Article In Brief
Scores on a set of five markers of inflammation predicted the cognitive scores of patients with Parkinson's disease at 36 months. Two measures of cellular senescence likewise predicted cognitive outcomes.
Blood-derived biomarkers of senescence and inflammation look promising as predictors of cognitive decline in Parkinson's disease (PD), according to a January 13 study in the Journal of Parkinson's Disease.
At diagnosis, patients' summary scores on a set of five markers of inflammation predicted their cognitive scores at 36 months. Two measures of cellular senescence likewise predicted cognitive outcomes.
While significant, the results from the Incidence of Cognitive Impairments in Cohorts with Longitudinal Evaluation-Parkinson's Disease (ICICLE-PD) study were not strong enough to be of immediate clinical utility. They add, however, to the existing literature on risk factors observable at diagnosis that should one day enable neurologists to offer patients a reliable prognosis of cognitive outcomes, said the study's senior author and independent commentators.
“The findings look robust,” commented David K. Simon, MD, PhD, professor of neurology at Harvard Medical School and director of the Parkinson's Disease and Movement Disorders Center at Beth Israel Deaconess Medical Center, who was not involved with the study.
“This is a very well organized, well conducted study. But we see a lot of studies about biomarkers like this, and we're not using them in the clinic yet. I'd want to see the results validated in a larger cohort before I use these biomarkers to tell a patient that they're likely to develop dementia.”
Study Details, Findings
The ongoing ICICLE-PD study, based in Newcastle upon Tyne and Gateshead, United Kingdom, includes154 newly diagnosed PD patients and 99 healthy, age-matched controls from the Newcastle area.
The inflammatory markers used in the study were C-reactive protein, TNF alpha, IL-6, IL-10, and IFN-gamma. For markers of senescence, the investigators used p21 and p16—cell-cycle regulators known to suppress tumors—as well as telomere length in blood cells. Found at the end of every chromosome, telomeres serve as protective caps. Because their length has previously been shown to shorten with age, stress and illness, the study sought to determine if a shorter length at baseline would distinguish patients from controls and predict worse cognitive outcomes at 18 and 36 months.
As expected, the study found that PD patients as a group had shorter average telomeres in blood cells at baseline and 18 months compared to age-matched healthy controls. Shorter telomere length among the PD patients was also correlated with an increased risk of PD dementia (PDD) at 36 months.
Results from measurements of p21 and p16 were not as clear. Overall, the best predictor of cognitive score over the 36 months of the study was a summary score of the five markers of inflammation, while the development of dementia best correlated with short telomeres.
“We think these different components might be related,” said the senior author of the paper, Gabriele Saretzki, PhD, a lecturer in aging research at the Biosciences Institute of Newcastle University.
As more patients in the ICICLE-PD cohort progress to dementia, she said, results should strengthen.
“For now we could only go to 36 months,” Dr. Saretzki said. “At 72 months, we should have three to four times more dementia cases, and that will allow us to make much better statistical associations.”
A spate of recent studies has sought to tease out which clinical and biological markers might be predictive of PDD. Last November, another paper in the journal Movement Disorders based on the ICICLE-PD cohort reported that pro-saccades appear to be a useful non-invasive biomarker for long-term PD cognitive decline.
For the new study, Dr. Saretzki and colleagues sought to use ICICLE-PD blood and serum samples in hopes of bringing clarity to contradictory data regarding telomeres. Some prior studies had found no evidence that their length in leukocytes or other cellular chromosomes held any value for predicting PDD. Other studies had found significant effects, although one, published in PLoS One in 2014, found the opposite of the expected direction: longer, rather than shorter, telomeres were associated with increased risk of dementia progression.
“The whole telomere field is very difficult,” said Dr. Saretzki. “I tell students, be very careful when you read papers in this area. Results can depend on not just genetics but on lifestyle factors, psychological or physical stress and inflammation. It's really complex.”
After correcting for multiple comparisons, Dr. Saretzki's group found that PD subjects had “significantly shorter telomere length at baseline compared with controls (p< 0.001)” as well as significantly faster shortening of their telomeres over the first 18 month period (p=0.002).Even so, the telomere lengths of some PD patients overlapped with those of controls. Importantly, however, they also found shorter telomeres at baseline in those PD patients who went on to develop an early dementia.
Even so, because only 11 of the PD patients developed dementia within 36 months, as measured by the Mini-Mental State Examination and Montreal Cognitive Assessment (MoCA), the small numbers “precluded a meaningful regression analysis of the predictive power of telomere length for PDD,” the paper concluded.
Contrary to expectations, PD participants displayed significantly lower levels of p21 gene expression than controls at baseline (p< 0.001), but there was no difference in change with time. For p16 expression, the differences between PD and controls at baseline or the rate of change per month of p16 expression levels did not reach statistical significance. Dr. Saretzki speculated that perhaps the gene expression measures her group conducted on p16 and p21 are less informative than the measures of protein levels previously published.
The baseline composite inflammatory score was only predictive of MoCA score (p=0.037) at 36 months, independently of age, gender, body mass index, and levodopa equivalent doses. Thus, the composite inflammatory score at baseline was best associated with cognitive score, but not with rate of change for any other clinical indicators over the follow-up period. What's more, none of the baseline biomarkers of senescence or inflammation significantly predicted motor function at 36 months or their rate of change.
Connie Marras, MD, PhD, associate professor of neurology at the University of Toronto and a neurologist at the Toronto Western Hospital Movement Disorders Centre, said that the study provides the basis for further investigations. But, she said, “Because they only had 11 people who developed dementia at 36 months, we have to consider these findings quite preliminary.”
Even so, she applauded the effort, given the importance to patients of a reasonably accurate prognosis of their cognitive status.
“Cognitive slowing is a big deal for patients,” she said. “My patients ask what they can expect. It affects their planning for their lives and their careers. We need better ways to predict who will progress faster in terms of cognitive decline.”
Elizabeth Bradshaw, PhD, the Adler assistant professor in neurological sciences at Columbia University's department of neurology, said she was excited to see the results on senescence. In particular, the findings regarding telomeres “look straightforward and quite exciting.”
But, she said, the findings for p16 and p21, as well as for the inflammatory markers, puzzled her.
“The p16 and p21 didn't go in the direction you might expect based on the telomere results,” Dr. Bradshaw said. “But they examined total blood. If they had been able to break it out into specific cell types, like monocytes or memory T cells, perhaps that would explain what was driving the reduction of p21. If it turned out to be in a memory T population, that would be consistent with data suggesting that there is a potentially pathogenic role for memory T cells.”
Dr. Simon agreed that the study's finding that increased risk of dementia was linked to lower, rather than higher, levels of p16 and p21 was the opposite of what was expected.
“The authors acknowledge in the paper that some of the results were paradoxical,” he said. “The bottom line is I wouldn't discount their findings, but I would like to see the results replicated in other studies.”
Dr. Marras has received fees for consulting for Acorda Therapeutics, serving on the advisory board of Denali Therapeutics, an honorarium for teaching from EMD Serono, and research contracts from Grey Matter Technologies.