A perspective on optical developments in microfluidic platforms for Caenorhabditis elegans research

Abstract

Microfluidics offers unique ways of handling and manipulating microorganisms, which has particularly benefited Caenorhabditis elegans research. Optics plays a major role in these microfluidic platforms, not only as a read-out for the biological systems of interest but also as a vehicle for applying perturbations to biological systems. Here, we describe different areas of research in C. elegans developmental biology and behavior neuroscience enabled by microfluidics combined with the optical components. In particular, we highlight the diversity of optical tools and methods in use and the strategies implemented in microfluidics to make the devices compatible with optical techniques. We also offer some thoughts on future challenges in adapting advancements in optics to microfluidic platforms.

INTRODUCTION

The nematode Caenorhabditis elegans (C. elegans) was selected in the 1960s by scientists including S. Brenner as a promising model organism to study animal development and behavior.^{1, 2} This soil dwelling nematode is well suited for genetic analysis with a rapid life cycle from egg to adult worm (3 days at 20 °C), self-fertilization that gives genetically identical progeny, a relatively simple anatomy (959 somatic nuclei in the adult hermaphrodite) and a small but complex genome (only 20× that of Escherichia coli, but 20 000 genes predicted). Additional key features that make C. elegans attractive include its transparency, which allows high-resolution optical imaging of fluorescent proteins in intact animals, its stereotyped anatomy from individual to individual, and its ease to cultivate in the laboratory. Since the initial work by Brenner et al., a tremendous body of knowledge has been collected on C. elegans including the wiring diagram of the entire neuronal circuitry,^{3, 4} the complete sequencing of its genome, and the understanding of its developmental program at the single-cell level. All these advances combined with the high degree of genetic homology with other model organisms continue to make C. elegans one of the major model organisms for many areas of biology.

Microfluidics, since becoming a tool for C. elegans research about a decade ago,^{5, 6} has offered several intrinsic advantages to solve major bottlenecks in the handling and manipulation of C. elegans.⁷ Microfluidics brings a tremendous improvement in worm handling and manipulation compared to conventional methods that are manual, laborious, and not well controlled. The length scales of microfluidic channels, ranging typically from a few microns to hundreds of microns, match the size of the worm throughout its developmental stages. Over the years, a plethora of microfluidic tools have been developed to perform sophisticated operations on-chip including the manipulation of worms for precise spatial positioning and control of physical, chemical, and biological environment that the animals are exposed to.⁷ A few common on-chip methods include controlling temperature⁸ and delivering chemical stimuli in a spatiotemporal controlled manner.^{9, 10} Furthermore, microfluidic platforms can enable high-throughput experimentation through parallelization and automation, and improve experimental reproducibility by minimizing human intervention.^{11, 12}

For most microfluidic platforms, in addition to the chips themselves, the experiments also require peripheral equipment such as fluid actuators, optical systems (e.g., microscope), computer, camera, and light sources. The optical components can serve a dual role in the system—to illuminate for imaging and also to stimulate the specimen optically. Imaging is crucial in model organism research for tracking the localization and quantifying biological markers of interest (transcriptional reporters or fusion protein reporters for example) and for tracking animal behavior such as locomotion; these phenotypes can also be further used to screen for mutants. Besides imaging, optical tools may also serve to alter the animals. In cell biology, laser-mediated nanosurgery enables a wide range of operations from whole cell ablation to organelle ablations;¹³ at lower intensities, light stimulation can also be used to control neuronal activity¹⁴ or activate signaling molecules¹⁵ in a non-destructive way by using optogenetics. Thus, combining microfluidics and optical components is a crucial step to creating and using innovative and powerful tools in the future.

This review discusses microfluidic advances dedicated to C. elegans research along with the optical tools involved. We present several examples of development in optics to address the needs for future microfluidic platforms.

C. ELEGANS RESEARCH ADDRESSED BY MICROFLUIDICS

Behavior

C. elegans has been used as a model system to answer fundamental biological questions such as how genes and environmental conditions impact animals' development and behavior. Traditional assays manually track behaviors such as locomotion, pharyngeal pumping, mating, feeding, gathering, or egg laying.¹⁶C. elegans behavior is also used in toxicology and drug discovery programs for metabolic and degenerative diseases wherein worms are exposed to different drugs and their effects are subsequently analyzed.¹⁷ Most assays are labor-intensive. For all these behavioral studies, microfluidics offers unique opportunities to modulate the environments of the worms in a spatio-temporally specific manner. Using laminar flow and diffusion-based mass transfer, typical at the microscale, Gray et al.⁵ established an oxygen gradient in a chamber and tracked worm positions. This controlled oxygen environment enabled the determination of a favorable oxygen level and probing of signaling pathways in oxygen sensation in worms. Similarly, Zhang et al.⁶ designed a central arena with several radial corridors at the end of which different odor-generating bacteria reside. Presenting odors from pathogenic and non-pathogenic bacteria allowed them to study the complex phenomenon of olfactory learning. To quantitatively understand chemosensory behaviors, Albrecht and Bargmann developed a system capable of delivering precise spatiotemporal stimulus patterns.¹⁰ In these studies, the optical read-out of the devices consists of tracking worms' positions on-chip via conventional low magnification microscopes, and therefore the requirement on the devices is only that they are optically transparent.

In general, imaging in microfluidic behavioral assays requires a broadband illumination source and a relatively large field of view. Broadband white sources are commonly used since the extracted data consists mostly of analyzing the worm movements. The field of view ranges from imaging a section of the worm body to imaging the full chip for monitoring a whole population. Dissecting scopes or zoom lenses directly adapted on cameras are excellent tools for behavior experiments. The low magnification range fits the needs for imaging large areas; moreover the upright configuration and long working distance make them user-friendly for handling the microfluidic device. Finally, imaging from the top is necessary in cases when the bottom material of the chip is diffusive, for example, in a set-up made out of polydimethylsiloxane (PDMS) on agar plates.^{5, 6}

While microfluidics offers an opportunity to perform high-throughput experiments, the large number of images generated can quickly become overwhelming. Image processing tools enable automation and reduce human bias; they are vital to handle the data collected in many experiments. For instance, Wormtracker video analysis tools have been developed to automatically track the position of a population of worms on a plate.¹⁸ Wormtracker analysis provides not only information on worm position but also many additional locomotion parameters such as velocity, body curvature, stop duration, or number of pirouettes. Most of these machine-vision tools can be used on worms crawling on an agar surface or in microfluidic devices as exemplified by Albrecht and Bargmann¹⁰ to monitor chemotaxis behaviors. In this set of experiments, the chip both mimicked the natural environment and offered the advantages of microfluidics in controlling odor gradients (Fig. 1a). Currently, analyzing individual traces among a population of worms still remains challenging as it is difficult to track individuals when animals run into each other. Some effort has been made in partitioning the worms into valve-based chambers¹⁹ or droplets^{20, 21, 22} that prevents losing track of the identity while keeping the ability of monitoring behavior of a large number of worms simultaneously, but this is an on-going research area because of the practical difficulty in keeping the microenvironment consistent for long periods of time.

Figure 1.

C. elegans research addressed by microfluidics and optical modalities. (a) Microfluidic platforms allow for monitoring worm behavior while controlling its environment. Bright field illumination is used with low magnification (from imaging a section of the worm body to imaging the full chip for monitoring a whole population). Left: microfluidic arena where a worm swims freely while being exposed to specific chemical stimuli. Right: the trace of a worm in time is automatically determined using image processing. Reprinted with permission from D. R. Albrecht and C. I. Bargmann, Nat. Methods 8, 599 (2011). Copyright 2011 by Macmillan Publishers, Ltd.¹⁰ (b) Neuronal functional imaging and optical manipulation are used for deciphering worm neuronal circuits. An illumination source is required for exciting the fluorophores and opsins. Selective illumination may be used to target specific cells. Top: a worm is trapped in a channel and its body curvature can be adjusted using a valve or via optogenetic manipulation. Reprinted with permission from Wen et al., Neuron 76, 750 (2012). Copyright 2012 by Elsevier.³⁹ Bottom: neuronal activity of AVA interneuron is recorded as the worm transits from forward motion to backward motion. Reprinted with permission from Chronis et al., Nat. Methods 4, 727 (2007). Copyright 2007 by Macmillan Publishers, Ltd.⁹ (c) Laser-based surgery enables axotomy and cell ablation. The image illustrates a typical set-up where the worm is immobilized on chip during surgery. Reprinted with permission from Allen et al., J. Neurosci. Methods 173, 20 (2008). Copyright 2008 by Elsevier.¹⁴ (d) Microfluidic platforms for genetic screening allow for reducing screening time via automation of worm handling. Fluorescence imaging at high magnification is generally used to discriminate fine phenotypes. Left: screening platform for trapping worms in a specific dorso-ventral orientation. Center and right: fluorescence images showing fine neuronal pattern. Reprinted with permission from Caceres et al., Plos One 7, e35037 (2012).⁴⁹ (e) On-chip long-term culture allows for high-resolution imaging over long-period of time. Top: schematics of culture chambers. Control channels regulate food supply and Pluronic supply, a reversible thermosensitive polymer used for repetitive immobilization. Bottom: L2 stage animals loaded in chambers. Reprinted with permission from J. Krajniak and H. Lu, Lab Chip 10, 1862 (2010). Copyright 2010 by The Royal Society of Chemistry.⁵⁸

Neural functional imaging and optical manipulation

Besides gross behavior, because C. elegans has a small nervous system, it is an attractive experimental model to decipher the genetic basis and the neuronal origin of behavior in response to environmental input.^{4, 23, 24} To do so requires methods for monitoring, exciting, and/or inhibiting specific neurons in a non-invasive manner. Here, we discuss several optical methods for recording and/or optically modulating neuronal activities via genetically encoded reagents.

Monitoring calcium transients into targeted neurons is usually accomplished with genetically encoded calcium indicators (GECI) because it can be done in a cell-specific manner and generally not invasive.²⁵ This method is well adapted for in vivo recording of neuronal activity, particularly in live and behaving animals. This technique requires maintaining the targeted neuron in focus, which is non-trivial in physiologically active animals. Microfluidics provides a restriction method wherein the worm partially confined in a small channel has restricted space in the z focus plane (Fig. 1b).⁹ This approach prevents potential physiological perturbations that can occur while using cooling or drugs for complete worm immobilization. Chronis et al.⁹ also demonstrated the ability to deliver the odors with precise temporal control while monitoring the neuronal activity, which is difficult with conventional imaging techniques. Upgraded versions of this design can now be used for the automation of the process from worm loading to image acquisition increasing the throughput up to 30 worms/h,²⁶ and the ability of delivering multiple chemical stimuli in specific temporal patterns.²⁷ These chips have had a great impact on calcium imaging experiments particularly in the dissection of the olfactory behavior circuitry.^{28, 29, 30, 31}

In general, as a fluorescence technique, calcium imaging of neurons or muscles in C. elegans is easily implemented in combination with microfluidics. An excitation source for exciting the photo-excitable proteins and proper optical filters are needed. The magnification varies depending on the desired resolution, typically from 20× to 63×. An inverted scope is preferred to upright in general as high-magnification immersion objectives require a small working distance and having connection pins and tubing facing up is more practical.

Besides recording neuronal activities, one could use optical manipulations to excite or inhibit neuron (or muscle) activities, usually via reagents such as channelrhodopsins.^{32, 14, 33, 34} This technique allows neural circuits to be interrogated in vivo, sometimes to correlate with behavior in physiologically active animals.^{35, 36} In C. elegans research, microfluidics gives control over the location of the animals and the delivery of physical and chemical stimuli. In one application, multiple worms were manipulated in parallel channels, enabling an easy and efficient way to carry out optogenetic study about synaptic transmission at the neuromuscular junction;³⁷ the experimental throughput was improved by several orders of magnitude compared to the standard approach.³⁸ In another study Wen et al.³⁹ used selective illumination on a worm trapped at its center and having head and tail free (Fig. 1b). This set-up allows studying worm locomotion in a worm that does not move away from the field of view.

In this type of problems, microfluidics will be able to bring different modalities of natural stimuli to the animals; with the future development of a larger panel of spectrally diverse GECI and opsins and the development of new optical platforms, the types of assays and experiments that can be performed to probe the functions of the nervous system will be greatly expanded. We envision this type of approach will enable the identification of new functional circuits and the understanding of dynamics of neural circuitry.

Axotomy and cell ablation

In neuroscience, short-pulsed laser surgery has been used for precise ablation of cellular and subcellular structures.⁴⁰ The ablation removes a node in the circuit permanently and therefore allows one to interrogate the function of the cell in the circuit. Laser-based surgery also has the precision suitable for ablating subcellular elements without compromising the cellular viability; this is usually accomplished by using ultra-short laser pulses, which targets ultra-small volume with little heating effect. Some major applications include studying the kinetics of neural regeneration or specific synaptic functions by targeting axons or synapses.⁴¹ Microfluidics in these applications offers the ability to immobilize the animals during surgery, and streamlines the high-throughput handling of the animals.⁴² Chung et al. designed a chip to perform neuron ablation on Larvae stage 1 worms that are difficult to handle due to their small size (less than 250 μm long and 15 μm in diameter); they used a cooling system to immobilize the worms during the surgery.¹¹ Allen et al. used a UV ns-pulsed laser to remove single synapses without damaging the axon (Fig. 1c)⁴³ while several groups used infrared fs-pulsed lasers to realize axon ablations.^{44, 45, 46} Because the chips were all designed with UV/near infrared transparent glass coverglass at the bottom, they are optically accessible. In some applications, the increased throughput allows genetic or chemical screening to identify molecular mechanisms or chemicals that promote or alter neurite regeneration.^{47, 46}

Genetic screening

One of the most powerful assets of microfluidics resides in the ability of being automated. For instance, using microfluidics to streamline imaging and perform genetic screens has allowed reducing screening time by several orders of magnitude compared to traditional assays where worms are individually and manually picked.^{8, 48, 49, 50} In a chip, thousands of worms can be successively loaded, imaged, phenotyped, and sorted. Generally, fluorescent reporters are used to discriminate worms of different phenotypes (Fig. 1d); epi-fluorescence microscopy with a high-magnification objective is the typical method as it is sufficient in resolution for most problems and fast (as compared to confocal scans) although the demand on image processing can be greater because of the image quality.

Screening for fine phenotypes based on micron-sized structures typically uses high-resolution imaging (at high magnification), which requires the immobilization of the nematode. To address this, technical solutions using suction,^{51, 52} mechanical restriction by a valve membrane,⁵³ channel constriction,⁵⁴ and CO₂ anesthesia⁵³ have been demonstrated; a common and reliable microfluidic method is to use valves to trap the worm in a region and use in situ temperature control to cool (to ∼4 °C) and immobilize the worm.^{8, 48, 49, 50} Cooling is fast, reversible, and effective. Compared to traditional immobilization methods that rely primarily on the use of anesthetic drugs, cooling-based immobilization is non-invasive, allows for recovering the worms after imaging and limits adverse developmental effects.

In addition to hardware engineering to allow for automation and streamlining of animal handling, software for image analysis is also important for phenotyping or detecting novel mutants.^{8, 12, 49} The first genetic screen of a mutagenized population of C. elegans in a microfluidic device was performed by Crane et al. in 2009.¹² More recently in the automated process, novel mutants were identified based on predefined criteria; using machine-learning techniques, the power of using software algorithms can be pushed one step further where the sorting criteria are determined by the software without human intervention.^{50, 48} Overall, genetic screening in microfluidics has become a mature technology; further development is now focused on enhancing the ability to detect more subtle phenotypes⁴⁸ and discovering new biological mechanisms.

Long-term culture

Conventional culture techniques and biological assays are labor-intensive and not well-suited for monitoring long-term developmental and dynamic processes in worms. This is because these experiments involve tracking single worms over extended periods of time (hours to days) in a controlled assay environment. Recently, a lifespan machine has been developed to automate the scoring of surviving worms.⁵⁵ The technology relies on a customized flatbed scanner that images sixteen 4-in. diameter agar plates. Imaging this extremely large field of view enhances the throughput. Moreover, the use of image processing to identify dead worms by detecting their absence of motion within image stacks releases the necessity for human supervision. One drawback, however, is that the 8-μm spatial resolution and 15-min image sampling limit the analysis to the survival curve of a worm population without other physiological details. In contrast, microfluidics platforms can allow for a higher-content analysis in long-term culture. Using microfluidic techniques, several ways to immobilize and release worms repeatedly have enabled high-resolution imaging over a long period of time. Ma et al.⁵⁶ used a valve-based compression technique and they ran a drug screen of adult worms over a period shorter than 4 h. Hulme et al.⁵⁷ combined chambers and tapered channels to study L4 and older worms over days. Krajniak and Lu⁵⁸ developed another method using a reversible thermo-sensitive Pluronic and by slightly tuning the temperature achieved the sol-gel transition to immobilize worms transiently for imaging (Fig. 1e); in this experiment, L1 stage worms were loaded and maintained in the chambers until adulthood. Future opportunities for developing long-term culture platforms include increasing the throughput of these systems as well as extending the period of observations of C. elegans over the full lifespan, which is challenging considering the need for supplying food and removing eggs.

INTEGRATING OPTICS AND MICROFLUIDICS FOR THE STUDY OF C. ELEGANS—OPPORTUNITIES AND CHALLENGES

Despite the tremendous progress in C. elegans research in the last few decades, many unknowns remain. Some of these questions such as deciphering the nervous system are currently largely bottlenecked by the lack of technologies for making observations and measurements, especially with high precisions and at high throughput. Further integrating microfluidics and advanced optical techniques could overcome some of these bottlenecks and lead to major discoveries. Here, we review several areas that have advanced significantly in the past few years, which can enrich the panel of techniques for studying C. elegans, and discuss some challenges.

Enhancing the resolution

Several super-resolution techniques have been developed in the last few years, which could tremendously enhance the ability to identify subtle phenotypes or study processes at a sub-cellular level not achieved before.^{59, 60} Resolution in conventional optical microscopy is limited by the diffraction limit, which is about 0.2 μm and 1.5 μm for lateral and axial resolutions with the best objective lenses. In fluorescence, the image quality is decreased due to scattered light from out-of-focus planes. Two-photon absorption microscopy or confocal microscopy allow for better axial resolution using non-linear absorption properties of the material or diaphragms to limit the volume imaged. Breaking the diffraction limit is now possible via super-resolution microscopy; features as small as 20 nm have been resolved. These techniques include deterministic methods, e.g., stimulated emission depletion microscopy (STED),⁶¹ and stochastic methods, e.g., photoactivated localization microscopy (PALM)⁶² or stochastic optical reconstruction microscopy (STORM).⁶³

Using these diverse microscopy techniques with C. elegans in microfluidics has challenges to various extents. The main challenges reside in dealing with the motion and the shape of the animals. The smaller the object, the more sensitive to movement; various techniques including anesthetics and cooling may be necessary to immobilize the worms. Choices of these techniques depend on the requirements of the experiment: imaging at a specific time point may tolerate the use of drugs and/or cooling. Some processes may tolerate slight movement that can be corrected by software afterwards; live imaging for a prolonged period may require feeding, therefore is the most challenging, and will require a combination of approaches to allow simultaneous feeding and partial restraint, an example of which is the continuous live imaging platform (CLIP),⁶⁴ but the drawback of such techniques may be the complexity required for the setup and robustness of the technique. The second challenge is optical—the cylindrical shape of the worm as well as the tissues having unmatching refractive indices with water and the PDMS walls of microfluidics induces aberration; potential solutions may include adjustment in channel design, and exploitation of new materials to make the chip. PALM and STORM also require the ability to introduce photoswitchable dyes in specific target cells or sub-cellular units, which may be challenging in C. elegans.⁶⁵ Future research is likely to focus on the development of techniques that are feasible to implement in regular biology labs without special equipment and materials, and techniques that are scalable.

Going 3D

Imaging in 3D is appealing for neuroscience and developmental studies because many problems benefit from the ability to acquire activity information from many cells (or locations) simultaneously. Confocal laser scanning microscopy and two-photon microscopy can be very helpful to probe small volumes but the acquisition time is the limiting factor for larger volumes. Alternative optical sectioning techniques based on planar illumination or structured illumination have emerged that allow high-speed, large field of view, and long-term imaging.⁶⁶ They have been successfully applied to other model organisms such as Drosophila embryos and zebrafish.^{67, 68} In the case of planar illumination, the illumination beam is shaped into a sheet so only a very thin volume is excited (Fig. 2a). The critical parameter in this technique is to maintain the plane of illumination since sample-induced aberrations or scattering broaden the light sheet. This is a major challenge for microfluidics to resolve in order to be suitable with this microscopy technique; finding the right design to allow transversal (i.e., 90°) illumination while preserving the quality of the planar beam. As a proof of feasibility, we adapted a microfluidic chip and imaged a transgenic worm that expresses fluorescence in its ASH sensory neurons (Fig. 2b).

Figure 2.

Optical challenges for microfluidic platforms. (a) Going 3D would allow for controlling or capturing signals from cells in different plane of views. If using planar illumination, chip design should maintain the shape of the light sheet to ensure fine axial resolution. 3D imaging is achieved by scanning the sample with the light sheet. Reprinted with permission from J. Mertz, Nat. Methods 8, 811 (2011). Copyright 2011 by Macmillan Publishers, Ltd.⁶⁶ (b) Preliminary results showing light-sheet microscopy imaging of a transgenic worm on a chip. Top: merging of bright-field and fluorescence images of a sensory neuron. Bottom: calcium imaging during chemical stimulation (data unpublished). (c) Multiple-objective imaging would allow for monitoring simultaneously multiple events. Chip design must be adjusted to allow room for several objectives or for imaging at different angles. Reprinted with permission from M. Weber and J. Huisken, Nat. Methods 9, 656 (2012). Copyright 2012 by Macmillan Publishers, Ltd.⁷³ (d) Integrating optics on chip may benefit from miniaturization, cost-saving and new functionalities. Imaging in optofluidic microscopy is lensless and relies on algorithms for reconstructing the images. Reprinted with permission from Bishara et al., Opt. Express 18, 27499 (2010). Copyright 2010 by The Optical Society.⁷⁸

A very recent method applies a 2-photon technique on a transgenic strain expressing a nuclear-localized genetically encoded calcium indicator.⁶⁹ Using wide-field temporal focusing, Schrödel et al. demonstrate near-simultaneous recording of activity of up to 70% of all head neurons. On-going efforts need to push the method to increase the coverage and identify the neurons. Future work should also see the application of stimuli by taking advantage of microfluidic controls.

Besides imaging, 3D selective illumination is an interesting method with applications in optogenetics, e.g., to excite spots in different z-planes. Using a spatial light modulator, Packer et al. demonstrated the independent or simultaneous excitation of two neurons located 20 μm apart in depth.⁷⁰ This method could be readily implemented for worms as well. A large number of neurons are located in the worm head, close to each other but in different planes. Instead of probing the neurons sequentially, the neurons may be activated simultaneously to study signal integration. Similar to before, when combined with microfluidics, this would open up new opportunities for studying functions of neural circuitry in the context of natural stimuli.

Multiple-objective and/or multiple-angle microscopy

In some instances, studying biological processes such as functional neuronal circuitry may involve tracking targets along the full body of C. elegans. Doing it at high-resolution is challenging considering the worm length of several hundreds of microns up to 1 mm; it requires the use of multiple high-magnification objectives. So far up to two objectives (a low-magnification and a high-magnification) have been mounted simultaneously to follow worms freely crawling on agar plates.^{71, 72} It is already possible using currently available microfluidic platforms and mounting two high-magnification objectives to perform high-resolution imaging in distant regions of interest. Introducing three objectives (Fig. 2c)⁷³ or a degree of freedom on the angle of an objective requires deeper modifications of both microscope mount and microfluidic chip. Conventional microscope body must be arranged to combine the objectives and the illumination sources or to allow for rotating movement. Because high-magnification objectives have short working distances and conventional glass-PDMS technologies allow imaging in a plane perpendicular to the glass slide, high-magnification imaging at large angles of view in glass-PDMS devices is made impossible. The chip must be adapted to allow room for multiple objectives or for ensuring a good image quality while imaging at different angles; one possibility is to explore other technologies like glass capillaries.

Integration of optical and electronic components on-chip

Besides making microfluidics chips compatible with conventional optical set-ups, one way to couple the two functions is to bring the optics to chips to reduce cost and the overall size of the instruments as some have demonstrated in optofluidic devices. Field applications benefit from the size reduction while all applications could potentially benefit from the cost reduction. Optofluidic microscopes developed so far have relied on bringing the sample as close as possible to the detector.^{74, 75, 76, 77, 78} They circumvent the use for an expensive optomechanic structure and expensive optics (as in conventional microscopes), and thus could be much more cost-effective and compact.

The first prototype based on shadow imaging offers a resolution larger than 10 μm.⁷⁴ A more sophisticated technology using submicron multi-apertures between the microchannel and the CMOS chip uses algorithms to rebuild an image of a worm flowing above the mask and increases the resolution up to 0.8 μm.^{75, 76, 77} The last strategy uses coherent inline holography and pixel super-resolution to create high-resolution amplitude and phase images with a comparable resolution (Fig. 2d).⁷⁸ Altogether, these technologies already offer an appreciable optical quality, the holographic imaging technology presents even a unique feature of large observation volume;⁷⁹ however, optofluidic microscopes are not as versatile as conventional microscopes in that they usually are suitable for one single type of applications. This is because most of the optofluidic devices work at fixed magnification. In addition, one would need a sensor and an algorithm to process these images, as in the case of the multi-aperture technology where one worm is imaged at a time. This type of indirect imaging techniques has a large processing time to rebuild the images. A cluster of graphics processing units would be powerful enough to reconstruct the image in real-time⁸⁰ but the cost and availability for such clusters prevent an easy access to the technology. These caveats make optofluidic microscopy unfit for the everyday maintenance of worm culture. However, these technologies can be excellent alternative tools to conventional ones in specific applications. For instance, the shadow imaging system allows for monitoring behavior of tens of worms simultaneously, which makes it potentially appropriate for an arena-like device.¹⁰ The multi-aperture array device images one worm at a time without stopping the flow, which could be advantageous for microfluidic screening platforms should the sample be amenable for the treatment required in this modality.

Pushing the integration one step further, a miniaturized microscope platform has been demonstrated on a cell-phone.⁸¹ The integrated light-emitting diode (LED) illumination source and wireless communication make the system highly portable. This lens-free digital microscope opens up the field of point-of-care diagnostics with applications targeting the detection of parasitic and pathogenic nematode species such as the Lymphatic filariae. In the future such a platform combined with microfluidics could perform not only imaging but also data analysis by taking advantage of the computing capacities of the cell-phone. Similar to the applications developed for smart phones, specialized applications could be developed to realize specific analysis of worm behavior at low resolution.

Beside miniaturizing microscopy, optofluidics also deals with using fluids to modulate functions of light sources and light paths.⁸² Illumination sources such as dye lasers or electrochemiluminescent sources, and passive components such as waveguides, lenses, and prisms have been successfully integrated on chip; refractive index measurements, biodetection in a laser cavity, and interferometers have been assembled and shown to function on-chip.⁸² These optical components and optical methods offer a panel of tools that wait for the development of new integrated platforms. In the future, these platforms need to demonstrate similar potential as conventional equipment but at a better cost or offer new optical functions. For example, the integration of a matrix of microlenses can increase locally the magnification while keeping a large field of view. The growing effort to develop liquid lenses⁸³ wherein the focus of each lens can be adapted may provide with a technological solution to this challenge, if one succeeds to create a dense array of liquid microlenses individually tunable. The fluorescence signals at high-resolution could be simultaneously monitored for a large number of worms, increasing the throughput in fluorescence-based assays.

CONCLUSION

Here, we have reviewed some microfluidic systems in combination with optical methods that have played a great role in the study of C. elegans. Various chambers, droplet arrays, and valve-based trapping systems have been developed to either process worms in parallel (for high throughput and longer experimental duration) or in series (for high throughput as well as high-resolution imaging); various on-chip immobilization techniques have been designed to accommodate the needs of imaging, optogenetic manipulation, and laser surgery. Additionally, microfluidics enables a high degree of control of the worm environment, unlike the conventional agar plates, as well as allowing precise handling of worms, which can be automated, therefore high-throughput and with reduced human bias. These diverse microfluidic methods used in C. elegans research highlights a trend for custom-made chips for specific applications. With the further development of optical technologies, it is conceivable that the next generation microfluidic designs will need to accommodate three-dimensional or to multiple angles of views, further improve the resolution, and potentially integrate on-chip optics. Particularly, expanding the imaging capability in 3D or to multiple angles of views would be of interest because it significantly enlarges the toolbox for cell biology, neurobiology, and developmental biology. For instance, in neurobiology, with the expanding toolkit of calcium fluorescent reporters, optogenetic effectors, and the use of selective illumination, entire neuronal circuitry could be monitored or manipulated. Putting all the tools together will greatly enhance our ability to decipher complex neuronal circuits and understand how genes and environment give rise to complex behavior.