Description

Abstract

---

For our project, we are developing a point-of care diagnostic test for Latent Tuberculosis (LTB) , the hidden form of the world's deadliest infectious disease. While a quarter of the population carries LTB, the current diagnostic tests fall short as being expensive , laboratory-dependent and cannot distinguish between latent and active tuberculosis. This leads to late diagnoses, longer treatment courses, and continued transmission.

Infecheck addresses the problem with early detection of LTB. We are developing a cheap, paper-strip diagnostic that uses a simple finger-prick blood sample to detect upregulated miRNAs specific to LTB. We explored two promising methods for exosomal miRNA detection: Toehold Switches and CRISPR-Cas13a . We envision our end-product to be in the form of a lateral flow test and to use microfluidics as the extraction method. The test delivers easy-to-read results without requiring lab-based equipment. Our approach prevents the progression from latent to active TB, therefore reducing patient suffering and TB spread.

We can't cure what we can't find.

Background Information

---

Tuberculosis (TB) is an airborne bacterial disease. It’s contracted when a person inhales Mycobacterium tuberculosis (Mtb) which settles in the lungs to grow. However, the growth may not always be accompanied by symptoms immediately - a state of TB known as latent Tuberculosis (LTB). Without treatment in its inactive state, those with LTB may develop active Tuberculosis (ATB) at any time which is when symptoms begin, and the germs are transmissible. At this stage, the body typically fails to prevent the growth of TB.

Project motivation

---

Breathlessness, sweatiness and a hacking cough. Some of the symptoms Lauren described before going to the doctors and discovering she has Active Tuberculosis (ATB) through a chest X-ray - all whilst pregnant. Her transition from latent to ATB was through her punctured lung accident, though she was under the assumption she was not latent after being tested negative twice. She had to be isolated from all immediate family for 2 weeks and alert close contacts to get tested for TB. That’s when the discrimination began. People didn’t want to be around her, interrogating her as to how she could contract a disease so “dirty”. In the course of 9 months, she took just under 2000 pills and needed to revisit the hospital multiple times despite her labored state. Finally, she gave birth to her son during treatment, and she was clear of TB 7 long months into treatment, a “relief” she needed, although worried for her newborn’s health.

Project motivation purpose

Lauren is not the only one with a survivor story. There are 10 million people like her, who are diagnosed with ATB, where many sufferers live in countries with decreased access to optimal healthcare. The WHO estimates that a quarter of the global population harbours Latent Tuberculosis (LTB), meaning 2 billion people are at risk of experiencing something similar. They further state that TB is currently the deadliest infectious disease, and has been a front-runner for decades. The cycle must be broken. But how? Let’s start from the root of the issue. The first issue is the cuts to TB funding. USAID (US Agency for International Development) is the “third-largest TB research funder” , according to WHO. In 2023, they contributed $41·6 million for TB research , helped 5 million people go through treatment , contributed most of the US government's US$406 million to global tuberculosis efforts and provided half the funding to the ‘’Stop TB’ partnership. However, funding from USAID has ceased as of July 2025, bottle-necking the progress for TB research, diagnostics and treatment. Consequently, the lack of financial aid for TB diagnostics has made current tools inaccessible for low and middle income countries (LMICs). Traditional diagnostic methods include the Tuberculin Skin Test (TST) and the Interferon Gamma Release Assay (IGRA). TST involves injecting a purified protein derivative of Mtb under the skin to cause swelling in ATB patients. The TST requires multiple hospital visits and takes around 72 hours for results. It also cross-reacts with the BCG vaccine, leading to false positives. The IGRA test is a blood test that mixes a patient's blood sample with antigens derived from Mtb. The IGRA is more accurate than TST as the antigens are specific only to Mtb, however that comes at a steep price and specialist training. The gap between the richest and poorest countries has never been wider. 98% of the global TB burden belongs to LMICs, with TB diagnosis and treatment costing approximately 20% of annual household income in those countries. The under-funding of TB simply worsens the severity of the TB burden on these countries, as they are more likely to suffer from adversities such as undernutrition and HIV outbreak - two major risk factors of who TB spreads to - yet it is the 2% who have more immediate accessibility to diagnostics and treatment. Pivotally, both tests cannot distinguish between LTB and ATB and neither test is universal. 2 billion people have LTB, how many of them know?

The solution

---

Infecheck is the next step towards universal accessibility to LTB diagnosis.

We aim to design a test that is specifically targeted towards LTB diagnosis, as we want to eliminate the root before it activates into ATB. When treated early, TB loses the ability to spread, and the patient can bypass the long treatment course and painful symptoms of ATB. Diagnosing LTB and raising awareness of TBs risk factors, is the optimal method in preventing the spread of TB and reducing those transitioning from LTB to ATB, so 2 billion people can "get the help before it becomes active".

Our Overall Plan:

We followed the REASSURED guideline (created by WHO) in making a POC diagnostic test. This ensures our test design was engineered with thought of its implementation into the real world.

Biomarker Choice

We first thought about the best biomarker for LTB and what samples they can be found in:

Final Biomarkers:

Here are our four final biomarkers that we chose to further model and experiment with. Alongside shows the fold change difference between healthy control (HC) and LTB patients (when normalised to the housekeeping gene U6) extracted directly from Cui X et al literature on LTB biomarkers (19):

Detection Method:

After choosing our biomarker and knowing what sample they are found in, we then explored current miRNA detection methods in the market:

Toehold Switch Challenges:

There was no expression observed when testing our designed toehold switches with their respective miRNAs. When analysing the reason for this, we found that some of the toehold regions within the sequences would interact with the reporter gene and hinder translation, but this was not the case for all toeholds we created. To assess the reason for why they did not work we would need to retest the method using a known working toehold switch and redesign our toehold switch but we were short on time. To find more about our troubleshooting with toehold switch , please go to the Engineering Page

We found numerous challenges when designing our initial toehold switch and when searching other literature methods(18):

• difficult to design it for miRNAs
• the hairpin loop of toehold switches normally are leaky and therefore would have a strong background signal
• the test is time-consuming for a signal output on a paper-based strip (2 hours)
• amplification of the miRNA is needed but this method did not include this

One lesson we learnt was a thorough literature review was needed before choosing a suitable miRNA detection method, which we failed to do as we immediately chose toehold switches. As our toehold switches didn't work and it was difficult to design, through an engineering cycle we chose a new miRNA detection method with the aim of picking a detection test that amplifies the miRNA, has high sensitivity, takes less than 30 minutes for a signal and can be applied in a lateral-flow test format.

Table comparison of Detection methods:

Ultimately, we chose CCDR as our final detection method because it can be done in 15 minutes total when applied to a Lateral Flow Test (LFT) format, was tested in a multiplex system before with different miRNAs, has good sensitivity with as low as 0.123 pM for miRNAs and also amplifies the miRNA signal whilst detecting it too. A limitation is a leakiness in design of the hairpin loop, but we can apply our learning from toehold switches to minimise leakiness and dNTPS can be added to reduce this further.

Cyclic Chain Displacement Reaction (CCDR)

Cyclic chain displacement reaction (CCDR) uses miRNA to trigger displacement reactions through hybridising to DNA hairpin loops which releases the miRNA to form a new DNA strand complex. The cycle repeats, thus amplifying the signal.

Firstly, a hairpin structure called DNA1 has biotin modified at the 5' end and is partially made up of the complementary sequence to the miRNA as well as the complementary sequence to a secondary hairpin structure called DNA 2. When the miRNA is present, it will unravel the hairpin loop and hybridize to the complementary sequence for the miRNA. This will trigger the opening of hairpin DNA2, which has FAM modified at the 3' end. Once unraveled, the complementary sequence to DNA 2 within DNA 1 will act as a primer as it hybridises to DNA 2 and forms a DNA1-DNA2 complex. The target miRNA is released by displacement reaction to become an available trigger again thus repeating the cycle so there are continuous DNA1/DNA2 hybridization reactions.

This complex can be detected in a LFT, with the conjugate pad having anti-FAM antibodies conjugated to red-fluorescent proteins and the test line having immobilised streptavidin. When the miRNA is present, DNA1/DNA2 complexes are formed. The FAM on these complexes can bind to anti-FAM conjugated antibodies and the biotin will bind to the streptavidin, thus forming a coloured sandwich at the test line. When the miRNA is not present, the DNA1 and DNA2 hairpin structures remain in an off structure and the control line has anti-mouse antibodies which can bind to the conjugated antibodies which are always present in the sample. This is an example of a direct test.

For our project, we designed DNA hairpin loops to be complimentary to our miRNAs. Three out of four of our miRNAs could be expressed and purified, so altogether we designed 6 hairpin loops (2 for each miRNA):

Implementation

---

We aim to implement our test into field diagnostics, helping those burdened by TB most. Our implementation strategy involves making the extraction process and final product design accessible to people in LMIC. To achieve this we wanted a simple fingerprick to collect the blood sample used because it's easy to use, and we focused on microfluidics for miRNA extraction from exosomes because it's a portable and cheap device that works on small sample volumes. For our product design we chose a low cost paper-strip test using lateral flow technology - a format native to most people. The LFT design can be made specific to LTB by converting it to a multiplex system, detecting multiple upregulated miRNAs associated with LTB. Overall, It's simple and quick, with a colour change readout showing your results

Sample Collection:

In order to collect a blood sample, we decided on the method of skin puncture by a lancet. This is a small medical needle that was used to prick the edge of the finger to draw a blood sample, purchased from AHP Medicals. We chose the use of lancets due to their cost – effective nature, efficient disposal and hygiene maintenance, as well as ease of use. Firstly, the hands of the subject were washed to ensure hygienic testing before finger pricking. The first drop of blood was not used, however, to avoid collecting a contaminated sample. A 1.5ml test tube was used to collect 50µl of the sample, which was subsequently pipetted onto the desired inlet of the microfluidic chip. Capillary force then moved the blood into the microchannels of the chip, enabling further investigation. The lancet was discarded into a sharps bin (56)(57).

Extraction Process:

We looked into a range of options for extracting miRNAs from exosomes.

Microfluidic chips are the novel method of working with small volumes of liquids. They mainly rely on laws of physics to carry out many essential tasks, such as filtration, cell lysis, DNA capture and more. We have decided to try to apply the technology of microfluidic chips to our project, since it presents a unique opportunity of creating a very cheap “lab in a box” style test that would eliminate the need for advanced equipment and trained staff. In this section we will go over 3 main functions that our microfluidic system will perform - plasma separation from whole blood, exosome isolation from plasma and extraction of miRNAs from the exosomes via lysis.

Plasma isolation

In order to isolate the miRNA from the exosome we first need to isolate the exosomes themselves. It is most easily done in plasma. Human blood contains many contaminants that will make the isolation of miRNAs from whole blood too challenging to achieve, so we decided to explore the possibility of plasma separation from the whole blood, rather than direct exosome isolation.

Exosome isolation

Exosome lysis

Final Product:

A lateral flow test (LFT) is a paper-based diagnostic test. A liquid sample with the target analyte is added to the sample pad and flows laterally via capillary action where it is detected by antibodies or nucleic acid probes to produce a visible signal (47). The WHO recommends 80% sensitivity and 97% specificity in a LFT (83). LFT meets the REASSURED guidelines for POCT because their simple and inexpensive design means they can be implemented into low-resource settings. However, the limitations of a LFT are its sensitivity and specificity when compared to lab-based detection methods. There are two types of lateral flow tests:

Our lateral Flow Test is made up of the following components:

Table 2: Description of what happens in the CCDR Lateral Flow Test at each component stage (47)(84)

Our product will be detected in a LFT. The conjugate pad having anti-FAM antibodies conjugated to gold-nanoparticles and the test line having immobilised streptavidin. When the miRNA is present, the DNA1/DNA2 complex is formed. The FAM on these complexes can bind to anti-FAM antibodies which are conjugated to gold nanoparticles and the biotin will bind to the streptavidin, thus forming a coloured sandwich at the test line. When the miRNA is not present, the DNA1 and DNA2 hairpin structures remain in an off structure and the control line has anti-mouse antibodies which can bind to the conjugated antibodies which are always present in the sample. This is an example of a direct test (47).

It’s important for the detection system to be as specific to LTB as possible, so we want multiple miRNAs to be detected because some miRNAs can be dysregulated in other diseases. Zhou P et al made a “triple-line lateral flow strip for the diagnosis of lung cancer” and in the same paper we extracted the methodology for the CCDR into a LFT, hence we would do the same method for our product but use our chosen miRNAs. For a multiplex system, an additional two DNA hairpin loops can be added to the system to detect an alternative miRNA. These DNA hairpin loops can be altered to have different chemicals modified at the 5’ end, for example as we have designed one DNA hairpin loop pair using Biotin and FAM, then for DNA 3 and DNA 4 we can have Digoxin and TAMRA modified at the 5’ end of these hairpin loops respectively. An additional conjugated antibody specific to TAMRA is in the conjugate pad, and a second test line is added which will have anti-digoxin antibodies immobilised. Both conjugated antibodies (anti-FAM and anti-TAMRA) will be from the same species, so on the control line anti-species antibodies can detect both. From the literature, this method takes 10 minutes so is quick and was shown to have similar sensitivity to PCR results. The disadvantage of implementing the test into a multiplex system is the additional resources needed — more antibodies and DNA-hairpin loops and a longer strip — thus the test will become more expensive the more miRNAs being detected (48).