Feature Article

DESIGN

Lab-on-a-chip systems

Lab-on-a-chip (LOC) refers to a chip that performs different laboratory functions on microscale. These can include:

• injecting a sample such as blood into a microreservoir
• adding labels that bind specifically to the target molecules, for example, cancer cells, disease-indicating proteins and virus molecules
• guiding the fluid through microchannels along a predefined route
• quantifying the labelled target molecules using specific sensors in the detection region
• visualising sensor measurements and read-out on a computer.

LOC systems hold the promise of a compact, easy-to-use and low cost tool for the detection of, for example, disease markers, environmental pollutants, or biological and chemical warfare agents. The use of LOC systems would not only mean a huge step forward for countries with few health care resources, but also for developed countries where medicine is evolving towards individualised diagnostics-based therapy.

The advantages of going magnetic

Super-paramagnetic particles have become a promising tool in LOC platforms for labelling the target molecules and an alternative to commonly used fluorescent labels. The use of magnetic beads means that the target molecule attached to the magnetic bead can be easily transported over the chip surface, for example, from purification to detection unit, using magnetic forces. A second advantage is that detection of the target/magnetic bead complex is performed with magnetic sensors that are known for their high sensitivity. The best known magnetic sensor is the magnetoresistive spinvalve sensor, which is widely used in hard disk read heads. Belgian researchers have created a new method to increase the sensitivity of a magnetoresistive spinvalve-based detection platform. To prove the effectiveness of the platform, the stroke-marker S100ββ was used as target protein.

The difficult to trace S100ββ protein

The diagnosis of stroke is becoming increasingly important: stroke is the third most frequent cause of death in the Western world, after cardiovascular disease and cancer. To be able to discriminate between ischemic and haemorrhagic stroke and apply the correct therapy in a timely way, the marker protein S100ββ is measured. It has also been demonstrated recently that the diagnosis of S100ββ could play an important role in eliminating the need for computed tomography scans following minor head injury. An important hurdle in the measurement of S100ββ is its low concentration in whole blood, that is, 1–100 pg/mL.

Currently, enzyme-linked immuno–sorbent-based assays (ELISA) allow the detection of S100ββ in the concentration range of 50 pg/mL to 3.5 ng/mL, but they require a lot of hands-on work performed in a specialised laboratory. A reliable, easy-to-use LOC system would allow for fast diagnosis at the patient’s side.

Magnetic spinvalve-based detection platform

At the heart of a diagnostic LOC system is the detection area. Recently, Belgian researchers have built a detection platform that is able to specifically detect S100ββ proteins down to 27 pg/mL, while maintaining a broad dynamic detection range of approximately two decades.

Figure 1. (click to enlarge) Schematic diagram showing the design of the LOC detection area. (A) the detection area consists of two rows of 12 sensors; (B, D) each sensor is made of 9 spinvalves that are electrically connected in parallel; (C) the sensing area is covered with gold, on which the magneto-sandwich assay is built up.

The detection platform consists of magnetoresistive spinvalve sensors (Figure 1). Each spinvalve sensor is made up of multiple alternating magnetic and nonmagnetic metal layers of 1 to 10 nm each. The resistance of the spinvalve sensor changes in response to an external magnetic field, for example, the presence of a magnetic bead on the sensing surface, and this can be measured as an output voltage change.

A “sandwich” with S100ββ stroke protein

Figure 2. (click to enlarge) Scheme of the magneto-sandwich assay on the spinvalve sensing area.

For good detection, it is crucial that all target molecules (in this case S100ββ proteins) that pass over the detection region specifically bind to its surface. For this reason, the gold surface is functionalised to achieve correct binding of the target protein to the sensing surface. A so-called magneto-sandwich assay was created. First, a golden sensing surface was coated with a self-assembling monolayer of alkane thiols, onto which the primary antibody was covalently immobilised. After this, a sample with S100ββ proteins was added, which binds to the antibody and thus to the sensor surface. Then, a secondary antibody was added, which binds with the S100ββ proteins to form the link with the magnetic beads that were added in a final stage (Figure 2; SAM=selfassembled monolayer).

Link–cut–move

The functionalised gold surface that binds the target proteins is typically larger than the spinvalve sensor-covered surface. Indeed, the larger the binding surface, the larger the amount of target proteins that are captured and thus the greater the sensitivity of the LOC. However, target proteins that are bound outside the sensor area cannot be detected by the sensors and will, therefore, not contribute to the final signal.

To circumvent this problem and benefit from the highest sensitivities offered by magnetic spinvalve sensors, a link–cut–move methodology was created. The steps involved are as follows:

• using the sandwich assay, S100ββ proteins are specifically bound to the chip surface and other compounds are washed away
• magnetic particle/S100ββ complexes are chemically detached from the surface
• current-carrying spinvalve sensors generate a magnetic stray field with a maximum at the sensor edges.

Figure 3. (click to enlarge) Scheme of the link–cut–move method. (A) the magnetic particles are specifically bound onto the sensor area in a random fashion (blue area on right figure); (B) the attached particles are released from the surface; (C) the particles are moved towards the edge of the spinvalve sensor (green lines in right figure), because of the higher magnetic field at this spot.

Because of this, the magnetic particle/S100ββ complexes will be concentrated towards the sensor edges for detection. In this link–cut–move methodology, spinvalve sensors function as both the alignment and detection elements (Figure 3).

Detection takes a step forward

Figure 4. (click to enlarge) Dose–response curve for the detection of S100ββ before (grey line) and after (black line) alignment. Left and right are microscopic images showing magnetic particles bound to the spinvalve surface in a sandwich assay format.

A dose–response curve for the detection of S100ββ clearly shows the effectiveness of this approach (Figure 4). For a random distribution of particles, the sensor signal approaches zero. Following the link–cut–move scheme, the sensor signal is strongly enhanced to allow the specific detection of S100ββ from 1 ng/mL down to 25 pg/mL. Thus, it is shown that a well thought out magneto-sandwich assay and sensor design allows the accurate measurement of the S100ββ stroke marker. This methodology can provide a universal tool for the highly sensitive and specific detection of biomolecules, which would mean a huge step forward for the medical world.

Els Parton, PhD, is Scientific Editor, Liesbet Lagae, PhD, is Group Leader, Nano-Enabled Systems Group, and Gustaaf Borghs PhD, is a Research Fellow and responsible for the long-term scientific strategy of the IMEC division working on micro- and nanosystems, materials research and packaging, IMEC, Kapeldreef 75, B-3001 Leuven, Belgium, tel. +32 16 28 12 11, e-mail: gustaaf.borghs@imec.be, www.imec.be