(683g) Systems Analysis of Components of PI3K/AKT Pathway Involved in Maintenance of Self-Renewal State of Human Embryonic Stem Cells

Introduction: Differentiation of human embryonic stem cells (hESCs) to the definitive endoderm lineage is a critical first step towards deriving important functional cell types of the pancreas and liver. The process of lineage specific differentiation is under the control of complex and non-linearly interacting parallel signaling pathways. The complexity of these interactions restricts a purely experimental analysis and necessitates a systems level mathematical approach. Recent investigations into the signaling pathways have identified the key role of insulin mediated PI3K/AKT pathway in maintaining self-renewal and influencing differentiation [1]. It is established that the levels of the molecule p-AKT correlate well to the state of self-renewal (pAKT ON) and differentiation (pAKT OFF). However most of these studies analyze key molecules of the pathway in isolation. What is still lacking is a systems level understanding of the regulatory interactions of molecules as a component of a pathway. In particular, the PI3K pathway involves several regulatory modules; namely (i) insulin receptor activation and trafficking in the endosomes, (ii) receptor mediated activation of intracellular IRS1/PI3K/AKT, (iii) negative feedback by serine phosphorylation of IRS1 by molecules like PKC-ζ (iv) negative regulators like PTEN, SHIP and PTP and (iv) positive feedback by AKT. In the current work, we have analyzed the dynamics of the pathway and the sensitivity of the pathway to the various regulatory interactions to identify the mechanism of maintenance of p-AKT levels in self-renewal and its implications during differentiation.

Materials and Methods: The dynamics of the key components of the PI3K/AKT pathway in hESCs was first analyzed by experimentally stimulating H1 cells with 100 nM insulin after growth factor starvation. The dynamics was then compared to in silico simulations using a detailed mechanistic differential equation model of insulin mediated PI3K/AKT pathway (20 state variables, 25 parameters in the current form) by Sedaghat et al [2]. Using clustering techniques, the dynamics predicted by the model was compared with the experimental profiles to identify the feasible parameter ranges. The steady-states of the model were further analyzed using global sensitivity analysis (GSA) to identify the key interactions in the pathway that affect the levels of molecules like p-AKT. We utilized a computationally efficient algorithm called Random Sampling High Dimensional Model Representation (RS-HDMR) to capture sensitivity under non-linear interactions between the large numbers of model parameters. The results from GSA were validated in hESCs using small molecule inhibitors of the sensitive nodes in the self-renewal state. Finally, we combined the validated GSA results with global parameter optimization using evolutionary strategies to track the pathway dynamics in hESCs. These methods allowed us to capture detailed dynamics and steady-states of the pathway nodes as well as the sensitive biochemical processes that are involved in self-renewal and differentiation.

Results and Discussion:We analyzed the dynamics and steady states of four major nodes of the PI3K/AKT pathway. These include the active insulin receptors (p-IR), tyrosine IRS1 (p-IRS1 (Y)), serine IRS1 (p-IRS1 (S)) and p-AKT. Dynamics from insulin stimulation experiments showed that intracellular components of PI3K/AKT in hESCs show the typical overshoot with steady-state behavior influenced by negative feedback by p-IRS1 (S). From a detailed rate parameter based GSA of the steady-states, we identified that many of the negative regulators of the pathway sensitized the pathway more than the positive regulators. These processes kept p-AKT levels considerably low in the explored high dimensional parameter space. We validated the influence of negative regulators in hESCs where it was found that inhibition of negative feedback by p-PKC-ζ increased p-AKT levels considerably and PTEN and PTP inhibition had a secondary effect. During parameter optimization, we found parameter ensembles that favor a low rate of negative regulation by PTEN and PTP but simultaneously favor a strong negative feedback by p-IRS1 (S) to best explain the hESC dynamics. Thus our analysis allowed us to identify the mechanism that regulates p-AKT levels in hESCs during self-renewal and which should be down-regulated during differentiation.

Conclusions: Our detailed mechanistic model and experimental analysis identified the mechanism of control of p-AKT levels in hESCs under self-renewal. During self-renewal, negative feedback by pIRS1 (S) controls p-AKT levels at a sufficient threshold. Therefore, our work has identified an important role for negative feedback to maintain the PI3K/AKT pathway in hESCs under homeostatic conditions which was previously unappreciated. We are currently investigating the role of this negative feedback in the context of cross-talks with TGF-β induced endoderm differentiation of hESCs.

References:

1.  Singh et al.Cell Stem Cell (2012) 10: 312-326

2.  Sedaghat et al.American J of Physiol-Endocrinol and Metabol (2002) 283: E1084-E1101