The last straw

What could Theranos not do?

Polina Loseva, Lidia Georgieva, N+1

In the United States, Elizabeth Holmes is being tried — a woman who created a startup at the age of 19, attracted millions of investments in it, and at 34 was behind bars. Theranos promised investors a new generation of diagnostics: a compact device that performs hundreds of tests on a single drop of blood. But he did not fulfill his promises. Does this mean that such a diagnosis is impossible in principle — or will someone have time to make a revolution in laboratory tests during the years that Holmes will spend in prison?

Everyone who recalled the meeting with Elizabeth Holmes — in conversations with journalists or testimony in court — agreed that she always sounded very convincing. Well-delivered speech, the habit of not blinking, looking into the eyes of the interlocutor. And a voice. Low, hoarse, almost baritone — you don't expect to hear this from a girl. However, evil tongues later said that Holmes allegedly specially trained him to make a "manly" impression — is it easy, they say, for a woman to make a career in Silicon Valley.

In 2003, when she registered her startup, she had no famous scientists or rich patrons in her team. Just the general idea of a "smart patch" and some money: the remnants of a study loan that became available when Holmes dropped out of the Stanford School of Chemical Engineering after her freshman year. The patch, which the girl decided to create instead of studying chemistry, had to take a drop of blood with a microneedle, analyze it, send the result to the doctor and then calculate how much medicine to inject to the patient.

Scientists and patrons of Holmes and her patch appeared quite quickly. At first, they were acquaintances from Stanford and influential friends of her father. This was enough to start creating a prototype of a miracle device. After some thought, Holmes refused the patch and replaced it with a project of a miniature machine the size of a pocket blood glucose meter. The startup was named "Teranos": from "therapy" and "diagnosis".

11 years later, five hundred people were already working for Holmes. Influential figures like former US Secretaries of State George Schultz and Henry Kissinger have appeared on the board of directors. Hundreds of millions of dollars were invested in Teranos, and the capitalization reached 9 billion — which made it one of the most expensive startups in Silicon Valley.

At the same time, Teranos has not published a single scientific article about its technology for more than ten years — investors and customers trusted the authority of the people on the board of directors and the voice of Holmes, who by that time had replaced baggy suits with a black turtleneck in the style of Steve Jobs. And while she was signing contracts for the supply of Teranos devices, the company's employees did not sleep at night, trying to make these devices work. Those who tried to draw the attention of the authorities to this discrepancy were fired.

By 2015, among the dismissed and dissatisfied, there were those who were ready to talk without fear of litigation and compete with the company. Former employees of Teranos contacted journalist John Carreira and said that the lack of scientific publications is no coincidence: the company really has big problems with technology, and its analyses are often wrong.

In 2015, Carreiro wrote the first article about it. By this time, the startup had already begun to be checked by the FDA, but the investigation caused a wave of new investigations and a big scandal. In 2018, the company closed, Carreiro wrote a book about it, and Holmes became a defendant in a fraud case — together with Ramesh Balwani, her ex-boyfriend and executive director of Teranos. In January 2022, the court found Holmes guilty, and in July Balwani. Since then, they have been awaiting sentencing: both face up to 20 years in prison.

And while the judges are thinking about how deliberately Elizabeth Holmes misled investors and clients, work on the miracle device does not stop. It's just that other companies and startups are already doing this. A compact analyzer that measures blood parameters one drop at a time, despite the failure of "Teranos", is still very much in demand.

Will there be enough drops?

"I'm definitely afraid of needles. This is the only thing that really scares me," Elizabeth Holmes said in an interview with the magazine Wired. And she added that 3-4 percent of people suffer from fear of blood and injections. Many people, Holmes said, do not go for tests after receiving an appointment from a doctor — because they do not want to donate blood from a vein. And for them, she continued, Teranos could offer a convenient solution.

It should have looked like this: a slight prick in the finger, a droplet appears on the skin, and a special cartridge draws it into itself. Inside it are two "nanotainers" — small test tubes with a length of only 1.3 centimeters, which could hardly be seen on the advertising posters of "Teranos". Then the laboratory assistant loads a cartridge with these test tubes into the device, and he himself measures all the necessary parameters in the blood.

But doctors in polyclinics prefer to take blood from a vein from patients for a reason. A finger puncture is a small, but an injury. This means that tissue fluid or blood micro-clots, which normally should not be in it, can get into the sample. This does not happen when blood is taken from a vein: there, the needle penetrates directly into the bloodstream.

But even when blood is taken from a finger, doctors are not limited to one drop. Firstly, because "drop" is an inaccurate concept. Depending on how and from whom it was taken, it can be of different sizes: from 15 to 150 microliters. Secondly, if there are few substances or cells in the blood, then there is a high probability that they simply will not fall into a randomly taken drop. Or they will — but not in the concentration in which they float in the rest of the bloodstream. For example, a 2015 study showed that different drops of blood may contain different amounts of hemoglobin, as well as platelets and leukocytes. The amount of DNA also varies: from 30 to 40 micrograms per milliliter of blood, depending on the donor.

That's how different the concentration of hemoglobin can be in six different drops of blood taken from one sample of venous blood (one color corresponds to one donor). Meaghan M. Bond et al. / American Journal of Clinical Pathology, 2015.

So clinical laboratories usually work not with microliters, but with milliliters of blood — as a rule, at least ten. And the current analytical instruments are designed for such volumes. If you put a small drop in them, they simply do not have enough sensitivity to recognize a signal, for example, a glow from antibodies that have bound some substance.

In order to work with individual drops, Teranos needed to build its own, more sensitive device. But by the time the company started selling its services, it actually failed to cope with this task, and there was not enough drop for analysis. But instead of taking more blood from patients, this drop was diluted with saline — so much so that the concentration of substances in them often turned out to be below the sensitivity threshold of the analyzer.

Naturally, working with such concentrations, the device "Teranos" was wrong every now and then. The Holmes staff sat at work all night long, trying to get the machine to give stable answers for the next presentation, but in vain. Later, when Teranos entered the market and began conducting commercial analyses, customers repeatedly complained about false results. Someone found out that the concentration of a malignant tumor marker in his blood is off the scale. And some healthy pregnant women were informed by the company that they had a miscarriage.

In most cases, these misunderstandings were resolved by repeated testing already in a normal laboratory — which showed that everything was fine with the health of the users of "Teranos". But reverse scenarios are also possible: when a sick person received a "healthy" analysis and did not go to the doctor. Although none of the clients reported such stories.

And yet some companies are still trying to teach their devices to do with one drop of blood. For example, a startup from Massachusetts-based YourBio Health has returned to Holmes' original idea of a smart patch and created a small plastic device the size of half a fist that is glued to the skin with sticky tape.

There are about 30 microneedles on this strip, each the size of an eyelash. Together they draw from the patient from 100 microliters of blood. This is about twice as much as the average drop, but it is suggested to take it from the shoulder: there are fewer nerve endings there than at the fingertips, so the procedure can be almost painless. This device has already received FDA approval for simple tests, for example, to measure blood sugar levels.

There are also projects that are already being tested in some American hospitals. The i-STAT device only needs two or three drops of blood applied to the cartridge - it measures the content of the main blood ions, glucose and urea. Now it is used in intensive care — to assess the condition of patients in a few minutes and adjust treatment.

But Elizabeth Holmes wanted her device to be in every home.

Is there enough for everyone?

By the end of the third year of the covid pandemic, we all appreciated the convenience of self-testing at home, which previously had to deal mainly with people with diabetes. Doing a covid analysis for yourself is at least quick and easy. It's enough to stick a stick in your nose — and it's immediately clear whether it's worth being quarantined or you can go out in public. There is no need to allocate time, gather strength, go to the doctor and worry.

Besides, home tests are cheap. You don't have to spend money on the work of the nurse who takes the analysis and the clinical laboratory that processes it. Holmes promised that the tests on her device would cost literally a couple of dollars, the most expensive — about twenty. This, she said, would open a window for many people into the world of surveys and checkups. And she cited her uncle as an example: if he had access to cheap tests, he could have noticed the incipient skin cancer in time and lived longer.

The medical analysis market is huge: there are hundreds of millions of potential customers in the USA alone. "Teranos", however, did not go so far as to sell its development to private individuals. But even if his devices were in public places — clinics, pharmacies, grocery stores, fitness centers or just offices - it would already bring Teranos a huge profit. Holmes dreamed of equipping such mini-laboratories with The US Armed Forces to monitor the condition of the wounded right on the battlefield.

But for this, the Teranos device had to become really compact. Holmes insisted that it fit on the desktop or in the closet.

The first prototype was called Theranos 1.0, and it was built on a microcapillary system. A drop of blood had to be placed on a cartridge the size of a bank card ("nanotainers" were invented later), inside which the liquid spread through a branching network of microchannels, and they already led to different compartments with antibodies. The cartridge was already placed in the device, and it measured the glow of each compartment — that is, how many proteins from the sample bound to antibodies.

This is what "Theranos 1.0" looked like. JohnCarreyrou / twitter

The design turned out to be very complex. Different liquids had to pass through the capillaries in turn: blood, saline, and other reagents. These flows had to be separated, so the engineers of Teranos tried to supply an already intricate network of capillaries with valves that would open after a certain time and in a certain order. A fragile system came out, where every time something broke or leaked. In addition, the cartridges themselves cost more than $ 200 apiece.

Therefore, the concept changed, and the engineers of Teranos created a device called Edison. A prototype of a nanotainer has already appeared here — a miniature blood flask that was inserted into the device. There was a robot inside the Edison. He took the blood from the flask, dug it into compartments and mixed it with reagents. And there, inside, there were reactions — it turned out to be a whole medical laboratory in miniature.

Such a design, according to Holmes, was supposed to give out analyses of increased accuracy. The variation in the results, she explained to journalists, depends mainly on the human factor: they did not take the sample to work on time or handled it carelessly. And if you entrust this task to a technician, then a drop of blood should be enough: they say, all the extra milliliters are actually just insurance against the laboratory assistant's mistakes.

The robotic laboratory assistant of "Teranos" at work. TrungTPhan / twitter.

Then, however, it turned out that the robotic laboratory assistant also worked carelessly. In particular, because he did not even have a "specialized education". It was developed by another company and it was intended for applying glue. And in "Teranos" they simply attached a dispenser to it and wrote a program of movements inside the device.

Therefore, there were difficulties with the samples every now and then. The robot missed the cartridge, the pipettes broke, and their contents spilled over the insides of the Edison.

The third generation of Teranos developments was called miniLab. It was even more complicated than the Edison, and combined several analytical instruments in miniature at once — plus robotic manipulators that transferred reagents between them.

But the more elements there are in the system, the more problems there are with it. Mechanical pipettes and pumps eventually loosened and began to lose precious blood. And the measuring instruments, locked in a cramped case, overheated — which also started to fail.

The minilab device. Marilyn B. Nourse et al. / Bioengineering & Translational Medicine, 2017.

Minilab never even reached those with whom Holmes managed to sign a supply contract: the Walgreens pharmacy chain and Safeway supermarkets. And the partnership with the Ministry of Defense did not work out: officials did not trust the technology, about which there are no detailed reports or scientific publications.

But this does not mean that it is impossible to build a small analytical device in principle. Even standard laboratory instruments are often not very large — for example, the popular ADVIA 1800 analyzer from Siemens is the size of a large desk.

You can make it smaller. Abbott's Piccolo Xpress analyzer, the size of a shoe box, can measure the concentration of various enzymes, lipids and salts — a total of 16 different analyses. However, for work he needs not one drop, but two or three.

And the Israeli startup Sight Diagnostics is able to manage literally two drops of blood at all. The toaster-sized device he created can "digitize" blood into color microscopic images. And then these pictures are viewed by a neural network that counts different types of cells in the sample. However, this development is now at the same stage where Teranos is stuck — testing in laboratories.

But none of these devices looks like Elizabeth Holmes' dream. And it's not just that not all of them fit on the table. More importantly, they cannot be used alone. At least, their developers assume that a laboratory assistant will manage them. And taking blood is a nurse, especially in cases when you need more than a drop from your finger. But the main thing is that each of these devices can measure literally several, at best several dozen parameters of blood. And "Teranos" promised that it would be able to conduct more than 200 tests on one "Minilab".

Is it possible to do everything at once?

The device that stands in every home should be universal. One is interested in the glucose level in his blood, another is interested in cholesterol, a third is interested in tumor markers or viral DNA. Now, in order to measure it all at the same time, we donate several blood vials for analysis in different devices, which sometimes stand in different laboratories. But the Teranos device had to be able to do everything.

Although the company's website has been deleted, its archived copy still stores a complete list of analyses that its machine allegedly knows how to do. There are more than 240 items, including the concentration of drugs and drugs in the blood, tumor markers and hormones. All these are tasks for which different technical solutions are needed — because there are different groups of chemicals or cells in this list.

To count blood cells, you need to photograph a blood smear or run cells one by one through a cytometer. The concentration of proteins is measured using antibodies and immunofluorescence analysis. And in order to detect DNA or RNA (this is how viral infections are diagnosed, for example), PCR will be needed — for which an amplifier is needed.

The engineers of Teranos could not fit a photomultiplier, a spectrophotometer, a cytometer and an amplifier into one small container so that they did not interfere with each other's work. Therefore, the management went to the trick, deciding at the beginning to put "Edison" in pharmacies and supermarkets instead of "Minilabs". But since they worked unstable, and besides, not everyone was able to do it, nurses were generally sent to the points of collection of tests, who took blood from people (sometimes even from a vein), and then sent it to the laboratory of "Teranos". It was not like what Holmes described in her speeches and interviews and, in fact, was no different from ordinary medical practice.

And since "Minilab" worked poorly, and "Edison" was able to conduct only chemiluminescent analysis, then the laboratory assistants had to cheat. Those analyses that could not be obtained with the help of Edison were carried out on conventional commercial devices. This, however, was a secret of the company. During the inspections, outside observers were shown not the entire laboratory, but only commercial analyzers — explaining that they are here for comparison and help in the development of their own device. And "Edison" and "Minilabs" were stored in a separate room, where most of the staff did not have access.

Closeness in general was a characteristic feature of "Teranos" — all former employees complained about it, who then communicated with journalists. Elizabeth Holmes was obsessed with preserving trade secrets: in one department they did not know what they were doing in another, only their direct authors were familiar with the developments themselves, and before employees were fired, they were forced to erase all working correspondence and sign non-disclosure agreements. Because of this, even the employees of "Teranos" for a long time did not realize the scale of fraud, and when they guessed, they quickly quit.

But the company hasn't really come up with any new technology. The engineers who created the Minilab were simply miniaturizing existing solutions. Meanwhile, their competitors have achieved the same goals by simply betting on new methods of analysis.

For example, another California company, Maverick, managed to do this. Her device, certified in 2019, works on the basis of optical ring resonators. These are polymer rings coated with antibodies from the inside. When the light gets inside the ring, it starts to "run in a circle" there — as a result, some of its waves resonate with themselves. They overlap each other, their amplitude increases, and intense waves of only a certain frequency come out of the ring. It will depend on which antibody is sitting inside the ring. If an antigen binds to the antibodies, the frequency changes — and the device can notice this.

Optical resonator: Light runs inside the ring (A), and the frequency of the wave changes when additional molecules appear inside the ring (B). Sasi Mudumba et al. / Journal of Immunological Methods, 2017.

In total, there are 128 such rings inside on the surface of a single silicon microchip. A channel passes over the rings, blood flows through it, and the rings catch their antigens in it. After that, you can shine a light on the rings and measure the frequency of the light coming out of them. The more it has changed compared to the empty ring, the more the corresponding substance is in the blood. Thus, it turns out to measure the concentration of 128 markers at a time. This device was often compared to the machines of "Teranos". However, unlike the scandalous startup, its creators have both scientific publications and FDA approval.

Another example is a Finnish startup, Nightingale Health, which has bet on nuclear magnetic resonance (NMR). Its essence is that an external magnetic field causes the nuclei of atoms to change their spin with a certain frequency — and this frequency depends on what other neighboring atoms this nucleus has. Therefore, each functional group of several atoms has its own frequency, and so it is possible to recognize the concentration of small molecules with this functional group. That is, an NMR-based device cannot distinguish between long proteins or nucleic acids, but it can measure the concentration of cholesterol, glucose, individual amino acids and many other small metabolites in the blood, a total of two hundred different substances. And the Finns already have about 300 articles on his methodology.

NMR spectrum of blood. Top: peaks that correspond to individual chemical groups. Bottom: blood molecules, the concentration of which can be restored by the number of characteristic groups. Pasi Soininen et al. / Analyst, 2009.

The third competitor of Teranos, SomaLogic Operating Co., promises to measure up to 7000 proteins in one sample. To do this, they use substances that are somewhat similar to antibodies. These are aptamers — short DNA strands twisted into different complex shapes, each of which recognizes one specific protein. With the help of these aptamers, you can pick out the necessary proteins from the blood. Then separate only those aptamers that have found their protein, wash them from proteins and measure the concentration of the aptamers themselves. And this is already easy to do with the help of hybridization: there is a microchip on which DNA strands complementary to aptamers are anchored, and only those with which the aptamer has contacted glow.

But Elizabeth Holmes probably wouldn't have been satisfied with such successes either. Not all of these devices can do with one drop of blood, and they take up much more space than a shoe box. In addition, none of them knows how to recognize all possible types of biochemical markers: some specialize in proteins, others in small metabolites, and no one promises to search in the blood all together. And it takes a lot of processing time — for example, a Finnish device gives results on 300 samples within 6 weeks.

"Teranos" forever

The "Potemkin" essence of Elizabeth Holmes' startup was not immediately disclosed. Since its devices have not formally entered the market over the years (all analyses were done as part of tests), the startup did not need to license either the device or the analytical method. And to get the right to work with tests, it was enough to present a laboratory — in which there was plenty of licensed medical equipment from other companies.

In 2012, the FDA was interested in a startup when it was trying to sign a contract with the Ministry of Defense. But the laboratory check did not show anything — the inspector was also shown only other people's devices. "Teranos" pretended well that everything was going according to plan. In addition, the company had influential patrons, including in the White House. And the story came to naught.

Therefore, serious checks began only after the Carreiro investigation, in 2015. First of all, the FDA banned the use of a "nanotainer" — as a medical device that "has not been tested and approved." Now "Teranos" could not carry out analyses on its devices. Then the FDA banned Holmes from running the lab—even using commercial analyzers.

In 2017, the only scientific publication of "Teranos" dedicated to the work of "Minilab" was published. She confirmed that the device works — but with the use of a full-fledged blood sample, and it is also capable of only several types of tests, and does only one at a time. And then lawsuits rained down on the company.

During the year, Teranos paid penalties on its failed contracts. And in 2018, Elizabeth Holmes and her startup partner Ramesh Balwani were charged by Securities and Exchange Commission. In total, they were accused of 12 episodes of fraud. Of these, Holmes was found guilty on four, and Balwani — on all twelve. Each of which promises the leaders of the legendary startup up to 20 years in prison. The announcement of the verdict has been postponed several times and is now expected in late 2022 or early 2023.

No matter how many years the judge assigns to the leaders of Teranos, you can be sure that by the time they are released, the laboratory analysis market will be replenished with new startups and devices. The Covid pandemic has only spurred them on: many companies have received regulatory approval specifically for tests for coronavirus and antibodies to it.

But none of them sets themselves as ambitious goals as Elizabeth Holmes did. The creators of the new devices are well aware that sensitivity, compactness and versatility are like a wolf, goat and cabbage: only two people get into the boat, and they are not always compatible with each other. Therefore, some people assemble a device that will work on a drop of blood, and it even turns out to be compact — but it cannot count many parameters at once. Others are learning to do hundreds of tests at the same time on a drop of blood — but their device will never fit on the table.

Those who started at the same time as Elizabeth Holmes won not because she lost, but because they did not try to achieve everything at once. But they will not be able to discount it: every time they present their products to investors, they have to prove that they are not a new "Teranos".

Portal "Eternal youth" http://vechnayamolodost.ru