Final Report Summary - MIMIC (Minerals Integrated into Multiple Identifications of Cancer (MIMIC): a multidisciplinary approach for the ultra-sensitive detection of cancer biomarkers)
In recent years, several studies have been emphasising the requirement of utilising ultrasensitive detection systems in clinical diagnosis. For example, it has been claimed that cancer recurrence could be detected in patients that have undergone total prostatectomy when minute concentrations of the molecule prostate specific antigen (PSA) could be detected after the operation. It is crucial to detect cancer recurrence as soon as possible because it is easier to treat the tumour in a very early stage before it spreads to other tissues. Hence, and ultrasensitive detection system is needed in order to detect ultralow concentrations of the cancer biomarker, and therefore, to diagnose the cancer at the earliest stage when it is still curable. Furthermore, ultrasensitive detection systems are required in other situations, for example, when designing low-cost methods to detect virus infection. In particular, it has been proposed that inexpensive immunoassays could be used as an alternative method for the diagnosis of HIV infection in resource-constrained countries. Usually, HIV diagnosis is achieved with costly nucleic acid-based tests. Although this technique affords for ultrahigh sensitivity the costs associated to it might be too high for the screening of large populations of patients. Immunoassays based on detecting virus antigens are much more economical but lack the ultrahigh sensitivity required to compete with the expensive counterparts. In this context, an inexpensive and ultrasensitive immunoassay could be beneficial to allow the diagnosis of HIV-infection in resource-constrained settings.
Ultrasensitivity is not the only issue that one has to face when designing new detection methods for clinical diagnosis. In cancer diagnosis, it is known that some molecules called cancer biomarkers may be found at abnormal levels depending on the type of cancer. However, the presence of these cancer biomarkers may be originated by other factors not related to the presence of a tumour. Furthermore, every cancer is different, and not all cancer biomarkers are found in all cases. Therefore, the detection of a cancer biomarker is not enough in order to diagnose a cancer. This issue can be partially solved when the diagnosis is based on the detection of several biomarkers instead of only one. This increases the specificity of the diagnosis because the probability of finding several cancer biomarkers is much higher when the cancer is present. To achieve this goal efficiently, it is necessary to design high-throughput detection systems able to detect several disease biomarkers simultaneously. Furthermore, these high-throughput detection systems may allow testing samples from different patients simultaneously, which reduces the time and cost related to the diagnosis. This feature can be particularly useful when the available resources for clinical diagnosis are limited, for example in resource-constrained settings.
Finally, high-throughput ultrasensitive detection systems cannot make a great impact in healthcare unless they can be easily implemented in existing clinical settings. Sensors fabricated with sophisticated methods might be too costly for their implementation in healthcare. For example, the manufacturing costs associated to nanosensors fabricated with non-serial techniques are too high to use these detection systems routinely. Sometimes the instrumentation needed for the detection is very expensive or requires highly skilled personnel. For example, it has been reported that cancer cells can be detected by testing their mechanical properties with an atomic force microscope. However, this technique is still not widely used in clinical settings, and therefore it might be difficult to implement it as a routine test for cancer diagnosis. This issue is particularly acute in resource-constrained countries where sophisticated instruments and skilled personnel may not be available.
The objectives for MIMIC were formulated with this context in mind. Three main objectives were outlined. The first objective was to design detection systems capable of achieving ultrahigh sensitivity, that is, capable of detecting tiny concentrations of a target molecule. Such detection systems could be applied for the early detection of particular cancer biomarkers or as alternatives to expensive nucleic acid-based methods. Second, MIMIC should provide detection systems that allow for high throughput measurements. In other words, the method should be able to test several samples simultaneously. This feature could improve the diagnosis when the disease biomarkers are not very specific for a particular disease. Furthermore, it could reduce the costs associated to the test since samples from many patients could be analysed simultaneously. Finally, in order to ensure a real impact in healthcare, MIMIC should develop marketable products that can be easily implemented in existing clinical settings. To this end, the detection systems should be economical and easily performed with existing resources. All these objectives were fulfilled during the duration of MIMIC via different approaches that shared a common view: the utilisation of enzymes as both nanoreactors for crystal growth and labels in enzyme-linked immunoassays.
In a first approach, we focused on the issue of ultrasensitivity, that is, the capability of detecting analytes at ultralow concentrations. This was achieved by designing a new signal generation mechanism for enzyme-linked immunoassays. In this design, the enzyme glucose oxidase generates hydrogen peroxide that reduces silver ions on plasmonic transducers, in our case gold nanostars. The deposition of silver changes the optical properties of gold nanostars, which can be detected with a common optical readout. The key factor in this detection scheme is to control the epitaxial growth of silver on gold by means of the concentration of enzyme on the gold nanostars. It was found that high concentrations of enzyme resulted in elevated levels of hydrogen peroxide, which in turn resulted in the formation of silver nanoparticles not supported on the gold nanostars. However, low concentrations of enzyme resulted in slower kinetics of crystal growth the favoured the epitaxial growth of silver on gold nanostars. The latter scenario had a greater effect on the optical properties of the plasmonic transducers, which resulted in a new phenomenon that we called inverse sensitivity. With inverse sensitivity, the largest signal is generated when the analyte is found at ultralow concentrations, which makes this phenomenon particularly suitable for ultrasensitive measurements. By this approach, the cancer biomarker prostate specific antigen (PSA) was detected at the ultralow concentration of 1 ag/mL, therefore satisfying the requirement of ultrasensitivity initially outlined in MIMIC.
In a second approach, we addressed the three objectives of ultrasensitivity, simultaneous detection of several molecules and generations of a marketable product. To achieve this goal, we altered an existing detection platform that is commercial and widely used for high-throughput measurements. In particular, we designed a new signal generation mechanism for enzyme-linked immunosorbent assays (ELISA). In the new approach, the enzyme label controls the growth of gold nanoparticles with desired state of aggregation. In the absence of the analyte, gold nanoparticles grow as clusters, which results in the generation of blue-coloured solutions. In the presence of the analyte, gold nanoparticles do not grow as aggregates, which results in the generation of red-coloured solutions. With this approach, the cancer biomarker PSA and the HIV-related molecule antigen p24 were detected at the ultralow concentration of 1 ag/mL with the naked eye, therefore satisfying the requirement of ultrasensitivity. ELISA can be performed in multi-well plates that allow for high-throughput analyses. Furthermore, the new approach can be performed on commercial systems by simply changing the enzyme label and adding an additional step during the signal generation phase. Therefore, this approach satisfied the three main objectives outlined at the beginning of MIMIC.
In conclusion, we designed new detection systems for clinical diagnosis that allow for ultrasensitive measurements and high throughput analyses that could be easily commercialised. The key step of these methods is to use the enzyme label of an enzyme-linked immunoassay to control the growth of nanocrystals with particular morphology or state of aggregation. Since the physical properties of nanomaterials are intimately related to their morphology, this approach paves the way for the development of a new family of sensors in which the signal is generated by the growth of nanocrystals with desired morphology, and therefore with desired physical properties.
Ultrasensitivity is not the only issue that one has to face when designing new detection methods for clinical diagnosis. In cancer diagnosis, it is known that some molecules called cancer biomarkers may be found at abnormal levels depending on the type of cancer. However, the presence of these cancer biomarkers may be originated by other factors not related to the presence of a tumour. Furthermore, every cancer is different, and not all cancer biomarkers are found in all cases. Therefore, the detection of a cancer biomarker is not enough in order to diagnose a cancer. This issue can be partially solved when the diagnosis is based on the detection of several biomarkers instead of only one. This increases the specificity of the diagnosis because the probability of finding several cancer biomarkers is much higher when the cancer is present. To achieve this goal efficiently, it is necessary to design high-throughput detection systems able to detect several disease biomarkers simultaneously. Furthermore, these high-throughput detection systems may allow testing samples from different patients simultaneously, which reduces the time and cost related to the diagnosis. This feature can be particularly useful when the available resources for clinical diagnosis are limited, for example in resource-constrained settings.
Finally, high-throughput ultrasensitive detection systems cannot make a great impact in healthcare unless they can be easily implemented in existing clinical settings. Sensors fabricated with sophisticated methods might be too costly for their implementation in healthcare. For example, the manufacturing costs associated to nanosensors fabricated with non-serial techniques are too high to use these detection systems routinely. Sometimes the instrumentation needed for the detection is very expensive or requires highly skilled personnel. For example, it has been reported that cancer cells can be detected by testing their mechanical properties with an atomic force microscope. However, this technique is still not widely used in clinical settings, and therefore it might be difficult to implement it as a routine test for cancer diagnosis. This issue is particularly acute in resource-constrained countries where sophisticated instruments and skilled personnel may not be available.
The objectives for MIMIC were formulated with this context in mind. Three main objectives were outlined. The first objective was to design detection systems capable of achieving ultrahigh sensitivity, that is, capable of detecting tiny concentrations of a target molecule. Such detection systems could be applied for the early detection of particular cancer biomarkers or as alternatives to expensive nucleic acid-based methods. Second, MIMIC should provide detection systems that allow for high throughput measurements. In other words, the method should be able to test several samples simultaneously. This feature could improve the diagnosis when the disease biomarkers are not very specific for a particular disease. Furthermore, it could reduce the costs associated to the test since samples from many patients could be analysed simultaneously. Finally, in order to ensure a real impact in healthcare, MIMIC should develop marketable products that can be easily implemented in existing clinical settings. To this end, the detection systems should be economical and easily performed with existing resources. All these objectives were fulfilled during the duration of MIMIC via different approaches that shared a common view: the utilisation of enzymes as both nanoreactors for crystal growth and labels in enzyme-linked immunoassays.
In a first approach, we focused on the issue of ultrasensitivity, that is, the capability of detecting analytes at ultralow concentrations. This was achieved by designing a new signal generation mechanism for enzyme-linked immunoassays. In this design, the enzyme glucose oxidase generates hydrogen peroxide that reduces silver ions on plasmonic transducers, in our case gold nanostars. The deposition of silver changes the optical properties of gold nanostars, which can be detected with a common optical readout. The key factor in this detection scheme is to control the epitaxial growth of silver on gold by means of the concentration of enzyme on the gold nanostars. It was found that high concentrations of enzyme resulted in elevated levels of hydrogen peroxide, which in turn resulted in the formation of silver nanoparticles not supported on the gold nanostars. However, low concentrations of enzyme resulted in slower kinetics of crystal growth the favoured the epitaxial growth of silver on gold nanostars. The latter scenario had a greater effect on the optical properties of the plasmonic transducers, which resulted in a new phenomenon that we called inverse sensitivity. With inverse sensitivity, the largest signal is generated when the analyte is found at ultralow concentrations, which makes this phenomenon particularly suitable for ultrasensitive measurements. By this approach, the cancer biomarker prostate specific antigen (PSA) was detected at the ultralow concentration of 1 ag/mL, therefore satisfying the requirement of ultrasensitivity initially outlined in MIMIC.
In a second approach, we addressed the three objectives of ultrasensitivity, simultaneous detection of several molecules and generations of a marketable product. To achieve this goal, we altered an existing detection platform that is commercial and widely used for high-throughput measurements. In particular, we designed a new signal generation mechanism for enzyme-linked immunosorbent assays (ELISA). In the new approach, the enzyme label controls the growth of gold nanoparticles with desired state of aggregation. In the absence of the analyte, gold nanoparticles grow as clusters, which results in the generation of blue-coloured solutions. In the presence of the analyte, gold nanoparticles do not grow as aggregates, which results in the generation of red-coloured solutions. With this approach, the cancer biomarker PSA and the HIV-related molecule antigen p24 were detected at the ultralow concentration of 1 ag/mL with the naked eye, therefore satisfying the requirement of ultrasensitivity. ELISA can be performed in multi-well plates that allow for high-throughput analyses. Furthermore, the new approach can be performed on commercial systems by simply changing the enzyme label and adding an additional step during the signal generation phase. Therefore, this approach satisfied the three main objectives outlined at the beginning of MIMIC.
In conclusion, we designed new detection systems for clinical diagnosis that allow for ultrasensitive measurements and high throughput analyses that could be easily commercialised. The key step of these methods is to use the enzyme label of an enzyme-linked immunoassay to control the growth of nanocrystals with particular morphology or state of aggregation. Since the physical properties of nanomaterials are intimately related to their morphology, this approach paves the way for the development of a new family of sensors in which the signal is generated by the growth of nanocrystals with desired morphology, and therefore with desired physical properties.