Mitochondrial biogenesis, telomere length and cellular senescence in Parkinson’s disease and Lewy body dementia

Progressive age is the single major risk factor for neurodegenerative diseases. Cellular aging markers during Parkinson’s disease (PD) have been implicated in previous studies, however the majority of studies have investigated the association of individual cellular aging hallmarks with PD but not jointly. Here, we have studied the association of PD with three aging hallmarks (telomere attrition, mitochondrial dysfunction, and cellular senescence) in blood and the brain tissue. Our results show that PD patients had 20% lower mitochondrial DNA copies but 26% longer telomeres in blood compared to controls. Moreover, telomere length in blood was positively correlated with medication (Levodopa Equivalent Daily Dose, LEDD) and disease duration. Similar results were found in brain tissue, where patients with Parkinson’s disease (PD), Parkinson’s disease dementia (PDD) and Dementia with Lewy Bodies (DLB) showed (46–95%) depleted mtDNA copies, but (7–9%) longer telomeres compared to controls. In addition, patients had lower mitochondrial biogenesis (PGC-1α and PGC-1β) and higher load of a cellular senescence marker in postmortem prefrontal cortex tissue, with DLB showing the highest effect among the patient groups. Our results suggest that mitochondrial dysfunction (copy number and biogenesis) in blood might be a valuable marker to assess the risk of PD. However, further studies with larger sample size are needed to evaluate these findings.


Discussion
Using blood and prefrontal cortex brain tissues from two different cohorts we show that mitochondrial dysfunction (mtDNA copy number and mitochondria biogenesis gene expression) and cellular senescence, but not telomere shortening is associated with neurodegenerative diseases (PD, PDD and DLB). Our results suggest that mitochondrial dysfunction (copy number and biogenesis) in blood might be a valuable marker to assess the early risk of PD.
A single mitochondrion contains 2-10 copies of mtDNA, depending on the type of cell and tissue 24 . Under healthy circumstances, human cells contain thousands of copies of mtDNA that are usually identical (homoplasmy). However, during infection or disease settings, mtDNA frequently presents a mixture of wild-type mtDNA within each cell 25 , therefore, mutant mitochondrial genome accumulates in cells over time 4 . The number of mtDNA copies increases with age, as a compensatory mechanism, which maintains the amount of wild-type mtDNA and reverses the effect of defective mitochondria accumulation 26 . However, this compensatory mechanism declines in PD resulting in exhaustion of mtDNA copies, which, in turn, leads to respiratory deficiency in dopaminergic neurons 26 . Here, we report a significant reduction of mtDNA copy number in both blood and   [27][28][29] . We found similar mitochondrial reduction (20%) in whole blood in PD patients compared to 19.6% in PBMC previously reported by Pyle et al. 28 . However, we found lower mtDNA copy numbers (46.4%) in prefrontal cortex tissues, while Pyle et al. showed no significant difference of mtDNA copy numbers between PD patients and controls in frontal cortex 28 . Overall, our results are also in agreement with findings from other neurodegenerative diseases including Alzheimer's disease (AD) and Huntington's disease, where mitochondrial dysfunction is observed 28,30,31 . Mitochondrial copy number is strongly associated with mitochondrial function, which makes it an important aging marker 32 . MtDNA mutation and mitochondrial dysfunction, respectively, have been associated with neurodegenerative diseases such as PD and AD 5,33,34 . Our study also shows lower expression of PGC-1β gene (master regulators of mitochondrial biogenesis) in brain tissues of PD, PDD and DLB patients compared to healthy controls (Fig. 4B), However, the expression level of PGC-1α was similar between patients and controls. A recent study by Dölle et al. 2016 also showed no difference in PGC-1α between PD patients and controls 35 . Lack of of PGC-1α correlation could be explained due to the fact that PGC-1α also influences the expression of several other genes involved in metabolic pathways 36 , and therefore its expression might be highly regulated to avoid its deleterious side effects. Our study suggests that both lower mtDNA copy number and expression of PGC-1β in PD, PDD and DLB might indicate dysfunctional mitochondria in patients. An alternative explanation of low mitochondrial content could be due the presence a of high proportion of mutant mitochondrial genome. Our results show that PD patients have longer telomeres in blood compared to healthy controls (Fig. 1B). We found similar results in brain tissues where PD, PDD and DLB patients show longer telomeres compared to healthy controls (Fig. 3B). These results were contrary to our expectations as we expected that patients would show shorter telomeres compared to controls. So far, previous literature has shown mixed results, no association of telomere length with PD in blood 37-41 and brain tissue 42 . However, a study by Maeda et al., 2012, from Japanese women reported shorter blood telomere length in PD patients 43 . Similarly, DLB patients have been shown to have shorter telomeres compared to controls 44 . However, a recent nested case control study showed a positive association between PD and longer telomere length in leukocytes and PBMCs, where men with shorter telomere length were of lower risk of getting diagnosed with PD 45   www.nature.com/scientificreports/ patients who developed dementia within three years after diagnosis had longer telomere length at diagnosis compared to the other PD patients without early development of dementia 37 . Contradictory results of telomere association with PD could be due to the heterogeneity of the study setup (cross sectional vs nested case control), sample heterogeneity, quality and cell composition within each tissue, or differential methods for assessing telomere length. An alternative explanation could be the effect of PD medication on telomere length. Interestingly, our results show that blood telomere length was significantly positively correlated with Levodopa Equivalent Daily Dose (LEDD) medication in PD patients. Furthermore, we found a positive correlation of telomere length with disease duration (years since PD diagnosis) in PD patients, which may further reflect the cumulative effect of LEDD on telomere length. Previously it has been shown that Levodopa moderately induces nerve growth factor and growth hormone 46 . In addition, Levodopa increased homocysteine, which in turn may accelerate aging processes, such as neuropathy and dementia 47 . Our results show that disease duration is positively correlated with Levodopa Equivalent Daily Dose LEDD, which in turn positively correlated with telomere length in blood in PD patients. Our results show that DLB patients have longer telomeres in brain tissue than controls and other patient groups. We did not have Levodopa Equivalent Daily Dose (LEDD) nor other clinical information for the brain tissues as we did for the blood samples. We were, therefore, unable to investigate whether this was an effect of the medication, as generally DLB patients are treated with lower LEDD than PD patients. Our study is also limited in terms of varying cell composition within each brain tissue which might also affect our results. Nevertheless, to elucidate the relationship between neurodegenerative diseases and telomere length, and to pinpoint whether short/long telomeres are the cause or consequence of neurodegenerative diseases, a longitudinal study set-up is needed, with well-defined samples.
Here, we show a significantly higher expression of cyclin dependent kinase inhibitor 2A (CDKN2A) in prefrontal cortex brain tissue of DLB patients compared to healthy controls. CDKN2A reflects the increased load of cellular senescence and has been shown to be negatively associated with telomere length 15,21,22 . A previous study showed that expression of CDKN2A has been associated with mild cognitive decline in aging humans, where CDKN2A expression was positively associated with 3 repeat TAU (microtubule-associated protein) in blood 23 . However, contrary to previous findings we found a positive correlation between telomere length and CDKN2A expression in PD patients. Mechanisms behind such association are yet to be investigated. One strength of the study is that different tissues were compared (blood vs brain) in Parkinson's disease (PD). However, it is difficult to obtain large sample sizes when using brain tissue and therefore results should be taken with cautions.
Our results show that mitochondrial DNA copy number and telomere length are not correlated in blood. However, in a previous study with healthy individuals we showed that these two markers were positively correlated 48 . Telomere-mitochondria axis is argued to compromise metabolism and organ function, where telomere dysfunction triggers P53 and P16, which in turn affect mitochondrial biogenesis 49 . We show a negative correlation of CDKN2A (P16) expression with mitochondrial copy number in brain tissues. Lack of correlation between telomere length and mitochondria in brain tissues could depend upon frequent mtDNA mutation due to disease, which accumulate over the time 25 . Hence our primers only capture wild type mitochondria and not all mitochondrial DNA.
In conclusion, our results indicate that mitochondrial dysfunction and cellular senescence might be valuable markers to study neurodegenerative diseases (PD, PDD, DLB). Follow-up studies were more individuals, particularly healthy controls, are important to perform. The identification of blood biomarkers in neurodegenerative diseases could potentially facilitate the drug development process, as the utility of measuring such markers in the brain is limited. Our findings further extend our knowledge that mitochondrial copy number and function could be a viable biomarker in blood as an early indicator for the risk of neurodegenerative diseases.

Materials and methods
Blood samples from Swedish cohort. The blood samples were obtained from controls and PD patients from the Swedish BIOPARK cohort, following the Declaration of Helsinki and Good Clinical Practice standards and approved by the Swedish Ethical Review Authority, reference number 2019-04967 and all applied methods were carried out in accordance with relevant guidelines and regulations 50 . Patients were recruited in clinics within Stockholm region, Sweden, and from the Sunderby Hospital in Luleå, Sweden. Both verbal and written informed consent were obtained at the time of inclusion. Blood was drawn by venepuncture by trained personnel and collected in EDTA tubes. The mtDNA copy number and the telomere length were measured from DNA of whole blood of n = 112 individuals including 100 PD patients and 12 controls. Age range of patients diagnosed with PD was between 47-97 years and male/female ratio was 1.5, while controls had an age range between 54-73 and male/female ratio was 0.3. Clinical data was collected from PD patients including, disease duration (years since PD diagnosis), Movement Disorder Society Unified Parkinson's Disease Rating Scale part 3 (MDS-UPDRS III) for motor symptoms, Hoehn and Yahr for disease severity, Montgomery-Åsberg Depression Rating Scale (MADRS) for depression, Hospital Anxiety and Depression Scale sub scores for anxiety (HADS-Anxiety) and depression (HADS-Depression), Montreal Cognitive assessment (MoCA) for cognitive assessment, and Levodopa Equivalent Daily Dose (LEDD) as a standard measure for patients' dopaminergic medication.
Brain tissues samples from UK cohort. Postmortem human prefrontal cortex brain tissues were obtained from the MRC London Neurodegenerative Diseases Brain Bank, King's College London, United Kingdom. The permission to collect human brain tissue included participants informed consent for research purposes following the procedure corresponds to principles expressed in the Declaration of Helsinki and Good Clinical Practice standards and approved by the UK National Research Ethics Service (08/H1010/4 and KI IRB) 51 . All applied methods were carried out in accordance with relevant guidelines and regulations. In total 58 brain tissues were A reference genomic human DNA of known telomere length of 4.1 kb (or 369 kb, per diploid cell) and mitochondrial copy number (1200 ± 9 copies) was added on each plate. ΔC T for both telomere length and mitochondrial copy number was calculated using the formula C T target sample-C T reference sample after adjusting the PCR efficiency using Pfaffl method 52 . We then calculated the ∆∆C T for both telomere length and mitochondrial copy number using the formula (TEL∆C T -SCR∆C T ). Relative telomere length of target sample to reference sample was calculated as 2^ − ∆∆C T and the ratio was then multiplied with 369 kb to get telomere length per diploid cell. Telomere length of the diploid cell was divided by number of chromosomes ends (92) to get average telomere length of each chromosome end (2^ − ΔΔC T × 369/92). Mitochondria copy number per diploid cell of target sample to reference sample was calculated as 2^ − ∆∆C T and the ratio was then multiplied with 1200 mtDNA copy number for each sample (2^ − ΔΔC T × 1200), as described elsewhere 48 . Gene expression. 30 mg frozen human brain tissue was used to extract RNA using RNeasy Plus Mini Kit (Qiagen) according to manufacturer's protocol. RNA concentration was measured and evaluated for purity (260/280 nm ratio) using Nanodrop (Marshall Scientific). RNA integrity was confirmed using bioanalyzer. cDNA was synthesized by using QuantiTec Reverse Transcriptase kit (cat# 205,311) following the manufacturer guidelines. Thermal profile consisted of 10 min incubation at 25 °C, followed by 1 h at 42 °C for cDNA synthesis and 5 min at 85 °C to inactivate the enzyme on a Quant Studio 5 thermocycler. Relative gene expression of CDKN2A, PGC1 α and PGC-1β was determined using the comparative ∆C T method by calculating the C T values of the target genes (CDKN2a, PGC1 α and PGC-1β) against the C T values of the reference gene (GAPDH). Target genes and GAPDH were run in triplicates and amplified in the same wells. Respective C T values were averaged before performing the ∆C T calculation (∆C T = C T Target − C T GAPDH ). Gene expression values were converted into log 2 of ∆C T (2^ − ∆C T ). Cellular senescence and mitochondrial biogenesis. CDKN2A  www.nature.com/scientificreports/ Statistical analysis. Statistical analysis was performed using the statistical program JMP (version 16). We performed multivariate regression analysis to investigate the correlation of disease with three hallmarks of aging (telomere attrition, mitochondrial dysfunction, and cellular senescence) in blood and brain tissue separately. Age and sex were fitted as fixed factors in all analysis. For further comparison between different groups, we used multiple comparison test (Tukey's test). Pearson correlation was used to assess the correlation between different cellular aging markers. Fold change of PGC1α, PGC-1β and CDKN2A was calculated by dividing the individual values with the mean value of controls.
Institutional review board statement. The blood samples were obtained from PD patients included in the Swedish BIOPARK cohort (approved by the Swedish Ethical Review Authority, reference number 2019-04967) 50 . Patients were recruited in clinics within Stockholm region, Sweden, and from the Sunderby Hospital in Luleå, Sweden. Both verbal and written consent were obtained at the time of inclusion. Postmortem human prefrontal cortex brain tissues were obtained from the MRC London Neurodegenerative Diseases Brain Bank, King's College London, United Kingdom. The permission to collect human brain tissue included participants consent for research purposes and ethical approval was obtained from the UK National Research Ethics Service (08/H1010/4 and KI IRB) 51 .
Informed consent statement. Both verbal and written informed consent were obtained at the time of inclusion.

Data availability
All data are available upon reasonable request to corresponding author, Muhammad Asghar (asghar.muham-mad@biol.lu.se).