Center for Cell Control
Executive Summary
Treatment of HIV/AIDS with multiple drugs simultaneously has yielded great success in controlling this disease, and similar strategies are likely to introduce a new era of using drug combinations to treat patients with other diseases. Using multidrug cocktails remains a promising, but little utilized therapeutic approach. We are developing methods, based on engineering principles, to efficiently search among enormous number of possible combinations for the most potent combination to drive cells to behave in specific ways. Given five drugs (each with ten dosages) that can impact viral replication, our methodology reduces over 100,000 possible combinations to be evaluated to approximately 20 combination trials. The resulting optimal drug cocktail allows us to identify the most critical signaling pathways that achieve a desired cellular response. This will open up a new understanding of the signaling pathways contributing to the differentiated states with optimal drug combinations. The tools and techniques will be used for developing novel treatments for a wide range of diseases such as cancer and infections. Our approach should also enable us to better define and re-engineer cell signaling pathways by developing intracellular nanosensors and activators that monitor and eventually manipulate signaling activity to stimulate therapeutically favorable cellular responses.
Introduction
A stunning reduction of deaths from HIV/AIDS resulted from treating patients with a cocktail consisting of multiple drugs which, if administered individually, would not be very effective. The U.S. death rate from HIV/AIDS dropped 60 percent in two years after the first HIV drug combination was introduced in 1995. Except for a few cancer treatments, the clinical utility of drug cocktails for other diseases is limited in part because each drug has its own set of side effects. Moreover, defining an optimal cocktail is difficult to determine empirically from millions of possible combinations [1]. Just as HIV patients on multidrug therapy benefit from a combination of viral-targeted effects, cellular processes such as differentiation, replication, and growth potentially could be manipulated by drug combinations. In general, the biological processes that regulate viral infection, oncogenesis, or stem cell differentiation are driven by a multitude of signaling pathways that affect cellular behavior. We aim to define and re-engineer, via multiple drug stimulation, cell signaling pathways to yield therapeutically favorable cellular responses. Next, we will probe the intracellular protein complexes involved in these pathways by developing intracellular nano-sensors that record kinetics and amplitudes in the activated pathways. These sensors use different molecular substrates of phosphorylating enzymes to provide kinase specificity to help delineate the relevant pathways. For example, pathways such as the JAK-STAT pathway interconnect with a variety of additional pathways, including the PI3K/AKT and MAPK/ERK/JNK pathways; nano-sensors attached to molecular intermediates at critical junctions within these pathways will define the precise signaling kinetics and strengths that yield desired cell outcomes.
In addition to intracellular sensors, we are developing microfluidic circuits integrated with sensors, i.e. "smart"Petri dishes, that provide enabling approaches for exploring in vitro, and soon in vivo, responses of biological systems in real-time at the molecular level. We probe cellular functions primarily in an input-output engineering mode, as exemplified by the iterative feedback-control paradigm embraced by our center. Although our initial work has been to control cell outcomes with rapid determination of optimal drug cocktails, we are beginning to explore and understand the activity of intracellular signaling pathways in order to exert control over the signaling nanomachinery inside cells as well.
NDC Goals
The challenge to generalizing our methods of optimizing drug cocktails is that we must not only identify efficacious drugs but the proper concentrations of each. The possible combination of drugs and concentrations quickly becomes an unmanageable number. For example, six drugs and 10 different concentrations yields 106 possible combinations. We obviously cannot take a trial-and-error approach to searching for the optimal combinations. Utilizing the well established engineering principle of feedback control [2], our research demonstrated that smart search algorithms can efficiently identify the most potent drug cocktail in tens of iterations instead of millions. We also learned that the concentrations of the drugs combined together require significantly lower doses, one tenth or less, compared to single drug administration. We will establish the scientific basis for the utility of this unprecedented approach towards the regulation of a spectrum of cellular functions that will be useful in a variety of biological contexts such as cancer eradication, inhibition of viral replication, and stem cell differentiation.
The increasingly recognized therapeutic power of drug combinations lies in the synergetic effects on signaling pathways within the cell [6]. Interactions among the millions of intracellular molecules carry out these signaling cascades. It is through these complex inter-connected networks that a cell senses, processes, and adapts to its environment. In order to understand why a drug cocktail functions so well, our Center is developing a suite of nano-sensors [3] that can be inserted into live cells, to explore the responses of the network of signaling pathways under the influence of the drug combination.
CCC organizes a unique team of biologists, engineers and clinical doctors at the University of California, Los Angeles and the University of California at Berkeley for accomplishing this highly interdisciplinary task. Furthermore, we also collaborate with other NDCs to contribute to our common goal — pushing the envelope of nanomedicine.
Progress
We are establishing three fundamental capabilities essential for achieving our goals: 1) selecting cells that behave similarly from a variable population of cells , 2) use drug combinations to elicit desired cellular phenotypes, and 3) investigate the biochemical pathways for those specific phenotypes.
Unlike other centers, our outcome measures are not just basic biological processes but we begin with diseases. We intend to perform studies on three model systems: stem cells, viral infection, and cancer. At this early stage, selection of the best embryoid bodies (EB) is used to demonstrate the capability of sorting controllable and reliable progenitors for further stem cell differentiation. Inhibition of viral replication in cells will be the model endpoint for evaluating optimal drug combinations that can elicit a desired cell response. Finally, we will track the phosphorylation events in the JAK-STAT signaling pathway of cancer cell lines to demonstrate unique nano-modalities for quantitatively interrogating signaling cascades.
Selecting a subset samples from a collection of mixed cells
Cells display an inherent stochasticity from their biochemical processes that would generate significant diversity within any group of cells even when a cell line is established from genetically identical cells exposed to similar environments. This is one of the contributing factors of high noise in biological experiments. One of the tasks in CCC is to develop pre-selection techniques of samples for ensuring uniform starting point and reducing noise.
Sorting embryoid bodies (EB) for reliable and controllable progenitors for cell differentiation: Pluripotent human embryonic stem (hES) cells provide a powerful therapeutic opportunity for treating diseases such as diabetes, cancer and neurodegenerative disorders. However, culture systems that allow for a reliable and controllable cellular proliferation and differentiation in an animal and pathogen free environment are currently lacking and thus have hindered the development of efficacious therapeutics. Of the several ES cell differentiation methods, the formation of three dimensional cells aggregates known as embryoid bodies (EBs) is a common and critical intermediate prior to the induction of lineage specific differentiation. However, EBs generated using the traditional low adhesion culture dishes (LAC) are heterogeneous, unsynchronized and inefficient, which makes the production of a particular cell type in large scale rather difficult.
To improve EB-mediated ES cell differentiation, we have conducted a systematic study to evaluate whether EB size would influence proliferation and differentiation. By characterizing EBs according to their size distribution, our recent work demonstrated that EBs with an intermediate size (100-300um in diameter) are the most proliferative, display the greatest differentiation potential and have the lowest rate of cell death.
We also investigated other biocompatible surface materials for enriching intermediate size EBs. In an attempt to control EB formation and enrich for the intermediate size, we directly compared EBs formed on LAC with other biocompatible surface materials, such as agarose, polythethylene glycol silica sol gel (PEG-sol), polydimethylsiloxane (PDMS), and poly-hydroxyethylmethacrylate (poly-HEMA). Interestingly, PDMS coated hydrophobic surface promoted intermediate size EB formation by 2-fold. To further examine the role of hydrophobicity in EB development on a more chemically defined system, we elected to mimic the PDMS interface using self assembled monolayer (SAM) films of alkanethiolates formed on gold coated substrates. We found that the most hydrophobic surfaces appear to better promote EB development and uniformity (Figure 1). Interestingly, hydrophobic surfaces can significantly prevent ES cells adherent to the culture dish that may contribute to efficient EB formation in the suspension cultures. Therefore, surface condition can pre-select the intermediate size EB population may significantly improve the efficiency of ES cell differentiation and the yields of specific cell types.
-
- Figure 1. Hydrophobic surfaces enhance uniform EB formation. The size distribution generated by various surface EBs were cultured on various surfaces.
Microfluidic system for EB sorting: The hydrophobic surface can promote the uniformity of EB growth, but the size scattering is still fairly large. For ensuring high quality pre-selected EBs, we developed an ultra-compact microfluidic continuous flow sorter. By predefining the locations of the particles in the microchannel with a sheath flow, their paths can be differentiated by the pathlines distributed at different positions from the channel wall (Figure 2A). Experiments were performed to successfully separate EBs into two groups; 0 - 50 um and 50 - 100 um with 94% of the smaller particles (Figure 2B) and 100% of the larger ones (Figure 2C) going to their respective bins. Multi-level separation is possible by adding additional pillars within the microchannel. Simulations for multi-leveled separation are presented in Figure 2D for dividing EBs into five distinct-sized groups.
-
- Figure 2. (A) Computational simulation results showing the pathlines for separation intwo two groups: 0-50um and 50-100um. (B) Trace of mouse EBs 30um in size and (C) 100um in size. Images were generated by superimposing video frames. (D) Pathlines calculated for separation into five groups: 0-100um, 100-150um, 200-250um, and 250-300um.
Control and Elicit Cells toward Desired Phenotype
To regulate a complex biological system with the closed-loop optimization approach, input stimuli (drugs) are applied to stimulate the system of interest. Outputs (key biological responses) that contain valuable information about the system are then evaluated. With a search algorithm, we make use of this information to search for appropriate inputs that drive the systems towards desired outputs. This allows a systematic strategy, a protocol, for manipulating the regulatory networks and exploring unknown complex dynamics.
Global optimization of complex systems over large parameter spaces is often needed in scientific and engineering research. Effective search of the specific combination of input stimuli for optimizing certain properties of a system output is the challenge. The basic approach of a search algorithm is similar to the mountain-climbing procedure. A random starting location is picked to begin. Based on the system output and the chosen search algorithm, e.g. Gur game, differential evolution game or cross entropy game, a second set of stimuli is selected to move the system output to a neighboring position at a higher elevation of the mountain. Continue iterations may lead toward the peak of the mountain. Under the guidance of extensive studies in various search algorithms, we apply the methods in searching for the potent drug combination for directing the cellular system responses.
Elicit strong inhibition of viral infection in cells by searching for optimal drug combination stimulation: Viruses infect an individual by entering a cell as a Trojan horse, quietly entering a cell and subsequently taking over the cellular systems. By hijacking control of the cell, a virus is able to replicate itself using the cell's own machinery and metabolism, which can eventually lead to cellular necrosis thru lysis. Developing anti-viral treatments is a difficult endeavor for a variety of reasons. Any drug that is used to target a viral infection will invariably begin hurting normal cellular systems if used at a high enough dose. Consequently, anti-viral drugs are notorious for having side effects that range from minor to severe.
Another challenge in developing anti-viral therapies is that viruses are constantly mutating in order to survive. Just as cells change their surface receptors so viruses are unable to attach, viruses change their surface proteins so they can attach to the altered cell surface receptors. This can be deleterious to someone being treated for a viral infection since a viral mutation may render a previously successful therapy ineffective This problem was most commonly seen during the HIV epidemic when patients were no longer helped by certain anti-viral treatments they were taking.
An approach to combat these problems is to take a combination or "cocktail"of anti-viral drugs. This approach affords two benefits: first, the combinatory drugs synergistically work on several pathways at the same time, and two, you could use either drug at lower, more tolerable doses. This approach was highly effective in treating HIV infection but up until now has required tremendous amounts of time to develop suitable, effective combinations. To increase the rate of drug development we have developed a systematic process which uses a mathematical search algorithm to determine the optimal combination of anti-viral doses. Using this protocol we can quickly screen vast numbers of combinations and determine the optimal dosage of drugs to use to stop a viral infection.
We applied this system to determine the optimal anti-viral combination for the virus Vesicular Stomatitis Virus (VSV). Using our approach, we were able to quickly determine that when used in combination, the doses of individual anti-viral drugs were able to be decreased substantially, as much as ten fold. We have continued to harness this system by applying it to infection of Herpes Simplex virus (HSV). HSV is a common infection of humans and is the cause of the sporadic herpes sores that affect roughly 10 to 30% of individuals in the USA. Though the infection is common, to date, few effective treatments are available. Using our protocol we have already found one effective low dose combination in preventing viral infection in cellular experiment. We are currently translating these findings to a mouse model of HSV infection to validate the system as well as further the prospect of a more effective HSV treatment.
Interrogating the cascade of pathways under combinatory drug stimulations
Cells respond to signals from the environment by relaying signals from the cell surface to the nucleus to alter the expression of genes that control cell fate and function. Usually these signals are relayed by chemical modifications on proteins within a pathway, often by the transfer of inorganic phosphates. Complexity in this signaling mechanism is brought about by cross-talk between multiple distinct signaling pathways and by time-dependent signal generation and propagation considerations. Currently, there is no live cell method for quantitatively evaluating signal generation and propagation, as well as signaling network interactions, available for use. With a small number of iterations using the feedback control scheme, we have demonstrated that drug cocktails with ten-fold lower dosage can prevent viral infections. For understanding the overall responses of the cellular network machinery to the potent drug cocktail we are developing Surface Enhanced Raman Spectroscopy (SERS) based Tunable Nano Plasmon Resonators (TNPRs) [4] to monitor real-time cell signaling pathways.
-
- Figure 3. Schematic of TNPR.
Monitoring real-time phosphorylation using Tunable Nano Palsmon Resonators (TNPRs): SERS provides a method to identify chemical species in solution, without incorporating labels, through the monitoring of the vibration modes of chemical bonds. However, SERS signals are too weak for small number of events, such as that of phosphoryloation in signaling pathway. To amplify the signal, we apply the TNPRs. The TNPRs are vertically layered particles, alternating gold or silver and an insulator layer. This provides precise control over the resonance frequency of the TNPR allowing tuning of the resonant frequency. Furthermore, these resonators provide the highest reported SERS enhancement factors for single particles, on the order of 1010, making them ideal for biological studies.
In monitoring cell pathways we are examining phosphorylation of key peptides associated with signal transduction. Specifically, the JAK-STAT signaling pathway is being monitored via the observation of JAK3tide phosphorylation. JAK3tide was immobilized on the surface of the TNRPs and the SERS signal was acquired both before and after the addition of JAK3 kinase. Upon the addition of JAK3 kinase the peptide is phosphorylated and the SERS signal changes in a time and dose dependent manner that can be monitored in real-time without the addition of any labeling moieties.
Solution based TNPR will be used to perform the key intracellular measurements of phosphorylation based activity. Currently, we are working on solution based SERS measurements of JAK3tide phosphorylation both inside and outside of cells. This will lead to the ultimate goal of monitoring cell pathways in real-time in a time and dose dependent manner.
Single cell based nanoparticle tracking for quantitative evaluation of phosphorylation signals in pathway cascades: Until recently, biological research was accomplished by observing the reaction of a population of cells to a certain stimulus. Researchers at CCC are applying a unique tool that enables noninvasive cell manipulation, known as optoelectronic tweezers (OET) [5]. The OET device uses light patterns to control the positions of single cell, allowing researchers to trap, transport, and sort single cells. Multiple single cells can be controlled at the same time, making it possible to study many single cells at once, and compare the results. This technology was used to successfully manipulate live HeLa and Jurkat cells (two types of cells widely used in biological research) in phosphate-buffered saline solution (PBS) and Dulbecco's Modified Eagle Medium (DMEM), without affecting the cells' viability. Besides using this technique to manipulate single cells, we will apply it to develop an electroporation-based insertion of TNPR into live cells for quantitative phosphorylation measurement.
-
- Figure 4. (A) A 3x2 array of trapped single HeLa cells on the Ph-OET device. The light patterns form stable single cells box-shaped traps around the cells. (B) Fluorescent image of the array in (A) shows that the HeLa cells have incorporated a fluorescent dye that indicates the cells are viable.
To make sure that TNPRs are properly located inside cells, we have developed a particle tracking method called biomicrorheology [6] to make certain that the signaling molecules are in the same location as the TNPRs, facilitating detection. Recently we have determined uptake parameters that, for particles sized similarly to TNPRs (about 50-100 nm each), can obtain about 200 particles in a cell without damaging the cell, can track the particles to determine their intracellular positions and movement using optical microscopy and very fast CCD camera collection of fluorescently-functionalized particles, and have written and published analytic computerized methods for handling the copious data generated.
Future Applications
Human disease can be caused by unbalanced signaling pathways, which are triggered by pathological, external or internal factors, such as by invading microorganisms, accumulations of toxic substances, and uncontrolled cell division. The challenge of correcting such imbalances by driving a biological system to health is compounded by the complexity of the regulatory and signaling pathways governing the cell. Simple fixes which, for example, target individual molecular complexes or pathways are likely to fail for a variety of reasons, including the adaptability built into the signaling network through numerous pathway interconnections. Treatments that target multiple signaling networks are almost certainly required to control disease and re-establish health. The CCC addresses this challenge by first applying a well established engineering principle of system control to determine the optimal combination of drugs to steer cells toward a desired state. Our experiments on five different cell systems have demonstrated that the feedback control is an efficient way to lead the cells toward desired phenotype and our nano-sensors will define the cellular functions under multiple stimulations to achieve this integrated response. It does not escape our attention that this methodology can be applicable to a wide variety of biological systems.
In parallel with in vitro and in vivo cell control studies, CCC members will provide detailed nanoscale molecular studies using nano-tools currently under development to quantify the relevant signaling network components that achieve a desired cell state. In this regard, iterative feedback control will help to limit the number of pathway nodes requiring monitoring as drugs are reduced or even eliminated from successful cocktails that provide a desired cell outcome.
In the future, we will extend the approach to animal and clinical tests. At the same time, we need to recognize that the bar for simple translation into animals and ultimately people is very high. The new challenges are combinatorial drug toxicities, distinct bioavailabilities, differences in metabolism and excretion rates. One day powerful drug-combination treatments, including both FDA approved medicines and newly developed medicines targeting specific nanoscale molecular signaling complexes may become a reality, perhaps similar but more tightly quantified than those that are currently successful with HAART anti-HIV therapies.
References
Background Papers
- Dancey, Janet E. and Chen, Helen X., "Strategies for optimizing combinations of molecularly targeted anticancer agents", Nature Reviews Drug Discovery 5, 649-659, 2006
- Sontag, E.D., "Molecular Systems Biology and Control", European Journal of Control, 11, 1-40, 2005.
- Ho, D., Garcia, D., and Ho, C.M., "Nanomanufacturing and Characterization Modalities for Bio-Nano-Informatics Systems", J. Nanoscience and Nanotechnology 6, 875-891, 2006
- Su, K., Wei, Q. H., and Zhang, X., "Tunable and augmented plasmon resonances of Au/SiO2/Au nanodisks", Appl. Phys. Lett. 88, 063118. 2006
- Chiou, P. Y., Ohta, A. T. & Wu, M. C., "Massively parallel manipulation of single cells and microparticles using optical images", Nature, 436, 370-2. 2005
- Jayne M. Stommel, et al "Coactivation of Receptor Tyrosine Kinases Affects the Response of Tumor Cells to Targeted Therapies", Science 318, 287, 2007
NDC Supported Papers
- Weihs D., Teitell M.A., and Mason T.G., "Simulations of complex particle transport in heterogeneous active liquids", Microfluid Nanofluid 3:227-237. 2007
- Kuraishy, A.I., French, S.W., Sherman, M., Herling, M., Jones, D., Wall, R., and Teitell, M.A., "TORC2 Regulates Germinal Center Repression of the TCL1 Oncoprotein to Promote B Cell Development and Inhibit Transformation", Proceedings of the National Academy of Sciences, 104:10175-10180, 2007
- Wong, P.K., Yu, F., Sun, R., and Ho, C.M., "Unraveling Gene Regulatory Networks using an Integrated Microfluidic Platform", to appear in IEEE Transactions on Nanotechnology, 2008
- Weihs, D., Mason, T.G., and Teitell, M.A., Effects of Cytoskeletal Disruption on Transport, Structure, and Rheology within Mammalian Cells. Physics of Fluids, 19:103102-1, 2007
Contact Information
Principle Investigator
Chih-Ming Ho, Ph.D.
Ben Rich-Lockheed Martin Professor
UCLA Distinguished Professor
Director of the Center for Cell Control (CCC) and
Director of the Institute for Cell Mimetic Space Exploration (CMISE)
Mechanical and Aerospace Engineering Department
School of Engineering and Applied Science
Engineering IV, Room 38-137J
420 Westwood Plaza
Los Angeles, CA 90095-1597
Phone: 310-825-9993
Fax: 310-206-2302
Email: chihming@ucla.edu
Website
Co-Investigators
Genghong Cheng, Ph.D.
Professor, Dept. of Microbiology, Immunology and Molecular Genetics
University of California Los Angeles
8-240 Factor Building
10833 Le Conte Avenue
Los Angeles, CA 90095-1781
Tel: 310-825-8896
Fax: 310-206-5553
Email: genhongc@microbio.ucla.edu
Website
Hong Wu, M.D., Ph.D.
Professor, Molecular & Medical Pharmacology, UCLA
CHS 23-214
Tel: 310-825-5160
Fax:310 267 0242
Email: HWu@mednet.ucla.edu
Website
Michael Teitell, M.D., Ph.D.
Chief, Pediatric and Developmental Pathology
David Geffen School of Medicine at UCLA
675 Charles Young Dr. South
MRL 4-762
Los Angeles, CA 90095-1932
Tel: (310) 206-6754
Fax: (310) 267-0382
Email: mteitell@mednet.ucla.edu
Website
Ming Wu, Ph.D.
Professor
Department of Electrical Engineering and Computer Sciences
University of California, Berkeley
261M Cory Hall
#1770 Berkeley, CA 94720-1770
Tel: 510-643-0808
Fax: 510-666-2502
Email: wu@eecs.berkeley.edu
Website
Xiang Zhang, Ph.D.
Chancellor's Professor
5130 Etcheverry Hall, Mailstop 1740
University of California at Berkeley
Berkeley, CA 94720-1740
Tel: 510-643-4978
Email: xzhang@me.berkeley.edu
Website
