The signaling environment experienced by a single cell is highly dependent on its interactions with its neighbors, which may secrete locally acting factors or act via membrane bound receptor-ligand systems. Stem cells are particularly sensitive to these signals, which act in concert to regulate self-renewal and discrete transitions into distinct fates. In the case of the adult neural stem cell, decades of in vivo and in vitro work have illustrated specific roles for different types of niche cells and some of the molecular mediators of their instructive roles. However, these studies have not investigated the strengths of these fate-inducing or fate-repressing cues as presented by single neighboring cells at endogenous expression levels. In this work, we have developed and applied microengineered tools to examine the transmission of signals between small populations of cells and to assess their impact on neural stem cell fate.

First, we measured gap junction coupling between cells by using microfluidic flow focusing of calcein/AM dye and timelapse imaging to measure the dynamics of dye transfer through gap-junction coupled glioma cells. Secondly, we developed single-cell micropatterning methods to investigate the potential for single neural progenitor cells to influence their neighbor's fate. Size-matched circular and hourglass shaped polystyrene microwells achieved up to 80% efficiency in capturing single and paired neural progenitor cells, a large improvement over Poisson-distributed random seeding. In conjunction, we also developed a method to fabricate and align PDMS meshes to corral patterned cells into non-connected arenas so that single cells and cell pairs can be grown in isolation for six days before immunostaining for fate markers. Lastly, we applied a simplified micropatterning technique in conjunction with automated high-throughput imaging to enforce persistent interactions between initially patterned cells. We discovered that single neural progenitor cells can significantly bias their neighbor's fate in mixed differentiation medium. Although the initial number of cells on a pattern did not affect the overall distribution of different fates, we saw reduced fate asymmetry on patterns with initially paired cells, indicating that single cells may be inducing a similar fate in their neighbors, or repressing differentiation. By tracking cell populations on micropatterns on every day of the experiment via brightfield imaging, we also ascertained that the initial patterning state was maintained for approximately two days, suggesting that fate specification or repression occurred early in the process. Our discovery that a single as-yet undifferentiated neural stem cell can significantly bias its neighbor's fate paves the road for future work in elucidating how conflicting signals from various types of niche cells are arbitrated in a single recipient stem cell.

Over the past decade, there has been rapidly growing interest in wearable and implantable devices for a wide range of biomedical applications. For many applications involving prolonged contact with the body, devices that are compliant and can comfortably conform to and move with the patient are highly preferred. These flexible substrates (i.e. clothing, bandages, meshes, catheters, etc) can be instrumented to measure various physiological markers, such as temperature, pH, and oxygenation levels, to better inform clinical care. In this dissertation, I will discuss two examples of sensors designed to interact with flexible substrates for clinical monitoring and diagnosis applications.

I will first present the development of an electronic bandage to objectively monitor the progression of wound healing in pressure ulcers and other chronic wounds. Chronic skin wounds affect millions of people each year and take billions of dollars to treat. Pressure ulcers are a type of chronic skin wound that can be especially painful for patients and are tricky to treat because current monitoring solutions are subjective. We have developed an impedance sensing tool to objectively monitor tissue health in wounds. An electrode array is printed onto a flexible, polymeric substrate to form a “smart” bandage. With this sensor array, we can measure impedance of the underlying tissue and extract information on tissue health (i.e. size of wounds and tissue types) to inform the clinical course of treatment.

In the second half of the dissertation, I will discuss methods for instrumenting hernia mesh prosthetics to provide quantitative guidance to surgeons during hernia repair surgeries. Abdominal wall hernias are typically treated by suturing in a surgical mesh to cover and overlap the hernia defect. However, in 10-20% of patients, the hernia repair fails, resulting in recurrence of the hernia. 18% of these recurrences are attributed to mechanical failure of the mesh, often due to unequal stress distribution across the mesh surface resulting in high stress concentrations at the tissue-mesh interface. Strain across the mesh can be used as an indicator for how evenly stress is distributed across the surface of the mesh. I will discuss two methods, based on optical approach and magnetoelastic approaches, for instrumenting the hernia mesh prosthetic to measure the stress distribution during the surgical repair process and the postoperative healing period.

There is a growing need for technology that can control microscale oxygen gradients onto a tissue or culture sample in vitro. This dissertation introduces the oxygen microgradient chip (OMA), which employs electrolysis to generate oxygen microgradients within cell culture without forming bubbles. Dissolved oxygen generated at noble microelectrodes patterned on a chip surface diffuses through a gas-permeable silicone membrane and is dosed into cell culture. The amount of generated oxygen is directly proportional to a current flowing across the electrodes and thus can be controlled with high precision. By real-time modulation of the current, spatial profile of the oxygen concentration microgradient can be modified during an experiment. Distinct microgradients generated by different electrodes can be superimposed to pattern arbitrary and multi-dimensional oxygen concentration profiles with microscale resolution. Design of the OMA is based on simulation results of diffusion and physical considerations. Microfabrication, assembly process, and improvements of the early OMA are presented. Use of an oxygen-sensitive fluorophore film verifies that the measured microgradients in OMA match simulation results. Generation of reactive oxygen species during the electrolysis is also considered and quantified. This technology enables the biological study of gradient-related effects. Three different, well-known biological models are introduced to demonstrate how the microtechnology enables new types of experiments. Lastly, we explore the effect of oxygen gradients on muscle progenitor cells having half-deficiency of superoxide dismutase-1 (SOD1) enzyme. Another type of oxygen gradient chip that utilizes atmospheric oxygen is developed and used in this experiment. Myotube formation of the SOD1 heterozygote myoblasts decreased significantly compared to that of wild-type myoblasts, indicating that the oxygen stress on SOD1 heterozygote myoblasts induces defective cellular function on differentiation.

Molecular detection and analysis are of fundamental importance in disease prevention, disease diagnosis, medical treatment, drug delivery, food industry, and environmental monitoring. Conventional immunoassays require hours of incubation for low concentration analytes (femtomolar) since the rate-limiting step is the transport of target molecules to the biosensors. Thus, a new technology for fast and sensitive immunoassay is highly desirable for disease monitoring and personalized treatments. The work reported in this thesis is focused on developing silicon microfabrication technologies for sensors that can detect low concentration proteins (femtomolar) in a short time (tens of minutes).

We first demonstrate the use of germanium (Ge) films as sacrificial layers that allow the patterning of proteins onto surfaces with commonly used organic solvents. As researchers miniaturize biosensors and microfluidic devices down to submicron scales, high-resolution biomolecule conjugation compatible with these processes is highly desirable. The presented technique is scalable to high volume manufacturing and is compatible with nano- and microfabrication processes, including standard lithography. We achieved nanoscale resolution and misalignment with this technique.

We then discuss the development of ICP-based preconcentration devices on silica and silicon substrates. We report two nanofabrication strategies in this chapter that enhance the accumulation of protein molecules via 1) an increase in the number of nanochannels per microchannel and 2) an increase in the depth of nanochannels. Increasing the depth of nanochannels leads to the switch from silica substrates to silicon substrates, using deep reactive ion etch (DRIE). We also report the attempts for antibody immobilization in the ICP preconcentrators using previously reported Ge technique. However, due to the instability of antibodies in the drying process, we switched to bead-based immunoassays in our preconcentration devices.

After the demonstration of protein preconcentration in our devices, we present a scalable method for fabricating hundreds of vertical nanochannels (6.5 µm deep) employing ion concentration polarization (ICP) enrichment for fast analyte detection. Compared to horizontal nanochannels, massively paralleled vertical nanochannels not only provide comparable electrokinetic functions but also significantly reduce effective fluid resistance in each microfluidic channel, which enables microbead loading for sensing purposes. These nanochannels filter microbeads by size and preconcentrate analytes at the anodic side of the test area via electrokinetic entrapment. The device is capable of enriching protein molecules by >1000 fold in 10 min. We demonstrate fast detection of IL6 down to 7.4 pg/ml with only a 10 min enrichment period followed by a 5 min incubation, which is a 162-fold enhancement in sensitivity compared to that without enrichment. Our results demonstrate the possibility of using silicon/silica-based vertical nanochannels to mimic the function of polymer membranes for protein enrichment.

Implantable medical devices have tremendous therapeutic potential for both treatment of

disease states as well as monitoring for preventative care. To be effective, implants must

perturb host tissue as minimally as possible, while simultaneously being able to withstand

the tissue environment for a significant portion of a patient’s life. Advances in wireless

power transfer technology have made it possible to implant millimeter to sub-millimeter

scale devices fully within the body, but these devices have not been demonstrated to work

for decadal time spans.

Typical packaging material for medical implants are polymers such as silicone or parylene,

or titanium. However, polymers are not chronically hermetic due to their high water va-

por permeability and titanium is not easily amenable to electromagnetics-based wireless

communication. Ceramic materials, however, have a low water vapor permeability and are

transparent to radio-frequency radiation. Furthermore, they have had a long history in medi-

cal implants and semiconductor processing methods have greatly increased the compatibility

of ceramics processing with other materials.

This thesis explores the use of ceramics as packaging materials for wireless, miniaturized,

implantable medical devices. An alumina-titanium hybrid package is first demonstrated for

a millimeter-scale ultrasonically-coupled wireless implant. Sound propagation through solids

in two different modes is investigated, test packages are assembled, and performance is eval-

uated. Next, silicon carbide is explored as a potential packaging material for wireless radio-

frequency identification tags. A system for rapidly testing and aging silicon carbide thin-films

is designed and demonstrated, and progress towards building chronic silicon-carbide encap-

sulated radio-frequency identification tags is shown.

Temperature measurements find use in clinical settings for the monitoring of various disease states, and the monitoring of deep tissue temperature patterns in both space and time can provide valuable information about tissue health. Unfortunately, current clinical methods are not suitable for continuous monitoring of temperature measurements. Recently, acoustically-powered implants have been emerging as a method to enable small medical devices that can be implanted deep in the tissue. In this thesis, we demonstrated the development of two versions of implantable temperature sensors that we call thermal dust. Thermal dust motes are ultrasonically-powered, ultrasonically-communicating sensors, and were demonstrated as both a passive sensor mote and an active sensor mote. We designed passive thermal dust motes using only commercial off-the-shelf (COTS) components, and these sensors were fully sub-millimeter in all of their dimensions. We showed experimentally that passive thermal dust motes were able to encode temperature data using analog amplitudebackscatter modulation. Active thermal dust motes consist of a custom integrated circuit (IC) that we designed in a TSMC 0.18 μm biCMOS process. We experimentally demonstrated the operation of our thermal dust chip in the physiologically-relevant temperature range of 30 – 50°C as well as an elevated temperature range of 60 – 80°C to demonstrate the versatility of our device. Finally, we showed that the binary duty cycle modulation of the backscatter waveform envelope is a feasible method to encode temperature data, and we demonstrated fully wireless operation of an active thermal dust device.

Nanoscale machines which directly convert chemical energy into mechanical work are ubiquitous in nature and are employed to perform a diverse set of tasks such as transporting molecules, maintaining molecular gradients, and providing motion to organisms. Their widespread use in nature suggests that large technological rewards can be obtained by designing synthetic machines that use similar mechanisms.

This thesis addresses the technological adaptation of a specific mechanism known as the Brownian ratchet for the design of synthetic autonomous nanomachines. My efforts were focused more specifically on synthetic chemomechanical ratchets which I deem will be broadly applicable in the life sciences. In my work I have theoretically explored the biophysical mechanisms and energy landscapes that give rise to the ratcheting phenomena and devised devices that operate off these principles.

I demonstrate two generations of devices that produce mechanical force/deformation in response to a user specified ligand. The first generation devices, fabricatied using a combination nanoscale lithographic processes and bioconjugation techniques, were used to provide evidence that the proposed ratcheting phenomena can be exploited in synthetic architectures. Second generation devices fabricated using self-assembled DNA/hapten motifs were constructed to gain a precise understanding of ratcheting dynamics and design constraints. In addition, the self-assembled devices enabled fabrication en masse, which I feel will alleviate future experimental hurdles in analysis and facilitate its adaptation to technologies.

The product of these efforts is an architecture that has the potential to enable numerous technologies in biosensing and drug delivery. For example, the coupling of molecule-specific actuation to the release of drugs or signaling molecules from nanocapsules or porous materials could be transformative. Such architectures could provide possible avenues to pressing issues in biology and medicine: drugs could eventually be triggered to release in the presence of molecular signals indicative of diseased states, early disease detection could be achieved by examining the cell microenvironment then releasing imaging agents and generalized control could exerted over the free molecule signaling networks of cells.

This dissertation presents a millimeter-scale CMOS 64x64 single charged particle radiation detector system for cancer radiotherapy; this work is applicable to therapies including external beam radiotherapy (EBRT), targeted radionuclide therapy (TRT), and brachytherapy. We demonstrate 1 um x 1 um diode pixels capable of measuring energy deposition into a depletion region by single charged particles. These pixels can be arrayed to provide large detection areas; we demonstrate a 512 mm x 512 mm array. Instead of sensing the voltage drop caused by radiation, the system measures pulse width, i.e., the time it takes for the voltage to return to its baseline. This obviates the need for power-hungry and large analog-to-digital converters. A prototype ASIC is fabricated in TSMC 65 nm LP CMOS process and consumes the average static power of 0.535 mW under 1.2 V power supply. The functionality of the whole system is successfully verified in a clinical 67.5 MeV proton beam setting, as well as MeV range electrons and X-ray beam produced by linear accelerator.

Wearable sensors and instrumented implants have wide-ranging clinical applications, and their continued development in recent years has led to advances in personalized healthcare. Innovations in electronics and trends in mobile health are extending the capabilities of traditional medical practice. For a number of clinical conditions, there remains a lack of standardized methods for assessing patients, with physicians often relying on subjective physical examinations that can differ depending on medical training and experience. The work in this thesis focuses on developing objective tools using electrical impedance spectroscopy (EIS) to monitor surface and fracture wounds. EIS measures the frequency-dependent opposition to the flow of electrical current, and has been used to quantitatively characterize cellular changes. This work begins by examining pressure ulcers, tissue damage caused by prolonged pressure on the skin, which are commonly analyzed using photographs and ruler measurements. Prevention efforts involve nursing staff repositioning patients every few hours to relieve pressure on a particular area, which is highly time and labor intensive. A “smart bandage” was designed to non-invasively detect and monitor pressure ulcers by identifying changes in tissue health. By utilizing EIS over a flexible electrode array, visually intuitive maps were produced to determine areas with tissue damage and areas at risk for further injury. To extend this technology to deep tissue injuries, this work also details development of an instrumented implant to monitor bone fracture healing. X-ray radiographs are the most common method of assessing fracture repair, but are only useful at later stages because they rely on mineralization of tissue. Microscale sensors were designed and built to locally measure healing fractures in rodent models, and EIS was employed to robustly track longitudinal differences in healing trends. Advancements in miniaturized and flexible electronics have enabled these kinds of physiological monitoring devices, which can guide clinical decision-making and facilitate early intervention. By instrumenting items like bandages and fracture fixation implants already widely used in hospital settings, this work demonstrates that EIS technology can be integrated into current health management strategies with minimal disruption to clinical workflows.