Written by TKS Toronto student, Anupra Chandran (email: anupra.chandran@gmail.com)

With 86 billion neurons, the brain is one of the biggest mysteries here on Earth. I say, it’s time to change that.

My mission is to solve brain-related biology problems, so quality of life can go up, suffering can go down, and we can know more about the wacky world of the human brain. And there’s no problem scarier than dementia.

It’s the disease that causes your brain to deteriorate as you age, often times to the point of losing memory, thinking, behavior and the ability to do everyday activities. It affects people above the age of 65, but early-onset can happen as early as 30.

My mind goes crazy every time I imagine what it must be like to have this disease. From forgetting conversations that only happened a few minutes ago, to hallucinations, to losing the ability to talk, it’s no secret that no one should have to experience this.

But so many still do.

Right now, it affects 50 million people worldwide, but that number could TRIPLE by 2050. The total cost of dementia is $818B (larger than the GDP’s of some countries), meaning it’s one of the most expensive diseases out there.

TLDR: This disease needs to bounce. That’s Anupra talk for we need to get rid of this.

There’s one huge barrier (out of many) that’s stopping us from doing so. And it’s that a lot of the time, dementia is undiagnosed until it’s too late! As of 2011, the estimate was around 75% of cases that go undiagnosed in developed countries, where most cases happen (but this range could be anywhere from 50–80%). This number has stayed the same today, though it may have increased as the dementia rates rise and our population ages.

This is hugely because changes to the brain chemistry that cause Alzheimer’s happen years, or even DECADES, before you see the memory loss and other symptoms.

How can we expect to get rid of dementia if we can’t even detect it?

To tackle this barrier, my goal is to create a brain biosensor, eventually one that can continuously monitor any chemical change related to dementia in the brain. Kinda like a Fitbit, but for brain chemistry.

I focused on Alzheimer’s disease, the most common form of dementia (accounting for 60–70% of cases), to come up with a proof of concept. Right now, this project is in a proposal stage, and uses carbon nanotube field effect transistor biosensors (I promise that’s all English, it’ll make sense soon).

But wait, what are we actually monitoring?

Most dementias are caused by protein malfunctions in the brain. These are the changes that happen long before symptoms show.

If you think of your brain like a machine, your proteins are all the gears, wires, and parts that allow the machine to work. If something goes wrong with these proteins, then eventually the machine will break down.

In Alzheimer’s, protein interactions in the brain go south. This includes the protein called APP and other proteins called secretases causing the brain to lose function, by making plaques that block neuron connections. I do a deeper dive into this here, but the 🔑 rundown is:

  1. APP is is cut, or cleaved, by the beta-secretase (BACE1) enzyme.
  2. A gamma-secretase enzyme cleaves whatever fragments are left, creating more fragments called amyloid beta peptide fragments.
  3. The amyloid beta fragments clump together to form fibrils, or *plaques.*These are one of the key proteomic biomarkers of Alzheimer’s. In the “machine” of the brain, this is like a flashing red siren 🚨
  4. These plaques spread across the brain by breaking apart, and then repeating the process above, like a bunch of seeds planting themselves across the brain

https://miro.medium.com/max/888/1*7NZ6r-IY6ekd9nvdHfw5ug.gif

Although plaques are far from the only cause of the disease (mutations/changes in genes like APP, APOE, PSEN1 and PSEN2 increase the risk, and plaques inside the neurons called neurofibrillary tangles are more causes, see my other article mentioned before), they’re a major indicator of the disease.

The plaques destroy the brain, by disrupting the neurons. These are the cells that send electrical information to each other, holding instructions for bodily functions).

They clump together around neurons to disrupt the function/connections, and stimulate microglia (the immune cells of the brain, which act like a defense system against toxic particles). This creates strong immune responses like the brain becoming inflamed, and gliosis (overpopulation of microglial cells), which still doesn’t rid the brain of the plaques.

A brain with plaques is like a machine with a bunch of clogged pipes. In order to get rid of them, we need to know it’s there in the first place.

Right now, we mainly detect these plaques using:

PET scans, the invisible ink of medical imaging 💡

A drug called a tracer (or a slightly radioactive material attached to sugars that the body can take in) is ingested/injected, and collects where the plaques form. Kinda like invisible ink, where the brain is the paper.

Then, a PET scanner can detect where the tracer collects through light particles (photons) emitted by the tracer. This is like when you shine a light on the invisible ink, it can show the message written on the paper.

https://miro.medium.com/max/2400/1*H__UMm4QnU6kBfVdKWhMiQ.jpeg

PET is more specific than MRIs or CT scans, that only detect overall reduction in brain substance, which happen later than plaque formation.

https://miro.medium.com/max/782/1*Sd-4f0R9Q8wNmyBjjCQSew.jpeg

Alzheimer’s brain under PET vs normal brain

Tissue samples, color edition🌈

Biopsies or samples are mainly used in labs to quantify plaques in a sample of brain tissue, blood, urine, or cerebrospinal fluid (CSF) from the brain.

There are many different types of tests specifically for proteins, from the Bradford assay that detects the concentration of proteins in a solution (how much protein is in this substance), to ELISA, which detects the presence of specific proteins.

Blood collection is done through a needle in the vein of the patient’s arm, brain biopsy samples come from a hole in the skull, CSF is usually taken from your lower back area (lumbar spine), and urine is, well, pretty self explanatory.

Most protein assays add a solution to the sample, creating a reaction or response that allows a change to be seen (like if protein from solution A binds to one from solution B, the color changes, or a fluorescent dye binds to the already bound proteins). At a basic level: if the color changes, that signals that plaques are there. Dipsticks that change color in the presence of a protein are also used in clinical settings.

https://miro.medium.com/max/680/1*jk67GmcA48KNigf4W__trA.png

 

But these methods aren’t really cutting it. They’re:

  • Non-continuous: Neither of these systems can monitor the chemical changes overtime- the tracer would have to be continuously administered for PET scans, or a new sample would have to be taken for the biopsies/tests.
  • Side effects: Although blood tests are minimally invasive, and urine tests/PET scans aren’t invasive at all, lumbar puncture (to get CSF samples) and skull puncture to get brain samples can cause swelling, infection, headaches, nausea/vomiting, and more icky problems. The tracer for PET scans is only low-dose radiation, but can still be an allergy risk.
  • Reactive: These tests are only done if the patient is showing signs of Alzheimer’s, or has risk factors from their age or family history. They’re less likely to be done on seemingly healthy people, even if those people could develop the disease down the line. Amyloid testing is not part of the routine. We’re reacting to problems instead of preventing them.

Alzheimer’s is diagnosed based on other factors too, like overall brain atrophy (or a shrink in the matter of the brain, in brain imaging scans like MRI), or even just symptoms like behavior (when it’s often too late to make significant changes). Genetic testing based on mutations or errors in certain genes (like APP, PSEN1, PSEN2 and APOE) can give an estimate of the risk of getting Alzheimer’s. Even here, we’re still not detecting changes in the brain continuously, or at the specific proteomic level.

If you had plaques in your brain right now, it’s likely that you wouldn’t even suspect it. Imagine if you had a machine, but couldn’t tell how much battery it had, or what power setting it was at? You’d be missing really valuable info. I know I probably wouldn’t use that machine.

Yet that’s how we treat our brains every day.

https://miro.medium.com/max/700/1*_pGj3DixmBImxcAl_HWWdA.gif
Me, wondering why we don’t have a solution to this

 

Enter, the FitBit of the brain: nanosensors!

Nanotech has become a bit of a buzzword. But it’s no surprise why- it’s basically magic!

They’re on the nano-scale and are measured in nanometers. One human hair is 100,000 nanometers across!

https://miro.medium.com/max/2160/1*dCKxQwddutC_1PTEgYM2dg.jpeg

Good things come in small packages (my 5’0 self can definitely attest), but even better things come in nano-scale packages. Since nanomaterials are on the scale of atoms and molecules, they don’t follow the same laws of physics as larger materials.

The materials in the sensors in your car probably won’t be able to detect amyloid plaques (unless you have one sick car), but nanosensors can have these abilities. Especially since as the volume of a particle goes down, the surface area goes up, and more surface area = more cool properties like sensitivity to electrochemical changes, ability to bond to other particles through “functionalization”, and more!

Since these sensors are so small, they can pick up chemical changes at a detailed level, and can continuously be “swimming” around the brain (this is called in-vivo, or inside the body), even before Alzheimer’s is suspected.

My proposal is a carbon nanotube field effect transistor biosensor, attached to L-DOPA molecules so the sensor can cross the blood brain barrier, an RNA aptamer that binds to/detects the amyloid plaques, and PEG molecules to reduce toxicity. Let’s break it down (to the nanoscale 🙃)…

CNT-FET’s, the heart (or should I say brain) of the sensor 🧠

Carbon nanotubes (CNT) are small tubes made of carbon atoms. Each one can be 1–3 nanometers in width, but when put together, the sensor size will be around 40 nm across.

https://miro.medium.com/max/1500/1*W40Pm_sv9opF-vlrcWfzTg.jpeg

 

These nanotubes have the ability to pick up really small electrochemical changes. This means that when the sensor interacts with an amyloid plaque, it can turn this chemical information into electrical information, which can be sent to other devices.

A field effect transistor (FET) is used to control the flow of electrical current in devices like cellphones, and more. Carbon nanotubes are integrated into this FET structure.

https://miro.medium.com/max/1156/1*vrGZpRacZJdG5xYqVL1rxg.png

It has a base (in grey), source, drain, and gate (the CNT). Think of the base like a train, the electrons in the atoms like passengers, and the source, drain and gate as a ticket machine.

Here’s how all these parts fit together.

When the transistor is OFF, or no plaque is interacting with it:

  1. The base is made of a silicon wafer. It’s a semiconductor, meaning sometimes it lets electricity or electrons pass through it easily, and sometimes it doesn’t.

Silicon has 4 electrons in it’s outer shell that could be free to move, but instead are stuck. This is because this silicon atom is bonded to the other silicon atoms in the wafer, sharing their electrons.

4 electrons from 1 atom + 4 from another = 8, which is the amount of electrons in a full shell (valence shell). It’s like all the seats on the train are filled, so the passengers (electrons) won’t be moving around

https://miro.medium.com/max/2318/1*WeRF0M_Ue-ZEOuyEzXNCWA.png

  1. But, semiconductors are doped (no, not DOPE. Actually, I guess they’re pretty dope too). This means they’ve been injected with another element.

If it’s p-doped, or injected with an element like Boron with 3 outer-shell electrons, it has some extra “holes” or positive charges. This is like if there are extra seats on the train, so people (the electrons) can move around.

N-doped, or injected with phosphorus (a 5 outer-electron element), means there’s extra electrons moving around. That’s like if passengers got on the bus, realized there weren’t enough seats, and started pushing around to find anywhere they could sit.

Both doped materials are neutral, which means neither of them is positive or negative.

https://miro.medium.com/max/916/1*exe5t5ySxwyXgGEsJ8kP5A.gif

#dope

3. Doped silicon means our semiconductor is in conductor mode, with electrons freely moving, aka electricity. The electrons from the n-doped parts of the wafer on the ends will move to fill the holes of the p-doped parts in the middle. Passengers from train N will see there’s more seats on the P-train, and slide on over.

4. Now that the P side has more electrons (a negative particle), it turns from neutral to negative once all the holes are filled. The negative P-side + the negative electrons flowing repel each other. When the seats on the P-train are filled, the N-train people are out of luck.

5. This creates a “depletion layer”, like closing the door of the train. No more electricity happens, because there’s nothing to remove the depletion layer, or open the door.

Until…

The transistor is ON, and a plaque interacts!

On top of the silicon wafer, are 2 electrodes called a source and a drain, made of chromium (5 nm) and Gold (30 nm). In between that is the gate, made of our carbon nanotubes (bet you were wondering when those would come in)! The gate is on top of a silicon oxide layer, to insulate it from the silicon wafer.

  1. When the carbon nanotube picks up the presence of a plaque, it can harness that as a positive voltage.
  2. This overcomes the negative charge of the depletion layer, and allows the electrons to flow again.
  3. That electrical current flows between the source and the drain
  4. This change in electrical current is a signal that the plaque is found

It’s like if the train reached its destination (plaque-ville), and could open its doors to let the electron passengers out. New ones could also flow on by using a ticket (aka allowing the CNT to interact with the plaque). This on state is also called forward bias.

https://miro.medium.com/max/1024/1*LI5pn9qbc8P2hvFYXioCEQ.gif

**The nanosensor will also have a nano-antenna also made of CNTs,**which will send this signal from in vivo, to an external antenna made of the same material. It’ll be transmitted at the Terahertz frequency, which is not harmful to humans.

https://miro.medium.com/max/1438/1*tqpx2mmiOb8FAFtzAM45RQ.jpeg

A basic picture of a carbon nanotube antenna. This will pick up the terahertz frequency, between infrared and microwaves. Terahertz has been explored in biosensing because of its minimal effects on the body in small amounts.

Along with the sensor, I also proposed how to attach certain molecules on, that will help with the nitty-gritty of the functions.

One nano-sized leap for the biosensor, one giant leap for brain-kind: crossing the blood brain barrier ✈

The sensor can’t reach the brain can’t be reached by travelling through the bloodstream. The blood vessels are separated from the brain by a defense line of endothelial cells and glial cells, aka the Blood Brain Barrier (BBB).

https://miro.medium.com/max/3904/1*LJxJA0ZGp6Fg32d-qd9acw.png

 

The brain is almost on its own little island (I like to call it the Blood-Brain-Bahamas) 🏝

There are gaps between these cells for very small materials to pass through, called tight junctions. But a more reliable way to cross the BBB is to mimic materials that can cross the BBB through being packaged into bubbles called “vesicles” and being carried across. It’s the difference between sneaking on a flight to the Blood Brain Bahamas vs buying the ticket for a flight.

Particles are most likely to cross the BBB if they’re <200 nm, which as of now, my nanosensor is 😎

My CNT-FET is functionalized with, or attached to, L-DOPA particles. These are particles that are able to cross the BBB (or “tickets”). They’re the precursor to dopamine, the molecule in the brain that allows you to feel pleasure, and are used as a Parkinson’s drug.

https://miro.medium.com/max/2400/1*W1OBLeokCiqisDRF0atxHg.png

Structure of L-DOPA

Functionalization happens when the nanoparticles are dispersed in an L-DOPA solution and vigorously stirred. This forms (often covalent) bonds between the nanomaterial and the molecule, which means that the materials are sharing electrons to make a full shell.

This has been done for gold nanoparticles, but for this to work on a differnt material like CNTs, it’s likely that another molecule like a carboxyl will need to be functionalized first (in a similar process, using a solution with carboxyl groups), to allow this bond to happen.

https://miro.medium.com/max/2400/1*cSbz74KzW7IgEdA_wWatOw.png

Structure of a carboxyl group (this one is carboxylic acid)

Another molecule that can be functionalized is PEG, or polyethylene glycol. Functionalization with PEG, or PEGylation, has been proven to reduce the buildup of nanoparticles in the liver and spleen, and increase the solubility of the particles. This is mainly a way to make the nanoparticle more safe, since even though it is small, buildup in cells overtime can cause toxicity.

https://miro.medium.com/max/688/1*H-RcyapMv4w8yYHof1rp9w.png

Structure of PEG

RNA and amyloid beta, the ultimate match 🤝

Once we have the sensor, and it crosses into the brain, it needs to be clear when this sensor comes into contact with the plaque. The plaque needs something to latch on to. That “something” is an RNA aptamer.

RNA is a molecule that acts like a blueprint, corresponding to a sequence that has to be built into a protein.

But in this case, RNA is an aptamer, or a part of a nanosensor that binds to what we’re trying to detect (the plaque).

https://miro.medium.com/max/3200/1*ILUjovTGqpPqlfgBpVJrXQ.png

RNA aptamers are made using SELEX, a process to constantly duplicate and evolve RNA molecules until the best ones are left:

  1. A random “library” of RNA aptamers is chosen
  2. Each aptamer is tested to see if it will bind to the plaque
  3. The ones that bind are kept in the cycle.
  4. Reverse transcription (RT) is applied, which means the RNA sequence is used to build a DNA template. This DNA template is used to make more copies of the successful RNA, like another blueprint. A protein called reverse transcriptase makes this template
  5. The DNA template is now used to make a bunch of new RNAs. A protein called the Taq polymerase builds these new sequences.
  6. This process repeats, until the most effective aptamer remains. It’s basically survival of the fittest, for RNA molecules.

https://miro.medium.com/max/1674/1*7CM5baxqHyBmhirIgP5Siw.png

This aptamer also has to be attached, or functionalized, to the gate of the sensor. This works similar to the functionalization of L-DOPA. In this case, carboxyl groups also have to be functionalized first onto the sensor, and the aptamer can bind to the carboxyl groups.

By having CNT-FET biosensor, functionalized with L-DOPA, an RNA aptamer and PEG, we could have a device to:

  1. Travel around the brain in the CSF safely (PEG), after being injected into the bloodstream after crossing the BBB (L-DOPA)
  2. Come into contact with amyloid plaques (aptamer)
  3. Detect chemical changes (FET)
  4. Transmit those changes in real time (nano-antenna).

In other words, a real time, proactive biosensor to detect key changes related to Alzheimer’s.

The insane thing to me is that most of this can be done with materials that exist right now! I’m super pumped to do some of the earliest work with putting these materials together, and making a proof of concept that actually works. Some things I’m still figuring out are:

  • Most aptamers currently have a tendency to bind to a specific structure of plaques: β-sheet-rich fibrillar amyloid assemblies. How can a more versatile aptamer be made?
  • Logistical details: how long will the sensor stay in-vivo for, how many sensors will be in the brain at a time, what are the exact mechanisms allowing the sensor to reach the BBB from the blood, what will the notification for a sensor-plaque interaction look like, what devices will this send to other than the nano-antenna?
  • If this sensor is picking up information on such a detailed level, how can I filter out any noise or unnecessary info?
  • How can I get this into a lab and start testing, so I can make sure these materials can actually interact in the way I explained it, and don’t cause any major side effects? This was a theoretical proposal based on experiments, but my goal is to make this a reality.

I’m looking forward to a mind-blowing future 🤯

We might have a while to go before conquering dementia once and for all, let alone truly understanding the brain and all its other mysteries too. There’s still a lot to figure out, even within the space of detection.

But my proposal is only a piece of the puzzle. I’m looking forward to the rest of the pieces to come into place too.

Once we have this terrifying disease detected, we can take more measures to treat/cure the disease, or even prevent it altogether! What if we could have a similar nanosensor that could detect any chemical change relating to any disease, and release drugs to treat it too?

There’s so much left to do to get there, so many of those 86 billion neurons left to discover, but I couldn’t be more excited to do it.

🔑 Takeaways:

  • Dementia affects millions of people worldwide, and that number is set to grow immensely. But we can’t cure it unless we can detect it first, and 50–80% of cases go undiagnosed until too late.
  • Current methods of detection are not continuous, and are reactive. You need to go to the doctor with a concern before you’re tested.
  • I propose a nanosensor that can continuously monitor/detect amyloid plaques in the brain (a key protein involved with Alzheimer’s, the most common form of dementia). This sensor is on the scale of atoms and molecules, giving it properties that larger materials don’t have. It’s attached to molecules that can cross the blood brain barrier, bind to the plaques, and reduce toxicity.