Using CNT for Biosensors to detect, diagnose and treat disease

Basic Premise

It has been estimated that there are over one million Americans who are suffering from a serious, disabling, or even life-threatening disease but have been left without a diagnosis.

Now it is possible to use sequencing to unravel the molecular diagnosis of an unknown condition.

Tiny nanosensors have now been developed to detect DNA, RNA, protein, chemicals and autoantibody signals.

Nano Sensors can take on many forms:

  • Embedded sensors in the blood, which talk to one’s smartphone (or, in the case of children, to the parent’s device), may be especially helpful.
  • Or, At Scripps and Caltech, we’ve been collaboratively working on a bloodstream-embedded biosensor that picks up genomic signals, for applications in predicting a heart or an autoimmune attack or the earliest diagnosis of cancer.
  • Or, patients would receive an injection of the nanoparticles and urinate on a paper test strip coated with antibodies that detect the nanoparticles bound to abnormal cancer proteins. In practice, once validated, it would function a lot like a home pregnancy test.
  • Or, the sensors were implanted in the subcutaneous space (under the skin in the flank), mirroring the levels of the proteins found in the blood.
  • Or, Implantable optical nanosensors have been shown to continuously and accurately track glucose and electrolytes such as sodium or potassium.
  • Or, beyond embedding sensors in the bloodstream, wireless optoelectronic chips can be injected or imbedded into the tissue, such as the brain. And a group at MIT has developed a carbon nanotube, implanted below the surface of an animal’s skin, which detected levels of nitric oxide for over a year in order to monitor inflammation.
  • Or, at the University of California, Santa Barbara, an implanted microfluidic-electrochemical sensor has been demonstrated to provide continuous, real-time tracking of drug levels in animals. 58 Another LIB technology is magnetic resonance reflexometry, which uses antibody-coated magnetic particles; it has quantified the biomarkers of a heart attack (via assaying the protein troponin, released from dying heart cells) 59 and the principal adverse effect of the cancer chemotherapeutic agent doxorubicin (which can destroy heart muscle cells).

The data from these sensors can include the mutations, the structural variants (meaning the change in the number of copies of the gene present), the gene expression, the DNA methylation, the proteins (RPPA stands for reverse phase protein array), and clinical data. From the extensive omic profiling, the principal biologic pathway(s) of an individual’s cancer can be defined. This leads to the ability to match up a specific drug that targets a driver mutation or pathway.

Basic Diagnosis

The potential ranges from a point-of-care, simple genotype for detection of a prescription drug interaction for a patient, to rapid sequencing of a pathogen to determine the cause and optimal treatment for an infection, to actual sequencing of a region of the genome. A little mobile device for such purposes has been described as “a decentralized, universal diagnostic tool,” which could easily interface with the cloud for software interpretative apps. Here’s a partial roundup of some of the remarkably diverse lab-on-a-chip assays that have been or soon will be integrated with a smartphone. For blood, it includes glucose, hemoglobin, potassium, cholesterol, kidney function, liver function, thyroid function, brain natriuretic peptide (used to follow heart failure), toxins, and various pathogens (including malaria, tuberculosis, dengue, schistosomiasis, salmonella, and HIV with a capability of following CD4 + and CD8 + T Lymphocytes and Kaposi sarcoma virus). For urine, the list includes a full quantitative analysis, albumin, human chorionic gonadotropin (HCG, for monitoring preeclampsia in high-risk pregnancy), and urinary tract infections. 39 Testing saliva, there is the capability to detect strains of the influenza virus and strep throat.23 Perhaps most surprising is the spectrum of assays that are emerging from breath—lactate, alcohol, heart failure, drugs (cocaine, marijuana, amphetamines), and even some types of cancer.

Such rare genomic variants of increased risk represent sharp signals that are particularly informative for an individual, and may therefore be useful for prevention of a particular condition. But just knowing risk is not enough. We need to know when the condition will strike. And here is where biosensors come into play. If we knew, for example, that

Asthma

A child was at high risk of asthma, it would be ideal to use sensors to pick up incipient airway problems long before the first wheeze or symptoms. There are many conditions where we know there is genomic risk, but we have been clueless about when to intervene to prevent the event. Embedded sensors in the blood, which talk to one’s smartphone (or, in the case of children, to the parent’s device), may be especially helpful.

Diabetes

From genomics, we can identify children who have a high risk of autoimmune (Type 1) diabetes. We also know that it takes approximately five years before a critical proportion of the islet cells of the pancreas are cumulatively destroyed by autoimmune attacks, at which point diabetes manifests. Tiny nanosensors have now been developed to detect DNA, RNA, protein, and autoantibody signals.

Yet we do nothing clinically to understand the basis of a person with the diagnosis of diabetes and essentially try “hit or miss” treatments. With fourteen different drug classes to treat diabetes, a more intelligent GIS approach could be quite informative for effective treatment. There are probably at least as many molecular subtypes of diabetes as there are drug classes used to treat the condition. Besides genomic characterization, the use of a continuous glucose sensor, even for a limited time of days to weeks, would provide granular data on the individual’s glucose regulation. It’s surely not just diabetes that needs a new molecular taxonomy. There have been a number of exceptional omic studies that break down common diseases into discrete molecular subtypes; the list is ever growing and includes asthma, multiple sclerosis, rheumatoid arthritis, and colon and uterine cancer. It’s hard to imagine any common medical diagnosis that is not presently an oversimplified, reductionist umbrella term, unsuitable for an era of medicine, that is GIS ready.

Immune System

What if we had a blood-based sensor that detected immune system activation, and at that moment in time the immune system was down-regulated with a suitable drug? Perhaps the pancreas could be saved. This type of intervention is representative of a number of autoimmune diseases with sporadic attacks, such as multiple sclerosis, rheumatoid arthritis, or lupus.

Heart Attack

For preventing a heart attack, a genomic signal from the cells sloughing off from an artery lining (known as circulating endothelial cells) has indicated the smoldering process that antedates the actual event—formation of a clot in the artery and stoppage of blood flow to the heart muscle. By knowing precisely when an individual is incubating a heart attack, potent anticlotting medications could be given to block this from happening. And we know that in individuals affected by cancer, there is tumor DNA present in their plasma. This could be monitored during the course of therapy and prevent the need for expensive PET or CT scans that carry high radiation risk; however, an embedded biosensor could provide much tighter surveillance with the chance of picking up tumor recurrence, or, someday, even uncover the first signs of tumor long before there is a mass detected by a scan. In a similar vein, the concept of a “molecular stethoscope” is attractive.

Pathogens/Diseases

Infectious Diseases The use of whole genome sequencing along with mapping social networks has been applied to fighting multiple pathogen outbreaks, including Klebsiella pneumonia, methicillin-resistant Staphylococcus aureus, Clostridium difficile, and tuberculosis. For communicable diseases, this has been an extraordinary development to understand origin and transmission

Pathogen sequencing has potential application well beyond determining the origin of an epidemic. Still, today the typical workup of a patient with a serious infection involves taking blood cultures or other body fluids and waiting two days before the culture results come back, and further time to determine sensitivities of the pathogen to antibiotics. Need for time may be the enemy – During this two to three day period of time, the patient is usually given a blitzkrieg of potent, broad-spectrum antibiotics to “cover” all the possible pathogens that might be responsible for the infection. To understand the lifesaving power of sequencing for infectious disease

Engineers at MIT have developed a rapid, low-cost urine test that relies on nanoparticles interacting with tumor proteins for detection.

One important explanation for this recurrence or durable resistance to treatment involves the genetic heterogeneity of the cancer. When different parts of a tumor are sequenced, there are marked differences in the mutations found. This problem gets worse once a cancer metastasizes, as metastatic lesions have different mutations compared with the primary locus.

One company, Foundation Medicine, has initiated a commercial product of limited sequencing of about three hundred genes of the tumor to query the presence of likely driver mutations – Nano sensor at the tip of a probe.

 

Concussion/Head Trauma

Electronics are becoming more common in football player’s helmets, with impact-sensing accelerometers to quantify the extent of head injury. What is lacking on the sidelines are methods to instantly assess chemical and other biomarkers changes to more accurately assess the extent of a head injury. Beyond screening with a subjective assessment, beyond waiting for lab work from an on-field blood draw.

Same goes for baseball and ‘landed’ chin music, a soccer header gone awry, a cyclist whose head hits the pavement. Or, a fall coming down the stairs, an automobile accident or trauma from domestic abuse.

Monitor the Functioning of Any Organ

The most far-reaching component of the molecular stethoscope appears to be cell-free RNA, which can potentially be used to monitor any organ of the body.

Previously that was unthinkable in a healthy person. How could one possibly conceive of doing a brain or liver biopsy in someone as part of a normal checkup?

Using high-throughput sequencing of cell-free RNA in the blood, and sophisticated bioinformatic methods to analyze this data, Stephen Quake and his colleagues at Stanford were able to show it is possible to follow the gene expression from each of the body’s organs from a simple blood sample.

Beyond looking at cell-free DNA, the cell free RNA transcriptome has important potential for detecting medically relevant signals, as recently demonstrated by tracking pregnancy and fetal development or diagnosing Alzheimer’s disease.

Lab-on-a-Chip

The lab-on-a-chip work extends beyond the smartphone itself, as wearable patches with microneedles that get just beneath the skin or electrochemical chips adherent to the skin have been shown capable of assaying chemicals such as lactate in sweat, and the real-time data can be displayed via the smartphone. Similarly, contact lenses that can quantify glucose via tears, reflective of what would be level in the bloodstream, are being evaluated for wireless smartphone transmission and display.

 

With Deep Information Comes the Need for Deep Understanding, Analytics

And that is changing all the time in each of us. This is an ideal case for deep learning to determine what these dynamic genomic signatures mean, to determine what can be done to change the natural history of a disease in the making, and to develop the path for prevention. Furthermore, this molecular stethoscope could potentially be made into an embedded sensor. But whether this promising window to the molecular operations of the body will pan out or wind up like most of the 150,000 biomarkers that go nowhere remains to be seen. We hardly have to be reminded of how complex human biology is, and that comprehending all the interdependent interactions—systems medicine—for each individual will likely prove to be a tough nut to crack.

And that is changing all the time in each of us. This is an ideal case for deep learning to determine what these dynamic genomic signatures mean, to determine what can be done to change the natural history of a disease in the making, and to develop the path for prevention. Furthermore, this molecular stethoscope could potentially be made into an embedded sensor

 

Incorporating Genetics into Medicine – In Motion, At rest

Most common genomic variants, meaning they are present in more than 5 percent of the population, carry only a small risk. In contrast, rare variants, meaning they are found in less than 1 percent of people, are much more apt to be associated with substantial risk. Because whole genome sequencing has only been performed in a limited number of people to date, without diverse phenotypes or ancestry, we have a ways to go to pick up the important rare variants. Such rare genomic variants of increased risk represent sharp signals that are particularly informative for an individual, and may therefore be useful for prevention of a particular condition. But just knowing risk is not enough. We need to know when the condition will strike. And here is where biosensors come into play. If we knew, for example, that a child was at high risk of asthma, it would be ideal to use sensors to pick up incipient airway problems long before the first wheeze or symptoms. There are many conditions where we know there is genomic risk, but we have been clueless about when to intervene to prevent the event.

Misc

From genomics, we can identify children who have a high risk of autoimmune (Type 1) diabetes. We also know that it takes approximately five years before a critical proportion of the islet cells of the pancreas are cumulatively destroyed by autoimmune attacks, at which point diabetes manifests.

Tiny nanosensors have now been developed to detect DNA, RNA, protein, and autoantibody signals. What if we had a blood-based sensor that detected immune system activation, and at that moment in time the immune system was down-regulated with a suitable drug? Perhaps the pancreas could be saved. This type of intervention is representative of a number of autoimmune diseases with sporadic attacks, such as multiple sclerosis, rheumatoid arthritis, or lupus.

For preventing a heart attack, a genomic signal from the cells sloughing off from an artery lining (known as circulating endothelial cells) has indicated the smoldering process that antedates the actual event—formation of a clot in the artery and stoppage of blood flow to the heart muscle. By knowing precisely when an individual is incubating a heart attack, potent anticlotting medications could be given to block this from happening. And we know that in individuals affected by cancer, there is tumor DNA present in their plasma. This could be monitored during the course of therapy and prevent the need for expensive PET or CT scans that carry high radiation risk; however, an embedded biosensor could provide much tighter surveillance with the chance of picking up tumor recurrence, or, someday, even uncover the first signs of tumor long before there is a mass detected by a scan. In a similar vein, the concept of a “molecular stethoscope” is attractive.

Beyond looking at cell-free DNA, the cell-free RNA transcriptome has important potential for detecting medically relevant signals, as recently demonstrated by tracking pregnancy and fetal development or diagnosing Alzheimer’s disease.

 

 

This information has been sourced, reviewed and adapted from Topol, Eric (92015-01-06) “The Patient Will See You Now: The Future of Medicine is in Your Hands.” Basic Books. Kindle Edition.

0 replies

Leave a Reply

Want to join the discussion?
Feel free to contribute!

Leave a Reply

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