Integrity of Presynaptic Neurons in Long COVID and Relationship to Neuropsychiatric Symptoms
Yuhan Liu, Devina Persaud, Erica Vieira, Joeffre Braga, Pablo Rusjan, Laura Miler, Jennifer Rabin, Tina Mccluskey, Isabelle Boileau, Thomas Chao, Michael Bagby, Lucas Narciso, Lauren Gray, Neil Vasdev, Kim Desmond, Stefan Kloiber, Jerry Warsh, Ishrat Husain, Kelly Smart, Wei Wang and Jeffrey Meyer
Introduction
Long COVID with neuropsychiatric symptoms is common, and effective treatments are lacking. While markers of elevated microglial and astroglial activation are elevated in the brain of long COVID, more prominently in the striatum, it is largely unknown whether injury to subsets of neurons occurs.
Dopamine releasing nerve terminals in striatal regions may be vulnerable to injury from gliosis via excessive synaptic pruning and/or greater generation of reactive oxygen species. Moreover, dopaminergic neurons in striatum may also be vulnerable to direct infection from SARS-CoV-2 because their cell bodies in the ventral tegmental area and substantia nigra express a relatively high density of angiotensin converting enzyme 2 receptors.
Here we measured (+)[11C] DTBZ binding potential (BPND), a well-established marker of dopaminergic nerve terminal integrity in the ventral striatum (VS), dorsal putamen (DP), and dorsal caudate (DC). It was hypothesized that loss of (+)[11C]DTBZ BPND occurs in the ventral striatum, dorsal putamen, and dorsal caudate of Long COVID. Furthermore, reductions of (+)[11C]DTBZ BPND in the VS and DP were hypothesized to be associated with apathy and motor slowing respectively.
An exploratory hypothesis was that memory loss is associated with reductions of (+)[11C]DTBZ BPND in DC. The hypothesized relationships to symptoms are based on evidence from human illness and animal models showing that loss of dopamine neurons in these regions are associated with these particular symptoms, which are also frequent in long COVID.
To additionally investigate the possibility of generalized loss of synaptic terminals in long COVID, we also measured [18F]SynVes-T1 total volume of distribution (VT), another well-established index of general synaptic density, across cortical and subcortical grey matter regions, prioritizing the prefrontal cortex, anterior cingulate cortex, hippocampus, as well as the striatal regions where (+)[11C]DTBZ BPND is being investigated.
It was hypothesized that loss of [18F]SynVes-T1 VT occurs across prefrontal cortex, anterior cingulate cortex, and hippocampus regions. Reductions of [18F]SynVes-T1 VT in the prefrontal cortex and hippocampus were hypothesized to be associated with cognitive symptoms reflective of brain fog in long COVID.
Methods
Twenty-four unmedicated, non-smoking individuals with Long COVID (12 F, 12 M; mean age 32.2) and 43 healthy controls (25 F, 18 M; mean age 29.5) underwent (+)[¹¹C]DTBZ PET scanning. Regional (+)[¹¹C]DTBZ BPND was quantified using a simplified reference tissue model with the cerebellum as the reference region. Group differences were assessed with linear mixed-effects model.
The primary comparison was between 24 Long COVID and 24 age-matched healthy controls. A secondary comparison was between 24 Long COVID cases were compared to the larger set of similarly aged 43 healthy controls and an exploratory comparison between 24 healthy recovered from COVID-19 and 19 healthy who never had COVID-19 (scanned prior to COVID-19 epidemic) was conducted.
ROI-specific group effects were assessed using post hoc pairwise contrasts of estimated marginal means.In addition, correlations between striatal (+)[¹¹C]DTBZ BPND and Long COVID symptom measures of apathy, motor retardation, decline in delayed memory recall and cognitive concerns were evaluated using Pearson’s correlation coefficients.
Ten unmedicated, non-smoking individuals with Long COVID and 23 healthy controls underwent [18F]SynVes-T1 PET scanning. Regional [18F]SynVes-T1 VT was quantified using 1T compartment model with arterial input function. Group differences were assessed with linear mixed-effects model. ROI-specific group effects were assessed using post hoc pairwise contrasts of estimated marginal means. Correlations between [18F]SynVes-T1 and Long COVID symptom were evaluated using Pearson’s correlation coefficients.
Results
For the primary analysis, (+)[¹¹C]DTBZ BPND was reduced across all regions in 24 Long COVID compared to 24 age-matched healthy controls (P = 0.00004).
Patients with Long COVID had lower (+)[¹¹C]DTBZ BPND in the ventral striatum (20%, P = 0.0013), the dorsal putamen (16%, P = 0.00003), and the dorsal caudate (17%, P = 0.00006).
Moreover, (+)[¹¹C]DTBZ BPND was reduced across all regions in 24 Long COVID compared to 43 healthy controls (P = 0.0006).
No significant differences in (+)[¹¹C]DTBZ BPND were found across all regions between 24 individuals recovered from COVID-19 and 19 healthy controls never had COVID-19 (P = 0.77).
In the Long COVID group, lower (+)[11C]DTBZ BPND in the ventral striatum correlated with more severe symptoms on the Marin Apathy Evaluation Scale (r = -0.48, P = 0.018) and the Cognitive Failures Questionnaire (r = -0.52, P = 0.0013). Lower (+)[11C]DTBZ BPND in the dorsal putamen correlated with slower motor speed measured by T-scores on the Finger Tapping Test Dominant Hand (r = 0.51, P = 0.010). Lower (+)[11C]DTBZ BPND in the dorsal caudate correlated with impaired memory retrieval measured by T-scores on the Hopkins Verbal Learning Test-Revised Delayed Recall (r = 0.58, P = 0.0029).
These symptoms that correlated with lower (+)[11C]DTBZ BPND also differed between Long COVID cases and healthy controls who recovered well after COVID-19, as assessed by Welch’s t-tests.
Long COVID was associated with greater symptom severity on the Marin Apathy Evaluation Scale (P < 0.000001), Cognitive Failures Questionnaire (P = 0.00005), Finger Tapping Test Dominant Hand (P = 0.0001) and the Hopkins Verbal Learning Test–Revised Delayed Recall (P = 0.00006).
[18F]SynVes-T1 VT was reduced across all regions in 10 Long COVID compared to 23 healthy controls (P = 0.003).
Additional data is in the process of being analyzed will be included in the poster.
Conclusions
The most likely interpretation is that reduction in (+)[¹¹C]DTBZ BPND reflects a loss of dopaminergic synapses across ventral striatum, dorsal putamen and dorsal caudate. The correlations of (+)[¹¹C]DTBZ BPND with symptoms suggests that loss of dopaminergic nerve terminals may contribute to these symptoms.
This interpretation is supported by previous electrophysiology, optogenetic, viral tracing studies as well as post-mortem studies of neurodegenerative illness which collectively demonstrate that subsets of dopamine releasing neurons in nucleus accumbens participate in motivation; subsets in dorsal putamen participate in motor control; and subsets in caudate participate in memory retention.
Hence, these findings suggest that clinical trials for long COVID should include approaches for restoring dopamine releasing synapses and/or augmenting tonic and phasic dopamine release in ventral striatum, dorsal putamen and dorsal caudate.
A plausible interpretation for the reduction in [18F]SynVesT1 VT is that there is a generalized reduction in presynaptic nerve terminals across PFC and hippocampus, and the striatum in long COVID. Reductions in SV2A binding in brain regions participating in cognition are associated with cognitive decline in most neuropsychiatric illnesses.
Should the reduction in [18F]SynVesT1 VT reflect pathology contributing to cognitive decline, therapeutics to increase general presynaptic density could be tested in long COVID.
Web | Journal of Nuclear Medicine | Abstract
Yuhan Liu, Devina Persaud, Erica Vieira, Joeffre Braga, Pablo Rusjan, Laura Miler, Jennifer Rabin, Tina Mccluskey, Isabelle Boileau, Thomas Chao, Michael Bagby, Lucas Narciso, Lauren Gray, Neil Vasdev, Kim Desmond, Stefan Kloiber, Jerry Warsh, Ishrat Husain, Kelly Smart, Wei Wang and Jeffrey Meyer
Introduction
Long COVID with neuropsychiatric symptoms is common, and effective treatments are lacking. While markers of elevated microglial and astroglial activation are elevated in the brain of long COVID, more prominently in the striatum, it is largely unknown whether injury to subsets of neurons occurs.
Dopamine releasing nerve terminals in striatal regions may be vulnerable to injury from gliosis via excessive synaptic pruning and/or greater generation of reactive oxygen species. Moreover, dopaminergic neurons in striatum may also be vulnerable to direct infection from SARS-CoV-2 because their cell bodies in the ventral tegmental area and substantia nigra express a relatively high density of angiotensin converting enzyme 2 receptors.
Here we measured (+)[11C] DTBZ binding potential (BPND), a well-established marker of dopaminergic nerve terminal integrity in the ventral striatum (VS), dorsal putamen (DP), and dorsal caudate (DC). It was hypothesized that loss of (+)[11C]DTBZ BPND occurs in the ventral striatum, dorsal putamen, and dorsal caudate of Long COVID. Furthermore, reductions of (+)[11C]DTBZ BPND in the VS and DP were hypothesized to be associated with apathy and motor slowing respectively.
An exploratory hypothesis was that memory loss is associated with reductions of (+)[11C]DTBZ BPND in DC. The hypothesized relationships to symptoms are based on evidence from human illness and animal models showing that loss of dopamine neurons in these regions are associated with these particular symptoms, which are also frequent in long COVID.
To additionally investigate the possibility of generalized loss of synaptic terminals in long COVID, we also measured [18F]SynVes-T1 total volume of distribution (VT), another well-established index of general synaptic density, across cortical and subcortical grey matter regions, prioritizing the prefrontal cortex, anterior cingulate cortex, hippocampus, as well as the striatal regions where (+)[11C]DTBZ BPND is being investigated.
It was hypothesized that loss of [18F]SynVes-T1 VT occurs across prefrontal cortex, anterior cingulate cortex, and hippocampus regions. Reductions of [18F]SynVes-T1 VT in the prefrontal cortex and hippocampus were hypothesized to be associated with cognitive symptoms reflective of brain fog in long COVID.
Methods
Twenty-four unmedicated, non-smoking individuals with Long COVID (12 F, 12 M; mean age 32.2) and 43 healthy controls (25 F, 18 M; mean age 29.5) underwent (+)[¹¹C]DTBZ PET scanning. Regional (+)[¹¹C]DTBZ BPND was quantified using a simplified reference tissue model with the cerebellum as the reference region. Group differences were assessed with linear mixed-effects model.
The primary comparison was between 24 Long COVID and 24 age-matched healthy controls. A secondary comparison was between 24 Long COVID cases were compared to the larger set of similarly aged 43 healthy controls and an exploratory comparison between 24 healthy recovered from COVID-19 and 19 healthy who never had COVID-19 (scanned prior to COVID-19 epidemic) was conducted.
ROI-specific group effects were assessed using post hoc pairwise contrasts of estimated marginal means.In addition, correlations between striatal (+)[¹¹C]DTBZ BPND and Long COVID symptom measures of apathy, motor retardation, decline in delayed memory recall and cognitive concerns were evaluated using Pearson’s correlation coefficients.
Ten unmedicated, non-smoking individuals with Long COVID and 23 healthy controls underwent [18F]SynVes-T1 PET scanning. Regional [18F]SynVes-T1 VT was quantified using 1T compartment model with arterial input function. Group differences were assessed with linear mixed-effects model. ROI-specific group effects were assessed using post hoc pairwise contrasts of estimated marginal means. Correlations between [18F]SynVes-T1 and Long COVID symptom were evaluated using Pearson’s correlation coefficients.
Results
For the primary analysis, (+)[¹¹C]DTBZ BPND was reduced across all regions in 24 Long COVID compared to 24 age-matched healthy controls (P = 0.00004).
Patients with Long COVID had lower (+)[¹¹C]DTBZ BPND in the ventral striatum (20%, P = 0.0013), the dorsal putamen (16%, P = 0.00003), and the dorsal caudate (17%, P = 0.00006).
Moreover, (+)[¹¹C]DTBZ BPND was reduced across all regions in 24 Long COVID compared to 43 healthy controls (P = 0.0006).
No significant differences in (+)[¹¹C]DTBZ BPND were found across all regions between 24 individuals recovered from COVID-19 and 19 healthy controls never had COVID-19 (P = 0.77).
In the Long COVID group, lower (+)[11C]DTBZ BPND in the ventral striatum correlated with more severe symptoms on the Marin Apathy Evaluation Scale (r = -0.48, P = 0.018) and the Cognitive Failures Questionnaire (r = -0.52, P = 0.0013). Lower (+)[11C]DTBZ BPND in the dorsal putamen correlated with slower motor speed measured by T-scores on the Finger Tapping Test Dominant Hand (r = 0.51, P = 0.010). Lower (+)[11C]DTBZ BPND in the dorsal caudate correlated with impaired memory retrieval measured by T-scores on the Hopkins Verbal Learning Test-Revised Delayed Recall (r = 0.58, P = 0.0029).
These symptoms that correlated with lower (+)[11C]DTBZ BPND also differed between Long COVID cases and healthy controls who recovered well after COVID-19, as assessed by Welch’s t-tests.
Long COVID was associated with greater symptom severity on the Marin Apathy Evaluation Scale (P < 0.000001), Cognitive Failures Questionnaire (P = 0.00005), Finger Tapping Test Dominant Hand (P = 0.0001) and the Hopkins Verbal Learning Test–Revised Delayed Recall (P = 0.00006).
[18F]SynVes-T1 VT was reduced across all regions in 10 Long COVID compared to 23 healthy controls (P = 0.003).
Additional data is in the process of being analyzed will be included in the poster.
Conclusions
The most likely interpretation is that reduction in (+)[¹¹C]DTBZ BPND reflects a loss of dopaminergic synapses across ventral striatum, dorsal putamen and dorsal caudate. The correlations of (+)[¹¹C]DTBZ BPND with symptoms suggests that loss of dopaminergic nerve terminals may contribute to these symptoms.
This interpretation is supported by previous electrophysiology, optogenetic, viral tracing studies as well as post-mortem studies of neurodegenerative illness which collectively demonstrate that subsets of dopamine releasing neurons in nucleus accumbens participate in motivation; subsets in dorsal putamen participate in motor control; and subsets in caudate participate in memory retention.
Hence, these findings suggest that clinical trials for long COVID should include approaches for restoring dopamine releasing synapses and/or augmenting tonic and phasic dopamine release in ventral striatum, dorsal putamen and dorsal caudate.
A plausible interpretation for the reduction in [18F]SynVesT1 VT is that there is a generalized reduction in presynaptic nerve terminals across PFC and hippocampus, and the striatum in long COVID. Reductions in SV2A binding in brain regions participating in cognition are associated with cognitive decline in most neuropsychiatric illnesses.
Should the reduction in [18F]SynVesT1 VT reflect pathology contributing to cognitive decline, therapeutics to increase general presynaptic density could be tested in long COVID.
Web | Journal of Nuclear Medicine | Abstract
Last edited:
