The ten PSP phenotypes

Posted on by

“Pheno-“ is the Greek root for “outward appearance” and so far, PSP has ten of them. The differences arise from varying emphasis of the degenerative process among different parts of brain.  While standard laboratory methods show that at the cellular level the pathology among the ten phenotypes is identical, a few details are starting to emerge using more recent and sophisticated techniques at the molecular level. 

The most common PSP phenotype, called PSP-Richardson’s syndrome (PSP-RS), is the one Steele, Richardson and Olszewski originally described in 1964.  The others were published piecemeal starting in the early 2000s.  J.C. Richardson was the leader of the trio at the University of Toronto, a senior clinical neurologist who noticed an unusual form of Parkinsonism among his patients in the 1950s and 60s.  John C. Steele was his trainee and Jerzy Olszewski was the neuropathologist who described the corresponding microscopical abnormalities.  So, it’s altogether fitting and proper that Dr. Richardson should be honored in this way.

Prevalence of the phenotypes. The percentage of the PSP population with each phenotype has not been studied in a true community-based population.  The published percentages vary widely across centers and are all from referral-based populations at research institutions, where unusual forms of diseases are over-represented to varying degrees.  Even without that issue, several things make it hard to be sure of the prevalence of the various phenotypes:

  • It’s difficult to estimate the population prevalence of atypical cases of PSP from autopsy series because atypical cases are more likely to come to autopsy, and without autopsy, it’s hard to know that someone with atypical PSP really had PSP.
  • In their later years, all of the phenotypes tend to merge into a PSP-RS appearance, so the relative frequencies of the phenotypes may depend on the patients’ disease stage when the researchers evaluated them.
  • Clear diagnostic criteria do not yet exist for many of the phenotypes and many patients satisfy criteria for more than one, even in early stages.  A method has been published for how to deal with this, but most of the publications antedate or ignore it. 
  • Although the original differentiation of PSP-RS from PSP-P in 2005 did use a rigorous statistical technique called “factor analysis” to confirm that the two are distinct, this is not usually the case for the other phenotypes relative to PSP-RS or to one another.

Nevertheless, here are my very rough estimates of their contributions to PSP in general, based on a Gestalt impression of the literature:

Richardson’s syndrome45%PSP-RS
Parkinsonism25%PSP-P
Frontal10%PSP-F
Progressive gait freezing5%PSP-PGF
Speech/language5%PSP-SL
Corticobasal syndrome3%PSP-CBS
Postural instability3%PSP-PI
Ocular motor3%PSP-OM
Cerebellar<1%PSP-C
Primary lateral sclerosis<1%PSP-PLS

Many recent research articles group these into three categories based on their anatomical predilections in the brain: cortical vs subcortical.  PSP-RS falls into neither of these because it has approximately equal degrees of both cortical and subcortical features. One important, practical reason for the grouping of phenotypes is to have groups large enough for meaningful statistical analysis. 

  1. PSP-RS                      
  2. PSP-cortical: PSP-F, PSP-SL, PSP-CBS
  3. PSP-subcortical: PSP-P, PSP-PGF, PSP-PI, PSP-OM, PSP-C, PSP-PLS

PSP-Parkinsonism. The most common “atypical” (i.e., non-PSP-RS) phenotype of PSP is PSP-Parkinsonism (PSP-P). Relative to PSP-RS, it features more asymmetry, generalized bradykinesia, tremor, and levodopa responsiveness, and only later displays falls and cognitive loss.  It is usually initially misdiagnosed as Parkinson’s disease.  It has perhaps the slowest course among the PSP phenotypes, averaging about 9 years’ survival from symptom onset.  This compares with about 6 years for PSP-RS and intermediate figures for the other phenotypes.  In fact, the PSP-subcortical group as a whole has a similarly longer average survival duration than the PSP-cortical group as a whole or the PSP-RS + PSP-cortical groups.

PSP-progressive gait freezing. After PSP-P, the most common atypical phenotype is PSP-progressive gait freezing (PSP-PGF). In fact, most patients exhibiting only progressive gait freezing will eventually develop diagnostic features of PSP. The central feature of PSP-PGF is loss of ability to continue ongoing gait, especially after a pause, during a turn, or at a doorway threshold. In advanced cases, the patient cannot initiate gait at all. The picture also includes rapid, small handwriting and rapid, soft speech as frequent or severe features. The anatomic location of the pathology in such cases differs from that of PSP-RS in showing less involvement of the base of the pons (part of the brainstem) and of the dentate nuclei (part of the cerebellum).

PSP-speech/language.  This is a composite category.  In PSP-nonfluent/agrammatic variant of primary progressive aphasia (PSP-nfaPPA) speech is halting, with poor grammar, syntax, and pronunciation, but with normal comprehension and naming. A mirror-image variant called semantic-variant primary progressive aphasia (svPPA) features difficulty in naming with reduced vocabulary but with normal grammar and syntax. Together, PSP-svPPA and PSP-nfaPPA are referred to as PSP-speech/language disorder (PSP-SL).

PSP-corticobasal syndrome. CBS as a clinical syndrome (meaning a group of signs and symptoms that occur together, although the underlying disease may differ across patients) comprises highly asymmetric rigidity, slowed movement, and apraxia (loss of skilled movement), often with equally asymmetric dystonia (fixed postures), pyramidal findings (weakness and abnormal reflexes), myoclonus (small, rapid, irregular movements), and cortical sensory signs such as astereognosis (inability to identify objects by feeling them) and agraphesthesia (inability to identify figures traced on the skin). Aphasia (difficulty processing language) and other abnormalities localized to specific brain areas may also occur. Dysarthria (difficulty with pronunciation) can be prominent but gaze palsy, postural instability, and cognitive loss tend to be later and milder than in PSP-RS.

PSP-frontal. More formally called PSP-behavioral variant frontotemporal dementia (PSP-bvFTD or simply PSP-frontal), this phenotype features disinhibition, irritability, apathy, and loss of empathy for others, along with impairment in frontal “executive” functions such as ability to maintain attention, to follow instructions, to shift tasks on command, and to inhibit an ongoing action when appropriate. This is the core of the cognitive and behavioral deficits in PSP-RS, but when it appears first and remains worst, the term PSP-F is appropriate.

PSP-ocular motor and PSP-postural instability. Perhaps unsurprisingly given the cardinal features of PSP-RS, PSP can also take the form of a relatively pure ocular motor picture or a relatively pure picture of severe postural instability with falls and little else to suggest PSP. However, reported cases are very sparse to date. These have been designated PSP-ocular motor (PSP-OM) and PSP-postural instability (PSP-PI).

PSP-primary lateral sclerosis. The pathology of PSP can also produce the clinical picture of primary lateral sclerosis. PLS is one of the phenotypes of amyotrophic lateral sclerosis (ALS; Lou Gehrig disease) and can be difficult to distinguish from it, especially as ALS can produce frontal cognitive difficulties in many cases.  The clinical picture of PSP-PLS is highly asymmetric and resembles that of CBS but with little or no cortical sensory loss (spatial sensation ability), dystonia (fixed postures), or myoclonus (very quick, small, irregular involuntary movements).

PSP-cerebellar. The classic lurching gait of PSP-RS has a cerebellar appearance, the speech of PSP has an ataxic (or drunken-sounding) component in many cases, and the ocular square-wave jerks of PSP occur commonly in cerebellar disease. For decades, these were considered minor and inconsistent features, but in 2009, neurologists in Japan described a PSP phenotype of PSP with involvement of the cerebellum at autopsy and early, prominent ataxia of the trunk and limbs.  Although that original report found PSP-C in 14% of all people PSP at a Japanese center, the figure is much lower in Western populations for unclear reasons.

A word about drug trial eligibility.  Until the cause of PSP in general is better understood, neuroprotection trials — those aimed at the fundamental brain cell loss rather than merely at ameliorating the symptoms — will continue to recruit only patients with PSP-RS. Why?

  • We don’t yet know if the non-PSP-RS phenotypes share – either with PSP-PS or with one another — the same molecular abnormality being targeted by the drug.
  • While the classic PSP pathology underlies close to 100% of PSP-RS, that figure is much lower for some of the other nine phenotypes.  That means that non-PSP-RS may well be a non-PSP pathology, and that admitting participants with non-PSP-RS to a drug study runs a risk that some may not have PSP at all.  This could obscure any benefit the drug may have unless the trial is prohibitively large.  
  • The main outcome measure for nearly all PSP trials is the PSP Rating Scale, which was designed for, has been validated for, PSP-Richardson’s syndrome alone.
  • Trials to slow the progression of PSP use the rate of progression as the “outcome variable” of the trial.  As noted above, the phenotypes do vary in their expected survival durations, and by inference, their progression rates.  Therefore, including phenotypes with different inherent rates of progression would require a larger, longer and more expensive trial. One solution would be to make sure the active drug group and the placebo group receive similar proportions of the various phenotypes (a time-proven technique in trial design called “pseudo-randomization”).  But this doesn’t solve the four preceding problems.
  • As PSP-RS progresses faster than the other nine phenotypes, a trial enrolling only that phenotype can reach a result in a shorter time or with fewer participants.  This isn’t only financially advantageous for the trial’s sponsor.  If the drug is ineffective or harmful, fewer patients will have been exposed to it, and if it’s effective and safe, it will reach market approval more quickly.   
These graphs are from a review of records of 32 people with autopsy-proven PSP in Spain and Germany.  The lead author was Dr. Mar Guasp of Hospital Clínic de Barcelona and the senior author was Dr. Yaroslau Compta of the same department. They show the difference in survival from initial symptom to death between the PSP-RS and non-PSP-RS (top) and between [PSP-RS + PSP] and PSP-cortical (bottom). 
Graphs A and C: For each box, the horizontal line is the median, the upper and lower borders are the 25th and 75th percentiles and the ends of the “whiskers” are the highest and lowest values. 
Graphs B and D show the same thing in the form of “survival curves” or “Kaplan-Meier curves.” The vertical axis shows the fraction of the original patients still living at the time (post-onset) shown on the horizontal axis.
The p values are the likelihood that the difference could have happened by random chance.  The asterisk indicates that this likelihood is low enough for the difference between the groups to be considered “statistically significant.”  (Statistical veterans: Sorry to belabor this for the benefit of the statistical novices.)

To-do list:

  • Let’s figure out why the disease spreads through the brain in PSP-RS and the other PSP-cortical phenotypes more quickly than in PSP-subcortical.  Efforts to do that have in fact begun and could provide the key to the whole puzzle of PSP.
  • Let’s agree on a way to enroll people with non-PSP-RS phenotypes into clinical trials.  Current efforts to diagnose PSP using tau-based positron emission tomography (PET) and measures of tau in skin, blood or spinal fluid could potentially identify people with PSP other than PSP-RS who could potentially join a trial.
  • Let’s educate neurologists to identify, or at least to suspect, the non-PSP-RS phenotypes.  This would allow them to avoid or delay fruitless diagnostic testing and to provide their patients with useful prognostic information.

. . . and we have liftoff

Posted on by

The first of a raft of experimental PSP drugs in the pipeline has begun enrollment in a large treatment trial.

The ORION trial (don’t ask me what the acronym stands for) is sponsored by Amylyx Pharmaceuticals.  Their press release is here. A bit more information is at the company’s website. Their email address is clinicaltrials@amylyx.com.

Clinicaltrials.gov lists more info, including a phone number (in the UK):  +44 (808) 1642604. Only a few of the planned 33 sites in the US are open so far and no sites outside of the US has yet opened, to my knowledge. Those will be in Europe, Canada and Japan.

The product is actually two drugs called taurursodiol and sodium phenylbutyrate.  They have a variety of beneficial actions in brain cells affected by PSP, but it’s not known which would be most important.  Both drugs are already on the market, the first over-the-counter and the second with prescription.  Neither alone is intended for brain diseases and each alone has only a minimal effect on the brain.  But the two together have a synergistic effect on the brain and spinal cord that has proven modestly beneficial in ALS and has already received FDA approval for that disorder under the brand name Relyvrio.

There’s a lot more you’d want to know about the study’s design and the participants’ obligations, so please read the material at the links I’ve provided above and contact the company directly. 

Amylyx has hired a contract research organization called Cronos to actually run the trial.  They’re a subsidiary of a company called IQVIA.  So if you encounter those names, no worries.

One more little thing: full disclosure. I’m a paid consultant for Amylyx, but that’s only for advice in the design of the trial and instructing the examining doctors how to use the PSP Rating Scale. They don’t pay me to recruit patients and my fee does not depend on the trial’s enrollment. Nor do I own stock in the company or have any other financial incentive to see the drug succeed. (Of course, I do have an emotional incentive for that — no secret there.)

Seek and ye shall find

Posted on by

This week, our knowledge of the genetics of PSP has more than doubled.  First, as usual, some background:

Like many other complex conditions like atherosclerosis, schizophrenia and most cancers, PSP does run in families a bit more often than expected by chance.  But as in those diseases, the familial tendency is too weak to produce the classic dominant or recessive pattern associated with a single, strongly-acting gene variant as in Huntington’s, Tay-Sachs or sickle cell anemia. Besides, adding up the risks from the known PSP-related genes wouldn’t explain the incidence of the disease in the population, rare though it is.  That has prompted the theory that some unidentified external exposure or experience also has to play a role. 

Over the past 25 years or so, a number of gene variants have been found to confer slight risks for developing PSP.  The first-discovered and still the most important, called the “H1 haplotype,” is a complex set of variants in region of chromosome 17 that includes MAPT, the gene encoding the tau protein. Another four variants on other chromosomes were published in 2011 by CurePSP’s PSP Genetics Consortium. 

In the years since, nine other variants were added piecemeal by other researchers. Those first 14 were all discovered using a technique called “marker association,” which only identifies a region of about 100 genes where the culprit gene would be located.  The gene from those 100 that’s reported as a “hit” is generally the one with the best statistical association with the marker along with a scientifically rational reason to be associated with the disease under study.  A more finely-grained search would actually work out the sequence of the genetic code, comparing people with PSP to those without PSP.  That wasn’t practical back in 2011, but now it is.  It’s called “whole-genome sequencing” or WGS.

The new list of gene variants has been found by an international WGS collaboration that grew out of the original CurePSP-supported team.  They used DNA samples from 1,718 people with PSP, of whom 1,441 were autopsy-confirmed, and 2,944 samples from people without PSP as controls.  The leaders are at the University of Pennsylvania and UCLA, but 26 other research institutions in nine countries contributed.

They confirmed five of the six previously-identified variants (the sixth came very close) and added seven new ones. They also elucidated new details of the cluster of variants in the H1 region.  Most remarkably, they confirmed a previous, smaller study showing that PSP reverses the relationship of Alzheimer’s disease with the ApoE gene on chromosome 19.  In AD, the epsilon 4 variant of ApoE is over-represented relative to controls and the epsilon-2 variant is under-represented, while in PSP, it turns out that those proportions are reversed despite the fact that both AD and PSP are tauopathies.

So far, the research article is only posted on medRxiv (“med archive”), a website for manuscripts not yet through the peer review process at a journal.  (But my brain’s blogging center couldn’t restrain itself.)  The next steps for the authorship team are to gather online comments on the manuscript from other scientists and to submit the resulting revision to a regular journal.  There, the peer review may dictate other changes.  The next scientific step will be to figure out what the mutations are doing wrong, determine to what extent the variants increase or decrease the amount of the protein they encode (called “expression studies”), and look for proteins encoded by those genes (or for proteins they interact with) that might be modulated by drugs.

As far as I can tell, even the newly expanded list of risk variants doesn’t explain enough of the overall cause of PSP to be used as a diagnostic panel.  But it’s a start in that direction.

My canned lecture on PSP includes a slide on the two dozen or so most important scientific milestones in PSP research since the disease was first described in 1963.  This paper is going there.  As I learn more about the publication progress and clinical implications of this work, I’ll keep you all apprised.

Seeds of a revolution?

Posted on by

Decades ago, the discovery that specific proteins aggregated in the brain cells of specific neurodegenerative diseases was a major advance.  But like so many other scientific breakthroughs, it created another question: Why are there so many different clinical pictures among different people with the same neurodegenerative disease (like PSP) despite the fact that they all host the same aggregating protein (in this case, tau)? The ability of abnormal tau to “seed” the disease process into previously healthy brain areas is at the root of the disease process, but we’ve had scant clue as to how that works, exactly.

For PSP, the most important clinical variable is the eight subtypes (PSP-Richardson’s syndrome vs PSP-Parkinsonism vs PSP-progressive gait freezing, etc), and slightly less variable features are the onset age and rate of progression.  In the past year or two, it’s become clear that the different subtypes tend to emphasize different areas of the brain, but that doesn’t explain why two people with the same subtype can have different onset ages and rates of progression.

This mystery became even more mysterious recently when a new electron microscopy technique called “cryo-EM” proved able to visualize individual protein molecules. It showed that for everyone with a given disease, the protein for that disease had the same misfolded shape.  In other words, the tau molecule assumes the same rigid squiggle in everyone with PSP, a different rigid squiggle in everyone with Alzheimer’s, yet another in everyone with corticobasal degeneration, and so on.  But that raised the question as to the reason for the variability among patients of the PSP onset age and rate of progression.

Now, researchers at the University of Toronto’s Rossy Centre, an institution dedicated solely to PSP research at the , have found new evidence supporting the old idea that the key may be in the “oligomers” or “high-molecular weight tau” or “HMW tau.”  These are stacks of tau protein molecules small enough to remain dissolved in the brain’s fluids, as opposed to single molecules or the large, insoluble neurofibrillary tangles visible through a conventional microscope. 

The top-line result was that the patients with more rapidly-progressive PSP and brain regions with the worst damage had higher levels of HMW tau.  In a tour-de-force of lab experiments, the Toronto researchers also showed that:

  • HMW tau was more resistant to the brain’s mechanism for breaking down such protein clusters.
  • The study’s 25 PSP patients could be divided into high-, medium- and low-seeders based on the speed with which their tau converted healthy tau to their own misfolded form.
  • Tau with phosphate groups attached to amino acids 202 and 205 were least likely to form the HMW tau clusters.
  • The pattern of production of proteins (i.e., the “proteomics”) in the brain areas rich in HMW tau showed disruption of the brain’s adaptive immune system and two other cellular systems previously known to be related to neurodegeneration.

The importance of all this is that we now have a more specific idea of the structure of the most toxic form of tau aggregates and that boosting the brain’s adaptive immune system with medication could discourage the seeding of misfolded tau into healthy cells.

The study’s first author, Dr. Ivan Martinez-Valbuena, published an editorial in the journal Brain Pathology explaining all this in language that non-specialist scientists can understand.

The research paper itself is posted by the authors in bioRxiv (“bio-archive”) an on-line, open-access website for articles awaiting word from the peer-review process at a conventional journal. Its senior author is Dr. Gabor Kovacs, one of the world’s leading neuropathologists in the field of neurodegenerative diseases.

It’s awards season

Posted on by

A caregiver has asked me, as CurePSP’s Chief Clinical Officer, to list the most important clinical research advances in PSP of 2023. Happy to oblige. Here are my top five in no particular order. 

  • The FDA approved a combination of two drugs called taurursodiol and sodium phenylbutyrate with the brand name “Relyvrio” for use in amyotrophic lateral sclerosis (ALS; Lou Gehrig disease).  A trial in PSP has already started to recruit patients.  The drugs address an issue in the mitochondria shared by the two diseases in different sets of neurons.
  • Tau PET ligand APN-1607 received go-ahead from the FDA to proceed to a pivotal Phase 3 trial.  Such a trial began recruitment in December in the US and will involve multiple other countries as well.  The compound would allow a diagnosis of PSP in early or equivocal cases by being taken up by the abnormal tau protein in the brain and imaged.
  • A drug called TPN-101 was found to be safe and well-tolerated in a Phase 1 trial of 30 patients with PSP.  The drug counters inflammation in the brain by reducing the transcription of ancient viral DNA in our genome.  Next is a small trial for efficacy.
  • A simple, remote, gait-monitoring system with only three sensors proved able to distinguish the gaits of PSP and PD.  Further testing for its ability to document progression or improvement will follow.
  • PET imaging of frontal lobe synapses showed good correlation with the PSP Rating Scale and with the results of cognitive testing.  This is different from typical PET in neurodegenerative disease, which images glucose utilization or protein aggregates.  The work suggests that synaptic imaging could be a good diagnostic marker in the earliest, pre-symptomatic stages of PSP.

But the most important piece of news is that several drug companies are planning to start clinical treatment trials in the next year or two. I’ll report on all that as it happens.

Dizzy

Posted on by

Dizziness is a common but poorly understood symptom in PSP.  That word can mean at least three things:

  • a lightheadedness that seems a prelude to fainting, usually caused by a drop in blood pressure
  • a sensation of movement of the body or the environment (usually spinning, but sometimes rocking or gliding)
  • a vague sensation of being off balance. 

Impairment of autonomic function is common in the Parkinsonian disorders such as Parkinson’s disease, dementia with Lewy bodies and multiple system atrophy (in ascending order of likelihood).  There’s less evidence for, and research on, autonomic dysfunction in PSP. 

As background: The autonomic system is divided into sympathetic and parasympathetic portions.  Generally, the sympathetic system performs “fight or flight” functions such as raising blood pressure and heart rate in response to stress and the parasympathetic the “vegetative” functions such as powering the intestines to digest food.  When the sympathetic portion is damaged by a neurodegenerative disease, the most common symptom is lightheadedness upon standing.  We’ve all experienced this to a mild degree, but when severe, the result can be fainting, falls and injuries.

A well-designed project reported in the current issue of an obscure journal called The Polish Journal of Neurology and Neurosurgery has now compared PSP with PD and MSA with regard to various aspects of autonomic function including drops in blood pressure upon standing, or “orthostatic hypotension.”  The authors are Drs. Jakub Malkiewicz and Joanna Siuda of the Medical University of Silesia, in Poland.  (Disclosure: One of the two co-editors of the journal is an old friend and colleague of mine, Dr. Zbigniew Wszolek of the Mayo Clinic Jacksonville, who has special expertise in the atypical Parkinsonisms.)

The blood pressure was recorded in a very careful, standardized way:  It was measured after 15 minutes flat on a tilt table and then 5 minutes after being raised to a 60-degree angle.  The result was that none of the 25 patients with PSP or 20 healthy, age-matched controls experienced systolic blood pressure drops of more than 20 points.  However, this did occur in 20 (26%) of the 76 patients with PD and in 7 (58%) of the 12 with MSA.

This result is consistent with my own experience, where an office version of this test in my patients with early or moderate PSP experiencing dizziness rarely elicited much drop in blood pressure or intensification of the symptom.

How can this guide the care of people with PSP?  It means a few things:

  • In people with early PSP, the possibility of low blood pressure is highly unlikely as a cause of dizziness.
  • If the neurologist can rule out an inner-ear disturbance by the absence of a sensation of movement, nausea, a change in hearing, or a rhythmic, abnormal eye movement called nystagmus.  That leaves the possibility of a brain disturbance – either the PSP itself or an unrelated issue with the balance mechanism.
  • A neurologist whose patient with known PSP reports dizziness should provide the same careful assessment as anyone else with the same symptom, rather than simply to diagnose low blood pressure due to dehydration or an excessive dosage of antihypertensive medication. 

A “brain disturbance” causing vague dizziness could be things like:

  • a chronic subdural hematoma from a fall (which could be drained surgically)
  • a minor stroke (which could prompt the addition of stroke prevention measures and a workup for treatable arterial narrowing)
  • a small tumor (which could potentially be removed or irradiated)
  • the effects on the brain of a medication (which could be reduced or discontinued)
  • unusual seizures (which could be prevented with medication) – the most distant possibility

Nevertheless, in someone with advanced PSP (not the group studied by Malkiewicz and Siuda), it’s also important to seriously consider hypotension as the cause of dizziness.  Here’s why:

  • When you can’t swallow comfortably or get up for a glass of water on on your own, it’s easy to become dehydrated.  
  • Limiting fluids to avoid nighttime incontinence or bathroom trips can do the same. 
  • Dopamine-enhancing drugs for Parkinson’s prescribed for PSP can reduce the blood pressure. 
  • Some drugs for urinary incontinence can do the same by dilating the blood vessels. 
  • Diabetes, even if undiagnosed, can damage the sympathetic nerves (“autonomic polyneuropathy”). 
  • Diuretics prescribed for swollen ankles can reduce blood pressure.

Sorry to give you all these things to worry about, but use them to question your doctors about whether they’re doing everything they can to diagnose and maybe treat your dizziness.

My thanks to Drs. Malkiewicz and Siuda for directing our attention to this still-understudied issue.

Wired

Posted on by

With a nice handful of medications for PSP approaching clinical trials, it would be great to be able to assess the participants’ movement ability not just every few weeks to months at the research center, but also much more frequently at home.  The reduced need for clinic visits would ease participation for patients who for whatever reason have difficulty tolerating or obtaining travel.  It could also provide a more “real-world” picture of how the patient is doing in their home environment.

One relatively easy step in that direction arose from a project published last year (in which I, full disclosure, was senior author).   It modified the 28-item PSP Rating Scale, omitting the exam items that might not work well by video and used existing databases of PSPRS scores over time to assess the correlation between the modified and unmodified scores.  In short, the correlation was excellent.

But a PSP Rating Scale modified for video still requires a video connection, and that can be tough for the PSP age group and their caregivers, especially those where cell service is spotty.  Besides, video visits can’t happen every day or even close to it.  So, some other gadget would be nice.

Now, a group led by Dr. Alexander Pantelyat of Johns Hopkins and Dr. Anne-Marie Wills of Mass General (the co-senior authors) with Dr. Mansi Sharma of Mass General (the first author) have published a first-blush look at a simple gait monitoring system in PSP and Parkinson’s. 

Other versions of the same idea for PSP have had to be used in a lab at a research facility and required a complex array of sensors pasted to various parts of the body.  But this one is used in the patient’s home and requires only three sensors: one strapped to the lower back with a belt and one fastened to each shin by what looks like an old-fashioned garter strap like my father used to wear.  For reasons of safety, only patients with histories of very few falls and ability to walk unassisted qualified for this early trial.  Patients with PSP and Parkinson’s were compared on their performance of four standard gait tasks.  They received instructional videos and the three sensors communicated with an app on a tablet provided.

Of the 22 patients who qualified and consented, only two (both with Parkinson’s) couldn’t manage the technical requirements.  For the others (10 with PD and 10 with PSP), the device proved able to quantify and time the movements well and to differentiate PSP from Parkinson’s. Most important was that managing the experimental hardware and software while avoiding falls or other complications was perfect.

The next step will be to assess the device over a period of several months for its ability to track PSP progression.  This should be successful because the Spearman correlation coefficients of the three gait measures with the modified PSP Rating Scale, were pretty good: 0.62, 0.64 and 0.84; and we know that the PSPRS tracks PSP progression well.  (Correlation of 1.0 is perfect and 0 is random.) 

Another reason to be optimistic about the device to track progression is that it’s already been accomplished, although with a more complex, six-electrode device implemented in a research lab.

A reason for caution is that not every patient in a drug study walks unassisted at home as safely as these 20 hand-picked participants, especially toward the end of a one-year trial period.  Furthermore, using this device at home in routine clinical practice would involve patients at all levels of gait instability.  But for people in remote areas or whose caregiver can’t afford to take time off from work for a clinic visit, this could be the ticket to research trial participation.

The hindbrain steps forward

Posted on by

The cerebellum is gradually being understood as a contributor to cognitive and behavioral function in both in health and disease.  A new publication has teased out MRI changes in the cerebellum that differentiate PSP from other dementing disorders early in the disease.  This pattern could be developed into a diagnostic test and as a marker of disease progression and even as a guide to rehabilitation measures.

The cerebellum is classically thought of as a regulator of movement.  In its most simplistic essence, its job is to put a brake on voluntary movement instructions from the cerebrum.  The cerebellum is guided in this task by perception of the position and motion of the trunk, head and limbs, by the effect of gravity, all complemented by visual input.

More recently, the cerebellum has demonstrated a memory function when it comes to movement regulation (making “muscle memory” more than just a metaphorical expression), and damage to certain parts of the cerebellum can cause a behavior disinhibition and cognitive impulsivity similar to the frontal lobe damage seen in PSP. In that sense, the cerebellum still functions as a “brake,” but on behavior and cognition rather than just on movement.

Now, researchers from the University of California San Francisco have carefully analyzed routine MRI scans from people with dementia arising from a variety of neurodegenerative conditions including PSP.  They specifically quantified gray matter damage.  (Gray matter is brain tissue composed mostly of cell bodies — as opposed to white matter, which is mostly axons.  In the cerebellum, unlike the cerebrum, the gray matter is the deeper layer and the white matter is superficial.)

The figure below shows the principal results. Illustration from Chen Y, et al. Alzheimer’s & Dementia, 2023. The senior author is Dr. Katherine Rankin. Each MRI image has been reconstructed by computer from routine scans to show the cerebellum splayed out flat.  The randomly assigned colored areas represent a loss of gray matter relative to non-demented people of similar age (“Controls”).  Note that the pattern for PSP differs in obvious ways from the other diseases, though at present the differences are only between the averages for groups, not individual differences useful for diagnosis in routine care. 

Notes: The small type abbreviations are the sub-areas of the cerebellum.  AD=Alzheimer’s disease; CBD=corticobasal degeneration; LBD=Lewy body dementia; TDP=frontotemporal dementia with TDP-43 protein aggregation.  It comes in 3 types. “Pick’s” is a form of frontotemporal dementia.  LBD is combined with AD because at autopsy, the former is always accompanied by some of the latter.  This paper did not include Parkinson’s disease or multiple system atrophy, as those diseases rarely include dementia early in the course, the focus of the present study.

The authors conclude, “These findings suggest the potential for cerebellar neuroimaging as a non-invasive biomarker for differential diagnosis and monitoring.”  They hasten to add that to understand the reasons for these different patterns of cerebellar loss, future studies will have to image the areas of the cerebrum where brain cell activity has been lost and to correlate that with corresponding loss of activity in the cerebellum.  That’s called “functional neuroimaging” as opposed to the “structural neuroimaging” of the current study.   

These insights, aside from their qualitative and quantitative diagnostic value, could provide guidance for electrical or magnetic transcranial stimulation (i.e., delivered across the scalp and skull rather than by inserting hardware onto or into the brain) as symptomatic treatment for PSP and the other dementing disorders. 

Flashes and rumbles

Posted on by

One of this blog’s more frequent and thoughtful readers/commenters, “Mauraelisabeth3,” has asked a good question about the possibility auditory and visual sensory gamma-frequency stimulation as a treatment for PSP.  I responded by promising a blog post on the subject, and here it is:

As always, some scientific background first:  An electroencephalogram (EEG) is a recording of electrical waves emanating from the surface of the brain, as measured by wires pasted to the scalp.  The various brain waves are classified by their frequencies.  The ones relevant to ordinary patient care range from the slowest (i.e., lowest frequency), called “delta,” at 1-4 cycles per second (or “Hertz” or “Hz), to “theta” (4-8 Hz), “alpha” (8-13 Hz) and “beta” (13-30 Hz).  Most of the EEG activity in a healthy, relaxed but awake adult with eyes closed is alpha, and with increasing alertness or with eye opening, there’s more beta.  Theta and delta are important in normal sleep and in many kinds of brain diseases. 

But there’s a higher frequency called “gamma,” which can be found mainly in deeper areas of the brain not usually detected by routine EEG, or if it is detected, it’s hard to distinguish from artifact caused by scalp muscle activity.  It turns out that in people with Alzheimer’s disease, there’s a reduction in gamma activity in the areas deep in the brain that are the headquarters of the memory problem. 

Now to the matter at hand:  There’s a way to “entrain” the EEG activity of those memory-related areas to increase their gamma activity.  An AD mouse model called 5XFAD (with 5 mutations in 2 genes relevant to AD: amyloid precursor protein and presenilin 1) improves in multiple ways after such stimulation, including reducing its load of beta-amyloid, the main component of the amyloid plaques of AD.  A mouse with a mutated form of the tau protein shows improvement in some measures as well, and tau, of course, is the protein central to PSP. Here’s a technical review article on the topic, but it’s from 2018.

A company called Cognito Therapeutics, based in Cambridge, MA, was started by scientists at MIT who have performed much of the early lab work.  So far, the company has sponsored one small, controlled trial showing some improvement in people with AD, but it’s not published other than in very cursory form on the company’s website.  A year ago, in December 2022, Cognito started a Phase 3 trial in AD, meaning a large trial of the sort that, if successful, could win FDA approval for the device for AD.  It’s scheduled to conclude in 2025.  For one hour a day, participants wear glasses flashing a light at 40 Hz (the most relevant point in the gamma range) and headphones playing a tone at the same frequency.  A 40 Hz flash is just barely perceptible as flashing (50 Hz is the standard “fusion frequency”) and a 40 Hz tone is a low rumble.

Some caveats about the treatment: 

  • The most recent literature I found is far from unanimous on whether AD consistently has reduced gamma activity in relevant brain regions, and I found no evidence at all that PSP does. 
  • Although the small clinical trial to date found no adverse effects, there is evidence from the mouse experiments that the gamma stimulation increases the activity of the microglia, the brain’s main inflammatory cells.  That could be a good thing if it enhances the scavenging of unwanted, aggregation-prone protein. But it could be a bad thing if it aggravates the inflammation thought to comprise an important part of the pathogenetic process in many neurodegenerative diseases, including AD and PSP.
  • In the few clinical trials to date, an hour’s stimulation provides only about one day of measurable benefit.  Such a regimen might prove impractical in the real world.

Bottom line:  Would I recommend volunteering for a trial of 40-Hz sensory stimulation in PSP . . .

  • . . . if some more lab data in mice, or very early phase human data supported the benefit and safety of such a treatment in PSP, and
  • . . . if such a trial fully communicated the scientific uncertainties and safety concerns? 

Yes, I probably would.

Keep in mind that devices producing 40-Hz light and/or sound are already commercially available as meditation aids.  No clue here if they help, harm or neither, but until I know more, I’ll categorize them along with all the other placebos out there.

You gotta know when to fold ’em

Posted on by

The last few posts have been about things at the macro level, from clinical trials to government action.  Now, let’s dive back into some molecular biology — if you’re nerd enough for it.

Yesterday, a paper appeared from researchers at the University of Alberta, in Canada, led by Drs. Kerry T. Sun and Sue-Ann Mok, comparing the folding structure of normal and abnormal versions of the tau protein. 

First, some background.  You all know that proteins are strings of amino acids. The healthy adult human brain has six forms of the tau protein ranging in size from 352 to 441 amino acids.  Tau’s normal job is to maintain brain cells’ internal structure and some other housekeeping tasks.  Tau unattached to something else normally flops around in the cell’s fluid like a piece of overcooked spaghetti in boiling water.  In PSP and the other tau-related disorders, tau becomes abnormally folded onto itself and forms toxic clusters that eventually clump further into neurofibrillary tangles.  Those are visible through a microscope and are critical in the diagnosis of the “tauopathies” although the details of how misfolded or aggregated tau actually causes loss of brain cells remain unknown.

Some more background: Although over 99% of people with PSP have no mutations in the tau gene, there are 50 different mutations in tau that do cause neurodegenerative diseases, many of which closely resemble PSP.  The most widely used experimental animal model for PSP has received a copy of a human tau gene with one of these 50 mutations. 

The new project analyzed the folding structure of normal tau protein and samples of abnormal tau protein, each with one of the 37 most important tauopathy-causing mutations.  It found that, at least as far as this lab technique could determine, no structural difference between normal tau and two of the most popular abnormal versions of tau used in research, the P301S mutation (where the amino acid proline at position 301 is replaced by the amino acid serine) and the R406W (arginine to tryptophan).  Another mutation commonly used in animal models, P301L (proline to leucine) does alter the structure.  That’s the form of tau addressed by the two monoclonal antibodies that AbbVie and Biogen, respectively, recently found did not help PSP. 

Of the other 34 mutations tested, 12 produced no structural change and the location of the mutation had no discernible effect on the folding structure.  Nor did the rate of aggregation influence the resulting structure. 

Interestingly, one of those 12 producing detectable misfolding is the A152T (alanine to threonine) mutation, which is the only single-amino-acid substitution tau mutation we know of that increases the risk of “sporadic” (i.e., non-familial) PSP.   

There are some caveats:

  • This study does not examine the effects of post-translational modifications (PTMs) on the folding structure of tau.  Nor did it study the effects of the various mutations on the ability to accept PTMs.  PTM’s are small molecules such as phosphate, acetate, methyl groups, sugars, and ubiquitin that can be attached to the protein in health to regulate its function, or as an effect of disease processes like PSP. 
  • The study restricted itself to only one of the six adult human tau isoforms, called 0N4R.
  • The 0N4R form of tau has 383 amino acids (the others range from 352 to 441) and locations that can alter the folding pattern occur in only about 45 of those.  So, as you’d guess, an amino acid substitution can change the chemical properties of a protein without changing its folding pattern.  Another major issue is that many of those 45 misfolding spots are hidden inside the folded structure, obscuring them from the researchers’ analysis.

Despite these limitations, we can conclude that the various amino acid substitutions affect the misfolding pattern of tau in different ways.  Any explanation of the cause of ordinary, sporadic PSP at its most profound molecular level can be guided by studying all of those misfolding patterns for hereditary PSP but will also have to take account of whatever bad thing the A152T mutation is doing – and that thing, according to this paper, is NOT to directly cause tau to misfold.