Turning on the data tap at GP2

Posted on by

A new paper in the journal JAMA Neurology presents a progress report from the Global Parkinson Genetics Program, or “GP2.” That’s a world-wide collaboration to study the genetics not only Parkinson’s disease, but also dementia with Lewy bodies, multiple system atrophy, corticobasal degeneration — and PSP.  Started in 2020, GP2 eventually plans to enroll 250,000 participants at 415 academic centers in 70 countries.  Its funding is mostly from the Sergey Brin Family Foundation to Aligning Science Across Parkinson’s (ASAP), and the complicated nuts and bolts of the project are organized by the Michael J. Fox Foundation. The paper is entitled “Pathology and Genetics in a Global Cohort of Parkinsonian Disorders,” with first author Dr. Lesley Wu and senior author Dr. Huw Morris, both of UCL Queen Square Institute of Neurology in London.

Our knowledge of the genetics of PSP, PD and the other Parkinsonian disorders remains insufficient, and the little we do know is based almost entirely on studies of white, European-derived populations.  The goal of GP2 is to extend our knowledge of the genetics of the Parkinsonian disorders to include non-European-derived populations as well.  Why?  Two reasons:

  • The most obvious justification is to allow all to benefit from tests and treatments arising from genetic knowledge. 
  • Another is to Identify genetic variants associated with the disease in previously under-studied gene pools.  That could identify previously unsuspected brain chemical processes that could in turn point to new treatment targets — not only for members of those ethnic groups or with those gene variants, but for the diseases in general, world-wide.

Examples of the groundbreaking success of this approach in neurodegenerative diseases have been a family in rural Venezuela with a mutation in the huntingtin gene causing Huntington’s disease, and one in rural southern Italy with a mutation in the alpha-synuclein gene causing a form of Parkinson’s disease.  In each case, the gene was not previously known to have any relationship to its respective disease.  In each, the genetic insight has resulted in new scientific understanding, diagnostic tests and experimental neuroprotective treatment for the disease in anyone.  Both examples, however, are based on dominant-acting mutations manifesting mostly in the young or middle-aged, where a familial pattern is obvious.  But in the vast majority of neurodegenerative diseases, any genetic component is more subtle, with each genetic “risk variant” contributing only a smidge to the overall disease likelihood, and familial patterns are often undetectable by laypersons or ordinary medical testing.

One of the first tasks for the GP2 study is to use autopsies to make sure that the people whose DNA samples have been obtained really do have the disease their neurologist has diagnosed. 

The analysis included 3,403 autopsies. Of those individuals:

• 1,171 had been diagnosed during life with PD
• 399 with Parkinson’s disease dementia (PDD)
• 227 with dementia with Lewy bodies (DLB)
• 491 with PSP
• 244 with MSA
• 76 with CBS

Here’s a table comparing the pre-mortem (or “clinical”) diagnoses with autopsy results:  The columns are the clinical diagnoses and the rows are those made at autopsy.  So, for example, of those with corticobasal syndrome diagnosed during life, 16% turned out to have Alzheimer’s disease at autopsy.

Autopsy diagnosisPercentage with Clinical Diagnosis
PDPDDDLBPSPMSACBSControls
Lewy body disease90%96%94%7%17%8%4%
Alzheimer’s1%1%3%0%0%16%1%
PSP3%2%1%88%7%36%0%
MSA3%0%0%2%75%5%0%
CBD0%0%0%1%0%24%0%
Other neuro-degen.2%1%0%2%1%12%2%
No neuro-degen.0%0%1%0%0%0%93%

Footnotes for this table:

  • “Lewy body disease” is the autopsy picture underlying the clinical spectrum that includes Parkinson’s disease, Parkinson’s disease dementia and dementia with Lewy bodies.  Those three conditions are now widely considered to be sub-types of the same disease, just as PSP-Richardson syndrome and PSP-Parkinsonism are sub-types of PSP.
  • For visual simplicity, I’ve rounded all percentages to the nearest integer, so some of the “0%” figures are actually 0.1% to 0.4%.
  • “Other neurodegen.” comprises aging-related tau astrogliopathy, argyrophilic grain disease, chronic traumatic encephalopathy, primary age-related tauopathy, Pick disease, tauopathy not otherwise specified, and vascular pathology.

Take-homes from this table:

  • A clinical diagnosis of PSP is highly accurate, but far from optimal, with 88% confirmed at autopsy.  The next most common autopsy result in such individuals is Lewy body disease, at 7%.
  • Of those with a clinical diagnosis of CBS, only 24% proved to have CBD at autopsy.  A more typical figure from the literature is 40-50%. 
  • Of those with a clinical diagnosis of CBS, 36% proved to have PSP at autopsy.  A more typical figure from the literature is 25%.

The main point of the project, however, is the genetics.  The 20 genes analyzed in this paper were selected because of their known association with the Lewy body diseases.  However, six of them did appear in at least one of the 531 people with PSP who underwent genetic analysis.  They are:

  • The MAPT gene, which encodes the tau protein.  The H1/H1 haplotype (where each of the two copies of chromosome 17 carries the variant) occurred in 83% of people with PSP and in 48-72% of the other groups.  This was first discovered in 1998, so no news there.
  • The LRRK2 (“lark-two”) gene, which encodes an enzyme involved in breakdown of abnormal tau by the lysosomes, showed a variant in 2 people (0.4%) with PSP and in 0.8% with the Lewy body diseases.  This is also no different from previously known statistics.  The lysosomes are one of the cell’s most important mechanisms for disposing of defective, worn-out or excessive proteins such as tau.
  • 29 (16%) of the people with PSP carried one of the known disease-associated variants in the GBA1 gene.  That encodes the enzyme glucocerebrosidase, which like LRRK2, relates to the lysosomes.  This percentage was less than for LBD but did not differ from the other diseases.  Also no surprise, based on previous research.
  • Two results related to ethnicity: 
    • Ashkenazi Jews, regardless of diagnosis, were more likely than all other groups to carry variants in the GBA1 gene.  This has been known since the original description of a GBA1-Parkinson’s relationship in 1996. 
    • South Asians, regardless of genetic results, were more likely to have PSP than other ethnic groups.  This has been informally suspected, but as far as I know, this paper provides the first (admittedly meagre) actual statistics. A deeper dive into the association must now be done:  For example, are neurologists in South Asia better informed when it comes to PSP than neurologists elsewhere?  In the US, where South Asians are a relatively prosperous community with a disproportionate number of physicians, does that group tend have better access to PSP expertise than other ethnicities?  When I know, you’ll know.

  Overall take-homes:

  • This genetic analysis included only 20 genes, of which only two have been reported elsewhere to be related to PSP.  The GP2 study will eventually perform whole-genome sequencing in search of gene variations not previously known to associate with PSP and the other disorders.  To date, the number of DNA samples from under-studied populations remains too small for any results to be reported.
  • The comparison of clinical and autopsy diagnoses shows that while the positive predictive value of a clinical diagnosis of PSP is very good, it could be better.  (The PPV is the percentage of people with the clinical diagnosis who eventually prove to have the disease, in this case by autopsy.) 
  • For most of the Parkinsonian disorders, better biomarkers in living patients are urgently needed.  That, we knew.

One last take-home:  I’d say that the GP2 study shows medical science at its best, combining the latest technology with concern for – and active involvement of – populations and their physicians world-wide.  

We tauopathies have to stick together

Posted on by

A post from two months ago explained why some positive results from a trial of an antisense oligonucleotide (ASO) drug for amyotrophic lateral sclerosis (ALS) is good news for PSP. Now, there’s similar but even better news about an ASO drug for Alzheimer’s disease called “diranersen.” Why better? Because unlike ALS, Alzheimer’s disease is a tau-based disorder, like PSP.

Quickie review: ASO drugs interfere with the action of messenger RNA, which is produced by a gene in the cell’s nucleus and carries it out to the ribosomes, where it’s translated into that gene’s specific protein. Trials of “neuroprotective” treatments hope to demonstrate a decline in the rate of worsening relative to placebo, affording the participant more time at each disease stage. Neuroprotective treatments like ASOs are not designed to produce improvement relative to the study’s baseline.

Yesterday (July 14, 2026) the results of a Phase 2 trial were released at the Alzheimer’s Association’s annual research conference in London. The double-blind design included 406 people with either “mild cognitive impairment” (MCI) or mild dementia caused by Alzheimer’s. (MCI does not include difficulties in performing daily activities and dementia does. For the purposes of the trial, MCI is defined as a score of 21-27 on the 30-point Mini-Mental Status Exam and dementia as 20 or less.) The participants were randomly assigned to one of three dosage levels or placebo for the 76 weeks of the double-blind period. The improvement was calculated as the difference between the first and last visit scores divided by the the first visit score.

The testing included five standard cognitive tests, tau levels in the spinal fluid, and positron emission tomography (PET) scans to image brain tau deposits. The drug, like all ASOs to date, is administered by injection into the spinal fluid space at the base of the spine, as for a spinal tap. Of those seven tests, the “primary outcome measure” was the Clinical Dementia Rating Sum of Boxes (CDR-SB), presumably because it’s the one with the most experience behind it.

Officially, the trial was negative because the CDR-SB failed to show a statistically significant slowing of the rate of worsening. But for two of the other four cognitive measures, there were statistically significant degrees of slowing of 50% in one and 42% in the other. The benefits shown by the PET and spinal fluid look equally impressive to my eyeballs, but their degree of statistical significance was not provided.

Transient pain in the head, limbs and/or back from the spinal taps were by far the most common side effects, occurring in about half of all participants, but no worse on diranersen than on placebo. The only side effect clearly worse on diranersen was a temporary confusional state (about 25% vs 5% on placbo). In all cases, it was gone within a week.

You can see and download the company’s detailed announcement here. Keep in mind that the drug company sponsor, Biogen, wrote it for public relations purposes. It is not sufficiently detailed for a research journal and has not been peer-reviewed, but Biogen has announced plans for an expensive Phase 3 trial, proving that they’re willing to put their money where their mouth is.

Biogen hasn’t parted with further details on the Phase 3, but if it starts in 2027 and is similar in design to the Phase 2 (but larger), one might expect results in 2030 and if all goes well, FDA approval shortly thereafter. Too long, I know, and besides, that’s Alzheimer’s — not PSP.

Of course, the elephant in the room is the trial of NIO-752, the ASO from Novartis, currently in a Phase 3 trial for PSP and stated for completion in mid-2029. The success to date of diranersen in Alzheimer’s, is excellent news for the prospects of a similar anti-tau ASO for PSP. As far as I know, there’s no clinically relevant difference between the two ASOs.

On behalf of the PSP community, I’ll thank Novartis for its efforts behind NIO-752 and encourage Biogen to extend its so-far-favorable Alzheimer’s program to PSP. Any other Pharma companies interested? The opportunity is ripe!

—–

Disclosure: In the past, I have consulted for both Novartis and Biogen, but have never had a financial interest in either company. The only possible exception is that Rutgers University, which owns the rights to the PSP Rating Scale because I was a professor there when I developed it, shares a fraction of its licensing fees with me.

Zoë’s extraordinary statistical analysis

Posted on by

A new paper has analyzed data from 863 people with PSP from the failed Biogen and AbbVie monoclonal antibody trials from the late 2010s.  The purpose was not to re-sift the data for signs of benefit, but to characterize the participants’ cognitive deficits in detail and to measure their rate of cognitive worsening over the 12 months of the trials. 

This sort of thing has been done before, but not with as many patient or using those trials’ main cognitive measure, the Repeatable Battery for the Assessment of Neuropsychological Status (RBANS), a 12-component test that has become the standard cognitive test in drug trials of PSP and other diseases.  It includes tests of:

  • Immediate memory
  • Delayed memory
  • Visuospatial/constructional ability
  • Language
  • Attention. 

The present analysis also included four timed tests that other work has shown to be affected early in PSP and to be sensitive to its progression:

  • Phonemic and semantic fluency (saying as many words as possible in one minute meeting a condition such as things starting with the letter T or things found in a supermarket)
  • Wechsler Letter-Number sequencing test (connecting randomly arranged letters and digits in order: A-1-B-2, etc.)
  • Color Trails Test 1 (connecting small, randomly arranged circles in order of their enclosed numbers, ignoring their two background colors)
  • Color Trails Test 2 (similar, but alternating the background colors)  

Other features of this analysis worth noting:

  • A lot of other data about the participants’ PSP was gathered in standardized fashion, per the two trials’ protocols.  As you’d imagine, a test requiring writing and drawing could be affected by PSP’s motor and ocular deficits.  So, the statistical analysis adjusted each participant’s cognitive performance by their scores on the eye movement and/or limb movement PSP Rating Scale sections.
  • These participants were in the early-to-middle stages of PSP, at a mean of 3.3 years since symptom onset, with a standard deviation of only 1.4 years. So, one cannot assume that the conclusions of this study apply to people in earlier or later stages.  The same applies to their average age of 68.7 (sd 6.9) years.
  • The analysis combined the placebo and active-drug groups; it did not compare them to each other.  The authors felt that this would be fine because the original study’s comparison between the two groups’ responses to the drugs showed no differences, unfortunately.

The results showed that all scores worsened to a statistically significant degree over the 12 months except for story memory (immediately after the story), story recall (after a delay), list recall (immediate recitation of a word list) and list recognition (after a delay, indicating which words on a second list were also on the first).

The most rapidly progressing tasks were complex figure copying (with the original visible) and coding. (That’s where nine geometric symbols are presented beside the digits 1 to 9. The person is allowed to refer to the “code” on the same page while drawing the symbols corresponding to a list of random digits.)

The take-home for patients and families is that memory is relatively preserved in PSP, at least through the first four or five years.  This despite the habit of many laypersons of referring to any cognitive deficit as a “memory problem.”  So, cognitive activities relying on memory such as listening to music, comedy or stories may remain a source of entertainment and satisfaction for those with PSP well into the disease journey.  On the other hand, puzzles, crafts, drawing and strategy games are likely to prove frustrating.

Another use of these results is in the design of future clinical treatment trials.  The RBANS is a time-consuming, fatiguing test for both patient and clinician, requiring 45-60 minutes for those with PSP.  If the RBANS can be reduced to its most rapidly progressing, most informative components, the same (or better) information on the subject’s progression might be obtainable in a fraction of that time, without the fatigue.  This could improve the quality of the data not only on the cognitive test, but on the study’s other tests as well.  An excellent start at this was made a few years ago by scientists at Biogen.

The new paper’s first author was Zoë Cappella Cooper, a Harvard undergraduate headed for a stellar career.  (I know this because I’ve worked with her on other projects.)  The senior author was her mentor, Anne-Marie Wills, MD, a long-time colleague of mine at Harvard Medical School and Massachusetts General Hospital.

NIO-752 on the fast track

Posted on by

Some encouraging news from the AD/PD Conference held this past March in Copenhagen. NIO-752 is the anti-sense oligonucleotide drug being developed by Novartis for PSP and Alzheimer’s.  It reduces the ability of the messenger RNA encoded by the tau gene to be translated into tau protein.  The news is the results of a Phase I trial designed mainly to assess safety and tolerability of various dosage levels.  But the trial also included measures of efficacy, just in case something dramatic appeared despite the trial’s small size.  The PowerPoint slides from the presentation by Dr. Günter Höglinger of Munich can be downloaded here.

NIO-752 is a large molecule that can’t cross the blood-brain barrier, so it has to be injected directly into the spinal fluid via the same sort of needle insertion used in a diagnostic spinal tap.  In this trial, 80% of the 59 subjects (45 on active drug, 14 on placebo) received injections at baseline and at months 1, 2 and 3 at ascending dosage levels.  The other 20% received it at baseline and at months 3, 6 and 9. The final assessment for all subjects occurred at 12 months.  The trial ran from February 2021 to October 2024.

The drug caused very little by way of important side effects: confusion and or lethargy in two of the 25 patients on the two highest dosage levels and some brain inflammation in one patient on the highest level as evidenced by elevated white blood cells in the spinal fluid.  This is a very modest overall burden of adverse effects, and what’s more, the frequency of milder side effects was no different between the active drug (19 of 45) and placebo (7 of 14) groups. There’s lots more information at clinicaltrials.gov.   

An important goal of Phase I trials is to demonstrate “target engagement.”  That military-style term means the ability of a drug to accomplish its job in the body’s tissues regardless of whether it actually helps the person’s symptoms or long-term outcome.  NIO-752 did well on that score, reducing the spinal fluid tau levels, especially at the highest dose level.  (The level remained unchanged in the placebo subjects.)  It also prevented any rise in levels of neurofilament light chain (NfL), while the placebo group’s NfL rose by nearly 40%.  NfL is a sign of damage to axons (the long fibers emerging from brain cells) that is elevated in PSP and several other neurodegenerative diseases.

The question in your minds is, “But what about the rate of worsening of the symptoms and disabilities”?  That wasn’t reported because there weren’t enough patients to perform a proper statistical analysis.  If the worsening in the PSP Rating Scale had been less in the 45 participants on NIO-752 than in the 14 on placebo, that would be too few to exclude the possibility of confounders (a “Type I error” or false-positive) effect.  Similarly, if this small study failed to demonstrate a benefit, Novartis would not want anyone to draw negative conclusions about the drug without a properly powered trial.

Next step: Novartis will now skip directly to a Phase III trial and just yesterday they posted details on clinicaltrials.gov.  You can get a look here, but I’ll discuss that trial in a near-future blog post. The trial’s nickname is PRESERVE.

__

Full disclosure: I have consulted for Novartis on trial design in the past but not since 2023. I have no stock in the company nor any other financial interest in the success of NIO-752 or the company in general.

41 heads are better than one

Posted on by

CurePSP’s Centers of Care (CoC) network has just announced this year’s four recipients of its “Collaborative Approaches to Resources, Education and Support” (CARES) grants.

The CoCs, established in 2017 with 25 sites, have grown to include 41 sites at 34 academic medical centers in the US, four in Canada and three in Europe.  All have satisfied CurePSP’s criteria for expertise in the care of people with PSP and CBS, and many have also qualified for additional certification in MSA.  The group’s mission is not traditional research, but improving the quality and quantity of clinical care. 

In pursuit of that goal, the group has collaborated in writing and publishing review articles and opinion pieces on such topics as symptomatic treatment, biomarkers, speech impairment, and delays to appointments. Its four special-interest sub- groups meet regularly to discuss ways to improve care. Sites must be re-certified every three years, at which time they must demonstrate improvement in quantitative measures and cite new, relevant programs, publications and educational activities.

CoC member sites also are eligible to apply to CurePSP for grants supporting research into improving care delivery, or for specific care delivery projects.  Brief descriptions of the newest four projects are available on the CurePSP website. The funding for this year’s crop totals $155,000.  Parkinson Canada and CurePSP are sharing support of the two projects involving Canadian sites.

An unusual aspect of the CARES grant program is its requirement for collaboration among at least two CoC sites.  The point is to encourage cross-fertilization of ideas and to encourage the newer or smaller sites to learn what has worked for the more-established ones.  

Disclosure: I helped organize the CoC network back in 2016-17 and serve on its Steering Committee as an ex officio representative of the CurePSP staff.

The takeoff of your 747 has been delayed

Posted on by

I just heard that the PSP Trial Platform’s enrollment kick-off, set to happen this month, has been postponed to sometime later this year. 

You can read more about the PTP in this blog post of mine and on www.clinicaltrials.gov.  The delay, I’m told, is that the trial organizers and the sponsor of one of the three drugs have yet to nail down certain administrative issues.  That sponsor is Alzprotect, a French company, and its PSP drug is AZP-2006, an oral drug addressing abnormal tau disposal and inflammation.  The two drugs fully in place are AADvac1, a subcutaneously injected anti-tau vaccine from Axon; and LM11A-31, an orally administered nerve growth factor modulator from PharmatrophiX.   


A bit of good news is that the FDA has given the green light to the Alprotect drug to start its participation in the PTP trial.  That had been one of the causes of the delay, and now only those admin issues remain. 

So, the trial from start of this Phase II trial (the PTP) to full FDA approval, which if all goes well would have taken at least four years, will now need a few months longer.  But the PTP’s participating sites are enthusiastic, the financial support is in place, the underlying science is good, and the medical leadership is first-rate. I realize that these reassurances are scant comfort to those affected by PSP and their families, but they’re important.

I’ll pass along updates when I can.

The scale is tilting

Posted on by

The PSP Rating Scale has had quite the odyssey. I designed back in 1996 (30 years ago now!) to be used by neurologists or other clinicians working in movement disorders, and its specialized terminology reflects that.  But some people with PSP or their caregivers have attempted to apply the scale themselves.  That, and the continued importance of the scale in PSP research, prompt this update.

The scale and its tips/rules for its administration can be viewed and downloaded from the CurePSP website.

The PSPRS is intended mostly for use in routine clinical care and touches on all the common features of PSP.  It’s designed to answer the question, “So, how am I doing, Doc?” on a 100-point scale, with zero being perfectly normal.  It’s a measure of the severity of symptoms, not a set of diagnostic criteria.  The most recent such criteria, if you’re interested, are available here. They’re also pretty technical.

The PSPRS includes 28 items, 22 rated 0-4 and six rated 0-2, divided into six sections.  The first, “History” is rated via interview.  The others, all requiring a neurological exam, are Mentation, Bulbar, Ocular Motor, Limb Motor, and Midline/Gait.  (“Bulbar” means speech and swallowing, so named because the part of the brainstem where those things are controlled, the medulla oblongata, looks like a flower bulb or a fat scallion.)

The PSPRS scoring behaves thusly: The average person with PSP-Richardson syndrome, the most common PSP subtype, comprises about half of all patients) and worsens by an average of about 11 points per year, more slowly at the very start and end of their PSP journey.  PSP-Parkinsonism, accounting for about 20% of all PSP, progresses about two-thirds as quickly.  Most people entering clinical treatment trials (all with PSP-Richardson) have scores between 35 and 40, a figure heavily influenced by most trials’ rules that participants be in the early or middle disease stages.

Although I did not design it specifically for research use, the PSPRS became the world-wide standard measure in PSP research soon after its 2007 publication.  However, the time required to administer it has prevented it from attaining much popularity for its original purpose as a routine tool in patient care.  I can run through it in 10 minutes with a mildly affected patient, but physicians with less experience and patients with more complicated deficits require up to 20.  With only 20 or 30 minutes scheduled by most medical practices for follow-up visits, you can see why the PSPRS has not attained popularity outside of research projects, where the visits include many more tests and require several hours.

Another issue with the PSPRS is that some of its 28 items relate more to the neurologist’s concept of what’s going on in the patient’s brain rather than directly to daily disability, which, after all, is the primary target of both routine care and research.  Therefore, the FDA in the US has expressed a preference for its own 10-item version, which retains items only in the History, Bulbar and Midline/Gait sections and collapses some of the response options in most of the others.  Naturally, the lesser precision of the “PSPRS-10” requires that the study recruit more patients to provide the statistical power to demonstrate a given degree of drug benefit and provides no information at all on eye/eyelid movement or cognitive/behavioral deficits.  But researchers in Germany have applied a sophisticated statistical technique for drug trials called “item response theory” using all the study visits (typically about five over 12 months) rather than just the starting and ending visits.  This allows the PSPRS-10 to exceed the original in its statistical power to detect a drug benefit despite the lesser supply of information from each visit.

The anti-sense oligonucleotide (ASO) trial, called “PRESERVE” from Novartis is the first to use the FDA’s PSPRS-10 as its primary outcome measure.  The trial’s designers had little choice because as a Phase 3 trial, it is “pivotal” in the sense that it could make or break the drug in the eyes of the FDA.  However, PRESERVE will add the original 28-item version to its secondary measures. They can be considered by the FDA to clarify a weak result in the primary measure.  Its inclusion also makes life easier for future researchers seeking to integrate the study’s results into the large body of data available from previous drug trials.

Other abridged versions of the PSPRS-28 to date are:

  • A 14-item version omits the items less relevant to daily activities, with each remaining item rated 0-2.   It performs as well as the original for PSP-Richardson but not for the other variants, according to a trial by a different set of researchers.
  • A 15-item version, also a subset of the original, which uses the results of the unsuccessful PASSPORT trial to omit items less related to daily activities as well as statistically redundant items and response choices.  It offers better statistical power than the original or any of its other modified versions and is being used as the primary outcome measure in the PSP Trial Platform.  The PTP can defy the FDA’s recommendation to use its 10-item version because those are only Phase 2 trials, not the Phase 3 pivotal trials that the FDA uses to decide on a drug’s final approval. 
  • 21- and 24-item versions omitting PSPRS-28 items problematic for tele-health administration.  These have not yet reached widespread adoption for either routine practice or research, but they have excellent potential for use in pandemics or for patients with difficulty traveling to a neurologist for reasons of disability or distance.  The only potential problem is wrangling the software.

None of the PSPRS versions allows reliable administration by the patient or caregiver, but the Cortico-basal Ganglia Functional Scale (CBFS) does so.  In fact, it’s designed for use online or in the neurologist’s waiting room, with no instructions needed beyond what’s on the form. Despite its name, it has been validated for PSP as well as for corticobasal syndrome.  So, if you want to rate yourself at home, use the CBFS rather than any of the PSPRS versions, and coordinate with your neurologist so as to make best use of the results.

Despite all this, the original content and wording of the PSPRS-28 have never changed. I’ve only refined the instructions and tips published along with it.

Finally, a little perspective: Despite all this hard work on various rating scales, measuring atrophy of the brain’s frontal lob on MRI tracks progression of the disease better than any of them.  One careful analysis found that a trial designed to detect a 50% slowing of progression using the PSPRS-28 would need 112 subjects (half on active drug, half on placebo) and only 72 using the MRI measure.  Of course, the FDA would never accept such a measure in a pivotal trial, but at least one current, early-phase PSP trial is using MRI as its primary outcome measure.

I refer to the PSPRS-28 as my own work, but some of the credit belongs to my statistician colleague at Rutgers, Pamela Ohman-Strickland, PhD.  I couldn’t have validated the scale without her.  I know because I tried.

Disclosure: I (and Pam) have a financial interest in the PSPRS via a sharing policy at Rutgers University, which owns the copyright and licenses it to pharma companies for use in their drug trials.  Rutgers imposes no fee for its use in regular patient care or in research by non-profit organizations such as universities.

Neuroprotective tangles

Posted on by

Prompted by a reader’s comment a few days ago, I thought I should write about the difference between “neuroprotective” and “symptomatic” treatment.  The distinction is relevant to the PSP treatment trials about to start and to one’s decisions about whether to volunteer for them.

Neurodegenerative diseases, by definition, progress over time and neuroprotective treatments attempt to slow that process, not to provide relief from existing symptoms.  In theory, such treatments could be so miraculously effective as to halt the process in its tracks, but a realistic best-case target given our current understanding of these diseases is a 40%-50% slowing of the rate of future progression. Most PSP trials are designed to detect about a 25%-30% slowing.  A trial large enough to detect more subtle degrees of slowing would be prohibitively expensive.

For a drug to improve the existing PSP symptoms or disabilities, as levodopa improves those of Parkinson’s, for example, would require replacing a molecule deficient in the brain cells that still survive and function, or stimulating the surviving cells to work harder, or modulating the activity of other, healthy, brain cells to partly compensate for the effects of the damaged cells.  Such drugs do exist for PSP, but their benefits are modest and temporary, and the underlying neurodegenerative process continues.  In Parkinson’s, levodopa gives dramatic and long-lasting symptomatic benefit, but even there, the degenerative process continues unabated.

The graph below illustrates all this:

  • The vertical axis is the PSP Rating Scale, where 100 is the worse possible score and the average rate of worsening is about 11 points per year.  At a score of about 80, fatal complications such as pneumonia or severe urinary tract infections become very common.  Note that to avoid displaying blank space, the axis starts at 30 points, not zero.
  • The horizontal axis is years since the start of the treatment (the “baseline”).  You can see that for purposes of this illustration, I’ve chosen a baseline PSPRS score of 40, which conforms to experience with previous trials and to the observation that the average person with PSP doesn’t receive that diagnosis until about three years after symptom onset.
  • The blue line represents the course of the disease untreated.  In a drug trial, there’s usually a placebo effect, but to keep things simple, the graph ignores that.  Besides, that effect would dissipate over a couple of months at most.
  • The orange line represents the course 30% slower than that of the placebo group.  Note its shallower slope.  Again, for simplicity I show the effect as starting immediately upon receiving the drug even though some neuroprotective effects may take a few months to get going.
  • The green line shows a symptomatic effect, which in this example starts immediately and lasts years.  I’ve semi-arbitrarily chosen its magnitude to be five PSPRS points and the time to maximum benefit as six months.  At that point the rate of progression of the underlying disease continues unabated, but the five-point symptomatic benefit persists.  Note that the participants on such a drug are doing better than those on the successful neuroprotective drug until a bit after the three-year point, when the lines cross, and the advantage of the neuroprotection continues to widen.

HYPOTHETICAL COMPARISON OF EFFECTS OF PLACEBO,

NEUROPROTECTIVE AND SYMPTOMATIC TREATMENTS OF PSP

I’ll emphasize that while the 11 points per year rate of progression is based on real data, the 30% slowing of the rate of progression is only an illustrative example for the purpose of this instructional exercise. The five-point symptomatic improvement is analogous to the magnitude of improvement of Alzheimer’s disease treated with cholinesterase inhibitors such as donepezil, galantamine and rivastigmine.

The death of a brain cell isn’t like an incandescent light bulb suddenly burning out – it’s more like a slowly fading LED bulb.  During that “ill” phase, it might be possible for a candidate neuroprotective treatment to instead (or in addition) have a symptomatic effect. 

With all that as background, here are some conclusions:

As you’d imagine, it could be difficult to tell neuroprotective from symptomatic (or placebo) effects, as they’re both being measured by the same PSP Rating Scale.  But clinical trials in PSP try to anticipate this by testing for more objective evidence of slowing of brain cell loss, for example by assessing atrophy on MRI or spinal fluid levels of a protein called neurofilament light chain (NfL), which increases steadily in PSP and some other disorders. Placebo and symptomatic improvement would not be reflected in those diagnostic markers.

    Neuroprotection trials also perform their first repeat exams in the first few weeks and months to look for a rapidly-appearing difference in PSPRS scores between the active drug group and the placebo group. (MRI and NfL would not be useful so soon after baseline.)  However, it would be difficult to decide whether such a PSPRS improvement is placebo effect or symptomatic effect.

    There’s another way to distinguish a placebo effect from a physiologic effect (we avoid the term “real” because placebo effects are also real in their own way).  That’s to assess the participants for worsening couple of months after the trial’s end, when any symptomatic effect would have dissipated.  I haven’t seen the PRESERVE (Novartis’ NIO752 trial) protocol and its clinicaltrials.gov entry doesn’t address the matter, but I expect that it plans to do this, if only to monitor any adverse effects of the drug. 

    Bottom line: In a Phase 3 trial in a neurodegenerative disease, separating true neuroprotection from symptomatic and placebo effects is tricky. In future blog posts, I’ll try to sort out that tangle for you if I can.

    —

    PS #1: For an excellent, very recent review of the placebo effect, see this paper, which is written in language easily comprehensible to educated laypersons.

    PS #2: Disclosure: I consulted for Novartis from 2018 to 2020, but not since. I have never held stock or any other financial interest in the company. But I do hold a major emotional interest in seeing their drug work, so there’s that.

    A good problem to have

    Posted on by

    Now, this is progress.  Novartis just yesterday announced in clinicaltrials.gov that its Phase 3 trial of NIO752 is ready to accept volunteers for screening. The name of the trial is PRESERVE.  Good name.  No, it’s not an acronym for anything.

    So far, the company has only announced three trial sites (Rochester, MN; Englewood, CO; and Ulm, Germany) but dozens more will follow, with a total recruitment goal of 300.  Here are some details of probable interest to potential volunteers:

    • The drug is an antisense oligonucleotide, which as you’ve learned from this blog, interferes with the brain cells’ ability to translate the RNA from a specific gene into its protein.  In this case, the protein is tau, which lies at the heart of PSP.
    • As a very large molecule, NIO752 cannot pass the blood-brain barrier, so it has to be injected directly into the cerebrospinal fluid in the lower spine, using the same procedure as a diagnostic spinal tap (lumbar puncture).  This will be given every three months, assuming it follows the plan of the Phase 1 study.
    • A trial of the drug for safety in 45 people with PSP showed no important or permanent adverse effects from the procedure – just harmless and transient headaches or back pain in some.  Some transient confusion or lethargy occurred in three of the 45 – an effect of the drug, not the injection procedure.
    • The trial will enroll 300 participants overall, of whom 100 will be randomly chosen to receive a placebo injection.  That treatment assignment will be double-blind — not revealed to participant or neurological staff until the whole trial is over.
    • The duration of the double-blind period will be 72 weeks – about a year and a half.  After that, all the patients will be offered the opportunity to continue receiving the drug at no cost, as long as it has not been found to be harmful, and as long as Novartis is still manufacturing it.  That “open-label extension” program may end if and when the drug works and is on the market (let us pray).

    Your big question right now should be this: Should I volunteer for the PSP Trial Platform (PTP) or PRESERVE?  The scheduled start for PTP is next month (June 2026), and those sites will roll out gradually, just like the PRESERVE sites.  So, in theory, there’s no overall difference in the timing, though a site near you might open for one study well before the other, or there may be an accessible site for one and not the other.  All the PTP drugs and NIO752 are similarly and acceptably safe, in my view.

    Right now, I’d say volunteer for PRESERVE, though by a slim margin.  Two reasons, each minor:

    • Like any large, complicated project requiring approvals from government, private companies and academic institutions, the PSP Trial Platform has been subject to unforeseen delays.  (All major drug trials require collaboration among these three, but the PTP is more complicated than most.) In fact, the company sponsoring one of the three drugs planned for the PTP has still not finalized the arrangements, according to clinicaltrials.org. If that can’t be accomplished soon, the trial will start with only two drugs.  So, a bird in the hand . . .
    • The PRESERVE trial will have an open-label extension (see the caveats above), while the PTP has not yet decided on that, and it may differ across the different drugs.  Without an open-label extension, someone completing the Phase 2 trial would have to wait until the Phase 3 is finished and the drug approved before gaining access to it.  On the other hand, the FDA has been known to approve drugs for general use after only a Phase 2 if the need is great, and for PSP, it surely is. The PTP double-blind trials are 12 months long and the PRESERVE double-blind is nearly 18 months, so assuming both offer open-label extensions, someone on placebo in PRESERVE would have to wait six months longer to receive their active drug than someone in the PTP.

    Yes, there are other drugs whose sponsors are optimistic that trials will start within the next year or so.  Those include bepranemab (a monoclonal anti-tau antibody), GV1001 (an anti-inflammatory), ARV-102 (an enhancer of abnormal tau degradation) and TPN-101 (an inhibitor of a toxic protein called LINE-1).  But the timelines there are just too uncertain for someone with PSP to consider right now.

    Maybe the most important consideration is which drug is mostly likely to work.  I honestly don’t know, and the Phase 1 data don’t answer that question.  So that simplifies things a bit.

    A difficult choice, I know, but a good problem to have. 

    ASOs: sci-fi takes a step closer to reality

    Posted on by

    Great news from Biogen about an antisense oligonucleotide (ASO) designed to reduce production of the tau protein. 

    First, some background: Most of you have heard about ASOs, but for a refresher, see these posts of mine from 2022 and 2026.  Here’s a slightly more technical but cutely animated explanation of ASOs from Harvard Medical School:

    The elevator version is that an ASO is a short length of RNA that binds and inactivates the brain’s messenger RNA for a specific protein – or an abnormal version thereof – to prevent it from carrying the protein’s genetic code from the DNA to the protein-manufacturing machinery.  In theory, the production of any protein involved in the cause of a disease can be reduced by designing an appropriate ASO to bind to a segment of that protein’s messenger RNA.

    The FDA has approved only one ASO so far – for a childhood muscle disorder called spinal muscular atrophy – but dozens of other ASOs are in the development pipeline for other conditions, including tauopathies. Biogen is currently testing its anti-tau ASO, called diranersen (formerly BIIB-080) against Alzheimer’s, by far the most common tauopathy.  A few days ago, they announced the results of a Phase 2 study of its safety and tolerability.  Here’s Biogen’s press release and here’s the description of the trial (without results) in clinicaltrials.gov. 

    Diranersen was well-tolerated in people with Alzheimer’s, as expected based on the Phase 1 results.  The big news was that the rate of accumulation of abnormal tau protein aggregation actually did slow down, as measured by levels of tau in the spinal fluid and by positron emission tomographic (PET) images of the tau protein’s distribution in the brain. The press release didn’t say how much slowing occurred, but it was apparently enough to convince Biogen to proceed to a Phase 3 trial and to convince the FDA to let them do so. More details will be presented at the Alzheimer’s Association International Conference in London in July 2026.

    The trial was not primarily designed to assess slowing of progression of the participants’ actual cognitive loss, but it gathered that information anyway in various forms.  The primary such test, called the “Clinical Dementia Rating Scale Sum of Boxes,” measures memory, orientation, judgment/problem solving, community affairs, home/hobbies, and personal care. It did not show a statistically significant slowing of progression for diranersen in Alzheimer’s, but the press release hints that there was some slight, statistically non-significant, degree of slowing.  

    That’s all about Alzheimer’s. For PSP, a different company, Novartis, is testing a different anti-tau ASO (NIO-752).  It is also well tolerated, as demonstrated by a recently-completed Phase 1 trial.  If the brain’s accumulation of abnormal tau can be slowed down in Alzheimer’s disease, as the Biogen press release claims, then presumably PSP can achieve the same result.  Novartis says it’s still (as of May 15, 2026) analyzing its Phase 1 PSP efficacy results, but those would have to be spectacular to show statistical significance in so small a study.  That company will soon start testing NIO-752 in a Phase 2 PSP trial in the US and other countries, so keep an eye on clinicaltrials.gov for enrollment instructions.

    Given these new results of one anti-tau ASO in one tauopathy, what are the prospects for a different anti-tau ASO in a different tauopathy?  I’ll duck the issue and call them promising but far from a slam dunk. That will be the topic of a future post, but what I can say right now is even these modest, preliminary signs of success with ASOs in tauopathies would have been science fiction back when I was in med school 50 years ago.