pallidus, putamen, and thalamus that are also affected in Parkinson's disease were reported to show both GHM and BIA. Further, in PD, the region that was reported by most studies to be affected by GHM was the parietal lobe, a region involved in coordinating movement. However, this region does not accumulate alpha synuclein and Lewy bodies to the degree that the substantia nigra does, and its involvement in PD is not always appreciated. However, connecting this finding to recent literature on the human connectome model of brain function aids in understanding this finding. This model suggests that rather than individual parts of the brain performing specific functions, networks that span across different lobes of the brain are responsible for functions [81]. According to this model, regions such as the parietal lobe would be part of a complex network of neurons associated with movement. Thus, when these regions are starved for energy and show GHM in PD, patients suffer from motor symptoms. This study suggests that although the parietal lobe is not a region that typically contains Lewy bodies, it is also an important region to target in PD [82].
Several of the studies, included in the tables above, selected regions of interest based on regions where previous literature had reported the presence of GHM or BIA. If multiple previous studies did not report significant GHM or BIA in a certain brain region, subsequent studies would be unlikely to report on that brain region unless they observed significant GHM or BIA there. The results suggest that GHM and BIA exist independent of each other in several brain regions, including the putamen, occipital lobe, precuneus, and globus pallidus in AD and substantia nigra, red nucleus, cingulate gyrus, occipital lobe, and cerebellum in PD. However, hypothesis-driven FDG-PET and QSM-MRI imaging studies, where both are conducted in the same brain regions on each group of individuals with AD or PD, will be needed to confirm or disprove the tentative observations presented here about the potential relationship or lack thereof between GHM and BIA in AD and PD.
Our analysis was limited by the number of published studies on GHM and BIA in individuals with AD or PD. Furthermore, the published studies did not always identify the same regions of interest and did not consider the medications that the patients were taking at the time the FDG-PET or QSM-MRI brain imaging studies were performed. For example, subjects with levodopa-induced dyskinesia have more iron in the substantia nigra than those without it [83]. It is also apparent that the results would be more accurate if studies were identified where FDG-PET and QSM-MRI were both conducted on the same patients or on age-and sex-matched pairs. However, no such studies exist.
Brain biomarkers representing potential pathologic mechanisms, such as GHM and BIA, have also been considered as therapeutic targets. Multiple preclinical studies on intranasal deferoxamine, a high-affinity iron chelator, have demonstrated improvements in rodent models of AD and PD [28]. Intranasal deferoxamine has even been reported to improve memory from baseline, increase levels of HIF-1α, and inhibit GSK-3β activity in healthy C57 mice. A much older study reported that intramuscular administration of deferoxamine reduced cognitive decline in AD patients by 50% over a period of two years [84]. Intranasal insulin, a treatment targeting GHM, has been reported to be both safe and beneficial in humans. Multiple human clinical trials have demonstrated that intranasal insulin improves memory in normal healthy adults with no change in the blood levels of insulin or glucose [20,24,26]. In patients with mild cognitive impairment (MCI) or AD, a single dose of intranasal insulin improves memory [25], and intranasal insulin also improves memory with multiple treatments of patients with AD or MCI [20]. A pilot clinical trial on intranasal insulin in individuals with PD also showed promising results with indication of clinically relevant functional improvement including improved verbal fluency and movement, and decreased physical disability [27]. Therefore, it is becoming increasingly more important to learn more about these biomarkers. Since GHM and BIA often occur in different brain regions as shown here, it may be clinically more effective to treat AD and PD by administering a combination of drugs targeting both GHM and BIA.

Conclusions
The results suggest that GHM and BIA are observed independent of one another in various brain regions in both AD and PD. This suggests that GHM is not always necessary or sufficient to cause BIA in a brain region and vice versa. In AD, the brain regions most often reported to show both GHM and BIA in published studies are the temporal lobe and more specifically, the hippocampus, which is key to memory function and demonstrates marked degeneration and volume loss in patients with AD. In PD, while GHM was most often reported in the parietal, occipital, frontal, and temporal lobes, BIA was most often reported in the substantia nigra, red nucleus, and globus pallidus. Hypothesis-driven FDG-PET and QSM-MRI imaging studies, where both are conducted on each group of individuals with AD or PD, are needed to confirm or disprove the tentative observations presented here about the potential relationship or lack thereof between GHM and BIA in AD and PD.