John G. Albeck

Non-genetic origins of cell-to-cell variability in TRAIL-induced apoptosis

In microorganisms, noise in gene expression gives rise to cell-to-cell variability in protein concentrations. In mammalian cells, protein levels also vary and individual cells differ widely in their responsiveness to uniform physiological stimuli. In the case of apoptosis mediated by TRAIL (tumour necrosis factor (TNF)-related apoptosis-inducing ligand) it is common for some cells in a clonal population to die while others survive—a striking divergence in cell fate. Among cells that die, the time between TRAIL exposure and caspase activation is highly variable. Here we image sister cells expressing reporters of caspase activation and mitochondrial outer membrane permeabilization after exposure to TRAIL. We show that naturally occurring differences in the levels or states of proteins regulating receptor-mediated apoptosis are the primary causes of cell-to-cell variability in the timing and probability of death in human cell lines. Protein state is transmitted from mother to daughter, giving rise to transient heritability in fate, but protein synthesis promotes rapid divergence so that sister cells soon become no more similar to each other than pairs of cells chosen at random. Our results have implications for understanding ‘fractional killing’ of tumour cells after exposure to chemotherapy, and for variability in mammalian signal transduction in general.

A Systems Model of Signaling Identifies a Molecular Basis Set for Cytokine-Induced Apoptosis

Signal transduction pathways control cellular responses to stimuli, but it is unclear how molecular information is processed as a network. We constructed a systems model of 7980 intracellular signaling events that directly links measurements to 1440 response outputs associated with apoptosis. The model accurately predicted multiple time-dependent apoptotic responses induced by a combination of the death-inducing cytokine tumor necrosis factor with the prosurvival factors epidermal growth factor and insulin. By capturing the role of unsuspected autocrine circuits activated by transforming growth factor–α and interleukin-1α, the model revealed new molecular mechanisms connecting signaling to apoptosis. The model derived two groupings of intracellular signals that constitute fundamental dimensions (molecular “basis axes”) within the apoptotic signaling network. Projection along these axes captures the entire measured apoptotic network, suggesting that cell survival is determined by signaling through this canonical basis set.

Imaging Cytosolic NADH-NAD+ Redox State with a Genetically Encoded Fluorescent Biosensor

NADH is a key metabolic cofactor whose sensitive and specific detection in the cytosol of live cells has been difficult. We constructed a fluorescent biosensor of the cytosolic NADH-NAD(+) redox state by combining a circularly permuted GFP T-Sapphire with a bacterial NADH-binding protein, Rex. Although the initial construct reported [NADH] × [H(+)] / [NAD(+)], its pH sensitivity was eliminated by mutagenesis. The engineered biosensor Peredox reports cytosolic NADH:NAD(+) ratios and can be calibrated with exogenous lactate and pyruvate. We demonstrated its utility in several cultured and primary cell types. We found that glycolysis opposed the lactate dehydrogenase equilibrium to produce a reduced cytosolic NADH-NAD(+) redox state. We also observed different redox states in primary mouse astrocytes and neurons, consistent with hypothesized metabolic differences. Furthermore, using high-content image analysis, we monitored NADH responses to PI3K pathway inhibition in hundreds of live cells. As an NADH reporter, Peredox should enable better understanding of bioenergetics.

Modeling a snap-action, variable-delay switch controlling extrinsic cell death.

When exposed to tumor necrosis factor (TNF) or TNF-related apoptosis-inducing ligand (TRAIL), a closely related death ligand and investigational therapeutic, cells enter a protracted period of variable duration in which only upstream initiator caspases are active. A subsequent and sudden transition marks activation of the downstream effector caspases that rapidly dismantle the cell. Thus, extrinsic apoptosis is controlled by an unusual variable-delay, snap-action switch that enforces an unambiguous choice between life and death. To understand how the extrinsic apoptosis switch functions in quantitative terms, we constructed a mathematical model based on a mass-action representation of known reaction pathways. The model was trained against experimental data obtained by live-cell imaging, flow cytometry, and immunoblotting of cells perturbed by protein depletion and overexpression. The trained model accurately reproduces the behavior of normal and perturbed cells exposed to TRAIL, making it possible to study switching mechanisms in detail. Model analysis shows, and experiments confirm, that the duration of the delay prior to effector caspase activation is determined by initiator caspase-8 activity and the rates of other reactions lying immediately downstream of the TRAIL receptor. Sudden activation of effector caspases is achieved downstream by reactions involved in permeabilization of the mitochondrial membrane and relocalization of proteins such as Smac. We find that the pattern of interactions among Bcl-2 family members, the partitioning of Smac from its binding partner XIAP, and the mechanics of pore assembly are all critical for snap-action control.

Frequency-Modulated Pulses of ERK Activity Transmit Quantitative Proliferation Signals

The EGF-stimulated ERK/MAPK pathway is a key conduit for cellular proliferation signals and a therapeutic target in many cancers. Here, we characterize two central quantitative aspects of this pathway: the mechanism by which signal strength is encoded and the response curve relating signal output to proliferation. Under steady-state conditions, we find that ERK is activated in discrete, asynchronous pulses with frequency and duration determined by extracellular concentrations of EGF spanning the physiological range. In genetically identical sister cells, cell-to-cell variability in pulse dynamics influences the decision to enter S phase. While targeted inhibition of EGFR reduces the frequency of ERK activity pulses, inhibition of MEK reduces their amplitude. Continuous response curves measured in multiple cell lines reveal that proliferation is effectively silenced only when ERK pathway output falls below a threshold of ~10%, indicating that high-dose targeting of the pathway is necessary to achieve therapeutic efficacy.

A Compendium of Signals and Responses Triggered by Prodeath and Prosurvival Cytokines

Cell-signaling networks consist of proteins with a variety of functions (receptors, adaptor proteins, GTPases, kinases, proteases, and transcription factors) working together to control cell fate. Although much is known about the identities and biochemical activities of these signaling proteins, the ways in which they are combined into networks to process and transduce signals are poorly understood. Network-level understanding of signaling requires data on a wide variety of biochemical processes such as posttranslational modification, assembly of macromolecular complexes, enzymatic activity, and localization. No single method can gather such heterogeneous data in high throughput, and most studies of signal transduction therefore rely on series of small, discrete experiments. Inspired by the power of systematic datasets in genomics, we set out to build a systematic signaling dataset that would enable the construction of predictive models of cell-signaling networks. Here we describe the compilation and fusion of ∼10,000 signal and response measurements acquired from HT-29 cells treated with tumor necrosis factor-α, a proapoptotic cytokine, in combination with epidermal growth factor or insulin, two prosurvival growth factors. Nineteen protein signals were measured over a 24-h period using kinase activity assays, quantitative immunoblotting, and antibody microarrays. Four different measurements of apoptotic response were also collected by flow cytometry for each time course. Partial least squares regression models that relate signaling data to apoptotic response data reveal which aspects of compendium construction and analysis were important for the reproducibility, internal consistency, and accuracy of the fused set of signaling measurements. We conclude that it is possible to build self-consistent compendia of cell-signaling data that can be mined computationally to yield important insights into the control of mammalian cell responses.

Dose-dependent induction of distinct phenotypic responses to Notch pathway activation in mammary epithelial cells

Aberrant activation of Notch receptors has been implicated in breast cancer; however, the mechanisms contributing to Notch-dependent transformation remain elusive because Notch displays dichotomous functional activities, promoting both proliferation and growth arrest. We investigated the cellular basis for the heterogeneous responses to Notch pathway activation in 3D cultures of MCF-10A mammary epithelial cells. Expression of a constitutively active Notch-1 intracellular domain (NICD) was found to induce two distinct types of 3D structures: large, hyperproliferative structures and small, growth-arrested structures with reduced cell-to-matrix adhesion. Interestingly, we found that these heterogeneous phenotypes reflect differences in Notch pathway activation levels; high Notch activity caused down-regulation of multiple matrix-adhesion genes and inhibition of proliferation, whereas low Notch activity maintained matrix adhesion and provoked a strong hyperproliferative response. Moreover, microarray analyses implicated NICD-induced p63 down-regulation in loss of matrix adhesion. In addition, a reverse-phase protein array-based analysis and subsequent loss-of-function studies identified STAT3 as a dominant downstream mediator of the NICD-induced outgrowth. These results indicate that the phenotypic responses to Notch are determined by the dose of pathway activation; and this dose affects the balance between growth-stimulative and growth-suppressive effects. This unique feature of Notch signaling provides insights into mechanisms that contribute to the dichotomous effects of Notch during development and tumorigenesis.

Accurate information transmission through dynamic biochemical signaling networks

Stochasticity inherent to biochemical reactions (intrinsic noise) and variability in cellular states (extrinsic noise) degrade information transmitted through signaling networks. We analyzed the ability of temporal signal modulation—that is, dynamics—to reduce noise-induced information loss. In the extracellular signal–regulated kinase (ERK), calcium (Ca2+), and nuclear factor kappa-B (NF-κB) pathways, response dynamics resulted in significantly greater information transmission capacities compared to nondynamic responses. Theoretical analysis demonstrated that signaling dynamics has a key role in overcoming extrinsic noise. Experimental measurements of information transmission in the ERK network under varying signal-to-noise levels confirmed our predictions and showed that signaling dynamics mitigate, and can potentially eliminate, extrinsic noise–induced information loss. By curbing the information-degrading effects of cell-to-cell variability, dynamic responses substantially increase the accuracy of biochemical signaling networks. Signaling dynamics increase the information transmission capacity of biochemical signaling pathways. Dynamic signals enhance information transfer Cells need to process information about their external environment reliably to survive. However, variation, or noise, in biochemical reactions, or in the states of individual cells, make it hard for a cell to detect concentration, rather than just the presence or absence of an activating ligand. Selimkhanov et al. show that cellular signaling circuits get around this problem by continually monitoring signals over time. Such dynamic responses in cultured human cells more effectively distinguish signals from noise and thus avoid loss of information transmitted to the cell from external signals. Science, this issue p. 1370

Cue-Signal-Response Analysis of TNF-Induced Apoptosis by Partial Least Squares Regression of Dynamic Multivariate Data

Biological signaling networks process extracellular cues to control important cell decisions such as death-survival, growth-quiescence, and proliferation-differentiation. After receptor activation, intracellular signaling proteins change in abundance, modification state, and enzymatic activity. Many of the proteins in signaling networks have been identified, but it is not known how signaling molecules work together to control cell decisions. To begin to address this issue, we report the use of partial least squares regression as an analytical method to glean signal-response relationships from heterogeneous multivariate signaling data collected from HT-29 human colon carcinoma cells stimulated to undergo programmed cell death. By partial least squares modeling, we relate dynamic and quantitative measurements of 20-30 intracellular signals to cell survival after treatment with tumor necrosis factor alpha (a death factor) and insulin (a survival factor). We find that partial least squares models can distinguish highly informative signals from redundant uninformative signals to generate a reduced model that retains key signaling features and signal-response relationships. In these models, measurements of biochemical characteristics, based on very different techniques (Western blots, kinase assays, etc.), are grouped together as covariates, showing that heterogenous data have been effectively fused. Importantly, informative protein predictors of cell responses are always multivariate, demonstrating the multicomponent nature of the decision process.

Collecting and organizing systematic sets of protein data

Systems biology, particularly of mammalian cells, is data starved. However, technologies are now in place to obtain rich data, in a form suitable for model construction and validation, that describes the activities, states and locations of cell-signalling molecules. The key is to use several measurement technologies simultaneously and, recognizing each of their limits, to assemble a self-consistent compendium of systematic data.