Differential Diagnosis of Autism Spectrum Disorder: A Neurologist’s Perspective


– It is my pleasure to introduce
back to the podium again at this conference for, I
dunno, the third or third time, Dr. Elliott Sherr, speaking
on a neurologist’s perspective of the differential diagnosis of autism, and Elliott, as many of you know, is professor of neurology and pediatrics and with the Institute of Human Genetics. Dr. Sherr. (audience applauding) – Okay, let’s see if I can make this work. Okay, those are my disclosures. And so in 29 minutes, I’m
gonna answer the question, what is autism, what causes
autism, what looks like autism, how do we investigate,
how and when can we treat and how do we get to Tahiti
as quickly as possible. So I’m gonna touch on all of these topics and see if I can provide a perspective that’ll hopefully be complementary to Dr. Leventhal’s that’ll follow. So, you know, as somebody
who had the pleasure or not of majoring in philosophy
as an undergraduate, we always talked about
the null hypothesis, so that’s the thing that
you’re trying to disprove, but if you can’t disprove
it, then that might be true, so maybe autism doesn’t exist,
so that’s the first question, and then its mirror image, if
you have a different idea is that autism is a singular
biological entity, and obviously, both of these are straw men to get you to think about the fact that autism is a collection of conditions that have shared clinical features, and one sort of simplistic
way of thinking about this is at the top, you have the
clinical manifestations, and those are supported
or caused by a number of different layers that contribute to that final clinical phenotype that you’re observing in your patient. So one whole group is the
kinds of clinical phenotypes that are revealed by standardized
clinical assessments, so something like an ADOS
or an IQ test shed light on what you’re observing as a clinician. I’m not gonna talk a lot about it, but there’s a developing field
that’s looking at biomarkers, so other surrogates for
the final diagnosis. Obviously, one of the
things that I’m gonna talk zero about but I think is
probably incredibly important is gene by environment
interactions, and then finally, something I’m gonna
talk a little bit about, but you’ve probably heard
a lot about in the news are what we would call the primary causes of some cases of autism, and it turns out that some of the cases of autism have single causes or highly penetrant causes. So if we think about the
overall etiologies of autism, this is sort of my
breakdown of what I think is the likely set of conditions,
and so idiopathic polygenic, so what we would call
a typical autistic kid, that’s gonna make up about 50%, and again, I think that it’s polygenic
but probably also has other roles, other etiologies,
and then I’m gonna talk today more about the top 20%, the ones that we would call syndromic or monogenic, meaning a single gene
that’s highly penetrant that doesn’t present in a syndromic way but just presents clinically
as autism and nothing else. Okay, so as a neurologist, I’m called upon not to make a clinical
diagnosis of autism, although that happens sometimes,
but I’m typically asked to help out to answer the question, is there a single genetic
cause or biochemical cause or brain anatomic cause
that leads to autism and/or other clinical symptoms? And one of the main goals of
this kind of differential, this kind of investigation, is to make it as efficient as possible,
so if I tested every kid with autism with the full
battery of tests at my disposal, I would probably have a
low yield on those tests, and some of those tests would have a yield as low as one percent
or even less than that, and so when it gets to
that level of yield, you begin to wonder
whether it has any utility, but rather, what I’m gonna
show you here in this slide, and this is actually kind
of like a springboard for the rest of the talk, is can you use your clinical skills
to enhance that yield? So can you identify the kids that are much more likely to have single gene or other single biochemical
or other causes, and so what I’ve done here is I’ve tried to outline some of those
things that I’m gonna look at. So one of them that I think
is very, very important is regression, so meaning that
a child clearly had language, clearly had social interactions,
and then lost them. Now, there’s a kind of
autism where people say it’s regression where
they have a few words around a year of age
and then they lose them. I’m leaving that out because I think that that’s probably a
different biological process, but I’m talking about the kid who, at three years of age or
two and a half years of age, who had a couple hundred
words, starts losing them. That’s when you pick up the
phone and call me that day. Another term that we like to
use, but it’s kind of like an overused term, but encephalopathy, so is the kid alert and interactive or does the kid seem kind of zoned out? And if the latter is true,
that can be concerning for a number of different reasons. Obviously, if the child is having seizures or other neurologic manifestations, that should also trigger
this kind of an evaluation. Things like spasticity or actual weakness, not just hypotonia, but true weakness, or if the weakness is asymmetric, so one side hemiparesis, one
side is weaker than the other. Exam findings, things
like dysmorphic features or changes in the head
size should also trigger an evaluation by a neurologist, and then the other one is involvement of other functional
systems, so if the kid has lung malformations or
has a heart condition, those are other things that make you think maybe there’s a metabolic
or a genetic cause that underlies all of those conditions in sort of Occam’s razor
as a unified whole, and then also, even though there can be positive family histories
in idiopathic autism, I think if there’s a
positive family history that that also requires
a genetic evaluation. Okay, so keep all of that in
mind as we move along here, and so now that you have those tools, you can then sort of use that information to decide who’s gonna come
see a child neurologist and if they come to see
me, what am I gonna do? Well, you know, it’s
like basically saying, if you have a carpenter and
he has hammers and nails, what is he gonna do? He’s, you know, gonna use
them, and so as a neurologist, I’m gonna use the tools that
are at my disposal besides, obviously, a careful
examination and history taking, I’m gonna use other tools
to arrive at a diagnosis, and I’ve listed some of these here. A first pass, so if you
look at the guidelines from the American Academy of Neurology or the American Academy of
Pediatrics on evaluating a kid with neurodevelopmental
disorders writ large, not just autism, but all
neurodevelopmental disorders, that includes a brain MRI. It includes some basic genetic testing, which I’ve listed here,
microarray and fragile X. Oftentimes, but not always,
depending on the judgment of the clinician, it involves
biochemical evaluations and can also, depending on the patient, for instance, if there’s
encephalopathy or regression, one of the things that you
should be thinking about are abnormal electrical activity
that may not present as a seizure overtly, but may
actually cause regression. Okay, and all of these, as
you see, lead to referrals of specialists, whether
that be a child neurologist or my colleagues in genetics
who run an exome clinic, and I’ll talk about
exomes in just a moment, or a referral to a metabolic specialist if there are metabolic concerns. Okay, so here are just
three simple examples of brain imaging that, if you found this in a kid that had impaired cognition, you would go down a very different pathway in terms of working that kid up. Now, you still might want to
treat that child’s behavior with behavioral intervention
or with medication, but there are other things that
you would wanna do as well. So the first one here is an example called X-linked adrenoleukodystrophy. This is a rare condition
but it presents in early school age children with
loss of cognitive milestones. That’s sort of the
hallmark of this condition. Behavioral changes, loss
of developmental skills, and that should trigger
an immediate evaluation by a neurologist because if I catch this sort of at the phase that’s shown here with these white areas
in the back of the head, then that child would be eligible for a bone marrow transplantation, which would arrest the
progression of the disease. If you don’t do that, then
the child can progress to basically a very, very
severe case of spastic cerebral palsy with profound
impairment in cognition. This now is being screened
for in the newborn blood spots as of literally just
this week, so we’ll start to pick these kids up early
and we’ll follow them along and then right as they’re
starting to show these symptoms, ’cause to make life complicated, not everyone with a
genetic condition goes on to have the clinical disease,
we will then intervene with a bone marrow transplantation. Here’s another example, so
this is a little bit subtler. Again, these are axial images, so this is the front of
the head, this is the back. This is the right side and
the left side of the brain. You can see this kind
of fuzzy white area here and you had Dr. Joe
Sullivan earlier in the day and I’m sure that he probably talked about focal cortical dysplasias. This is a very, very
common cause of seizures, but its location in the frontal lobe can also have behavioral consequences, and if a child is seizing sub-clinically, meaning you’re not observing
them having physical movements but just maybe spacing out,
removing that lesion is frequently curative, 100% curative, and can cause those
behavioral problems to stop and for developmental
progress to ensue once again, and then this is another example. This is a child that I took care of who had a in utero stroke,
and presented, as I said, with hemiparesis, but he
also presented clinically with features that were
consistent with autism, so even though it was
motivated or caused by this anatomic change, it’s something
that we would wanna know because that’s gonna
change how we work him up. It turns out that he also had a heart condition and so he really needed the attention of lots of specialists. Okay, and then here I’ve listed
just a few different causes, and obviously, given the time constraints, I’m not gonna go through all of them, but I just wanna talk
about the top two here. So the first one is
phenylketonuria, or PKU, and probably all of you
have heard about PKU. It’s a defect in an
enzyme that metabolizes the amino acid phenylalanine,
and in the days prior to newborn screening, these
children would present with seizures, intellectual
disability and autism, as well as pigmentary changes because of the phenylalanine conversion into melanin. But we don’t see that anymore
because, in the United States, all newborns are screened
with a blood spot and that’s sent to a state
lab and in California and most of the rest
of the country, we test for 50 different conditions
with that newborn blood spot, but as you can see, it’s quite variable. There’s only 12 tests in Germany,
two in the United Kingdom and many countries around
the world don’t test for PKU, and so a child with PKU could
still show up in your clinic and could be significantly
helped by being put on diet that’s restricting phenylalanine, and then there’s another deficiency. Again, this is also not common. It’s called arginase deficiency. It’s a disorder of the urea cycle, so the way that the body
metabolizes excess nitrogen, and kids who have this deficiency present initially as normally but then they start to regress when they’re toddlers, and they develop loss of
cognitive and motor skills and they develop spastic
presentation, so again, something that should prompt
a more aggressive evaluation. Okay, so I think that
Dr. Sullivan may have already talked a little bit about this. There’s a new technique
called whole exome sequencing, and what this is, is this
is the latest, greatest tool and what it allows us to do is to sequence most of the genes and most of the letters in most of the genes all at once, so in the old days, if I
had a specific hypothesis about, let’s just say Tay-Sachs disease, I could draw blood from
the patient and send it to a laboratory and they
could sequence the gene, hexosaminidase, that causes Tay-Sachs and see whether the
mutation was there or not, but now, instead of just
doing one gene at a time or if there was five
or 10 genes in a panel, I can do over 20,000 genes
simultaneously, looking for mutations that might account
for my patient’s symptoms, and it turns out that in about
10 to 20% of autism patients as well as other kids with
developmental disabilities, that there is a new change and
the term for that new change is de novo, so de novo
mutations, and so, you know, you have two alleles, one from your mom and one from your dad and
what’s depicted here is that when this chromosome is passed from either the sperm or the egg into the developing embryo,
that this gene that’s green, and so functioning normally, acquires by this little thunderbolt
here, a mutation, so oftentimes, a single
letter of DNA will change and that change alone is
sufficient to cause the symptoms, whether the symptoms are idiopathic autism by clinical appearance
or autism plus epilepsy or autism plus cerebral palsy. All of these combinations can be caused, depending on where it happens,
by literally a single change, a single letter of DNA being mutated. Okay, so there are companies
that do this kind of testing and you know, I don’t wanna
advertise any particular company but if you email me, I can send
you a list of different ones and you can pursue them
yourselves, but I’m gonna give you some reports to highlight
what these kinds of tests can give us in terms of information. Okay, so the first one is a case where we find a known syndrome, so a child presented to my clinic and he had irritability
that was pretty severe. Also had autism and had had a few seizures but not enough to be on
any seizure medication, and we did this genetic
test and we found a mutation in a gene that’s called GRIN2B, so here is the protein
consequence of this change. Here’s the cDNA, or the DNA consequence, and then here is their report
saying that it’s de novo, meaning that it’s not in
either mom or dad’s DNA, but their interpretation,
the laboratory’s, is that it is what’s causing
this child’s symptoms, and what is GRIN2B? Actually, let me just go back for a sec. GRIN2B is a receptor for
glutamate on nerve cells, so many, many millions
of nerve cells respond to the excitatory stimulation of glutamate in the nervous system, and when that is inappropriately regulated, you can get cognitive and behavioral
problems, as well as epilepsy, and what I’m showing you here is a drug that’s called memantine and
it can, in certain cases, block the pathological
activation of these receptors, so a drug that was not
designed for this purpose can actually be used to ameliorate
both the behavior, somewhat, and the seizures that
these children experience, so by having a specific diagnosis, I can target my therapies. Okay, sometimes, when we do these tests, we find something new, and so this is an example of another patient of mine who had a mutation in a gene called DDX3X. Gene names are often
cumbersome and so I apologize. Just the thing to remember is
that it’s on the X chromosome and so it turns out, there was a paper literally that came out like a week or two before I got these results back, so we were right plugged in
with when this was happening, and I took this information to the family and they reached out to other families that have children, and
these are almost all girls, have young girls that
are affected with autism and intellectual disability,
and they went together and formed a foundation
and a support group and this is the front page of that group that’s called DDX3X.org, and this is not the only case where this has happened. This is actually happening in dozens or even maybe now 100 or more
specific genetic syndromes around the country and around the world. Families are aggregating
together and trying to promote either clinical or basic science research on their children’s conditions
to advance our knowledge and to come up with better treatments. So that’s an example where,
you know, you can see already that there’s 95 known cases,
66 people are registered. They’re just getting started, but I think that they’ll have a lot of success. The families running this
are highly motivated. Okay, and then another
thing that you find out is something called incidental findings, and so there are approximately 50 genes that a lot of very, very clinical
savvy geneticists decided should always be on these tests and should always be returned
as reports to patients, and you can opt out as a family. You can check a box saying I don’t want that kind of information,
but if you don’t do that, they’ll report it to you,
and as you can see here, the gene is BRCA1, so BRCA1, one of the leading causes of breast cancer that’s known genetically,
and you can see here that there’s a known pathogenic mutation and that it’s present in my patient. He’s a male and young, so
he’s not at risk right now but there are things that
will have to be done for him, but his mom also was a
carrier of that mutation, and so it turned out that
the mom’s mother had died of breast cancer as a young woman, It was always in the back
of her mind to get tested, but had never gone through it, and this information,
obviously, pushed her to finally getting properly evaluated and so I think that, you
know, in an incidental way, we provided this family with
a lot of valuable information. Okay, so in addition
to some of these genes that I’ve just talked about,
you’ve probably all heard of some of the more
common or syndromic cases that are associated with autism and I’ve list sort of three
of the more well-described. They’ve been around longer
and they’ve been around longer simply because they are syndromic, so they have physical features
or other medical features that allow you to identify them, and so I’ve listed three of them here. Fragile X, obviously, that
mostly occurs in males. They have physical features. They have a long face and large ears, a tall forehead with a pointy chin. As they get older, they
have macroorchidism and many of these individuals have autism and intellectual disability, and Rett syndrome, as
you know, is a condition that shows up mostly in girls, and again, presents with more of a
regressive or progressive picture where they have stunted
growth of their heads and they have a number of
typical behavioral manifestations or physical movements, and then
finally, tuberous sclerosis is another condition that can
be associated with autism, both in cases where you
have severe epilepsy, and in other cases, where
the TS does not result in epilepsy but just
has behavioral changes, and then I’ve listed below
here two different categories. With the kind of broad
based genetics tools that I’ve told you about, there are labs across the country and
actually, one of them here, Matt State and Stephan Sanders in the Department of Psychiatry, have pushed forward this advance where they’ve taken many
thousands of individuals who all have autism and tested to see what kinds of genetic changes they had, and by doing it in such a large number, they could get the kind
of statistical evidence that they needed to be
able to say definitively that those genes or those
genetic changes lead to autism, and I’ve listed some of them here. These are some chromosomal regions, and then here are a list of single genes, and I just wanna highlight that the one that I just told you earlier, DDX3X, is gonna be more common, we
think, than all of these. 16p11.2, again, was not
found until people started using these microarrays,
these broad based tools of looking across a whole genome, and this probably accounts for
a full one percent of autism, so again, an important
condition to better understand, and so that’s what I wanna close with here in the next couple minutes,
is just to tell you that my lab and a few other
labs around the country have recruited many dozen
individual who have this genetic change on chromosome
16 and we’ve started to look at both their behavioral manifestations, which we kind of outline here, as well as some of these biomarkers
that I was telling you about, and so right up here is, these
are very funny looking images but basically, these
are fancy MRIs that are using a tool called
diffusion tensor imaging to map out the strength of connections in different parts of the brain, and if, just to sort of
color code it for you, if it’s in yellow, it means
that there’s a deficiency in the connections, and if it’s in blue, it means that there’s an increase, and it turns out that if
you carry the deletion on this region, so you only
have one copy, not two, that you have too few connections or these connections aren’t strong enough, and that if you have the duplication, you actually have the inverse. You have too many connections,
and so there are other people who have reported similar
findings in idiopathic autism, where over or under-connectivity can be a part of the picture of
autism, but we were able to see that very, very cleanly by using these genetically defined conditions, and there’s one other
thing that we saw here. Again, this is something called an M100, so when you hear sound, it
takes about 100 milliseconds for that sound to go through your ears, up into your brainstem and
into the part of the cortex that perceives sound, and
we showed that individuals who are normal have, as we’d expect, around 100 milliseconds, but children who are deletion carriers have over a 20 to 30 millisecond delay
in that perception of sound, so it’s not as if their
hearing is problematic, but it’s a cortical problem,
and it’s not surprising to me, but it was interesting to discover that those individuals also have
significant language problems and the degree of language
impairment correlates with how severe that delay
is in the perception of sound in the auditory cortex,
and again, my collaborators at CHOP in Philadelphia
who were doing this with us were only able to see in autism
a five millisecond delay, but by using a genetically
homogeneous group of volunteers, they saw this much larger,
robust delay, and now potentially have an angle to
think about treating them. Okay, I think I’m gonna skip that over. Just in closing, so you know,
not only is the field moving forward very rapidly in terms
of coming up with diagnoses. There are lots of people who
are working on novel treatments and I’ve just outlined
a couple of them here that I think are quite
potentially interesting. The two neuropeptides,
oxytocin and vasopressin, are both in clinical trials in the US, looking to see how they
might help kids with autism, and there’s some preliminary
data that’s suggesting that literally insufflating or
administering it intranasally leads to an improvement
in social cognition 45 minutes to an hour
after delivery of the drug, and you’ll be amused to
know that male volunteers also get better social
cognition, but not the women. And then finally, in terms of treatments, the other one that I wanna point out is that there are lots of disorders that are rare individually,
but collectively, they make a large group
and that they can be fixed, potentially, by replacing the enzyme, and there are ways of
getting these enzymes that are quite clever across
the blood-brain barrier into the brain and there are
a number of clinical trials going on right now to address that. Okay, so just in closing, genetics and some of these other
platforms that I talked about, biochemical methods, allow us to lead to earlier and earlier diagnoses
that will hopefully lead to earlier behavioral
interventions and maybe even leading to, you know, precision
medicine driven treatments and that even the things
that I’ve talked about, enzyme replacement and neuropeptide use, is really just the beginning. There’s a scientist across
the bay at UC Berkeley who’s probably on the shortlist
to get the Nobel Prize, Jennifer Doudna who along with
a couple other investigators came up with a way to
literally edit your genome and people are working on that
in cells and in animal models and I think it’s probably not too long before people think about
doing that in patients who have specific genetic conditions, and if you have any questions, I’m welcome to take them later or you’re welcome to contact me. I left my contact information
up there and thanks very much. (audience applauding)

2 thoughts on “Differential Diagnosis of Autism Spectrum Disorder: A Neurologist’s Perspective

  1. I'm glad I found this. Seems that regardless of the genetic changes, the body tries to maintain an equilibrium, and in doing so, the over or under compensations of say neuropeptides, result in the symptomatic manifestations. I'm not a student or doctor, just trying to make sense of my own neurology so I can live life again.

Leave a Reply

Your email address will not be published. Required fields are marked *