Imaging and Defining Emergent Behaviors of the Immune Response

Immune Sensitivity

Why Study the Dynamics of the Immune Synapse?

We’ve been studying, since the lab’s inception, how T cells physically interact with and detect opposing surfaces. In order for T cells to recognize foreign material, they use proteins called T cell receptors (TCR) to ‘touch’ pieces of that material (peptides) that are displayed on proteins (MHC) on the surfaces of antigen-presenting cells. This is the critical moment in immune recognition—the point at which a cell detects the foreign material and can begin the process of cell division and gene-expression that allows it to respond.

In Cai et al Science 2017, we demonstrated that T cells actively extend, retract and move their microvilli in order to efficiently scan opposing surfaces. The movement allows them to scan an entire surface in about 1 minute despite the observed fact that they only touch a given site on the opposing cell for about 5 seconds. That 5-second contact makes us rethink T cell antigen-detection since that timing means that receptors must coalesce on peptide-MHC complexes in that very short period. We also found that TCR engagement solidifies the microvillar projection, seemingly ‘gluing’ the contact in place. That this did not require ZAP-70 signaling or the actin cytoskeleton implies that the TCR ‘locks-in’ synaptic contacts without the need for cellular signaling.


Tracking T cell receptors and co-receptors during engagement

As a postdoc and just prior to starting our lab, I made the first fusions between GFP and TCR components and expressed them in live T cells. This showed differential movement patterns of CD3 and CD4 components. By simultaneously visualizing calcium levels in T cells, we were also able to show that signaling onset occurred prior to cSMAC formation and in a phase characterized by submicron cluster formation.  (Krummel et al. Science 2000)

We subsequently assessed costimulatory ligands CD28 (Andres et al. 2004) and CD40L/CD40 pair (Boisvert et al 2004) and found that both are recruited to the immunological synapse coincident with when clusters of TCR were forming but with different kinetics: CD28 appears in the earliest phases of our observations whereas CD40L/CD40 arrive after and likely as a consequence of the polarization induced by TCR signaling.

In Moss et al. PNAS 2002, we adapted segmentation algorithms and 3D quantization approaches to determine the velocities of membranes and receptors during the first moments of T-APC contact. This led us to be able to conclude that receptor movement was active, approximately 0.1-0.2um/min and not likely achieved via diffusion alone.  It also revealed a membrane wave that initiates during the T-APC contact.  The methods for mining data out of 3D datasets have been repeated in other systems by other groups and applied to similar approaches of membrane deformation.

To define TCR dynamics directly, as opposed to simply the associated CD3 chains, and as a means to analyze receptor dynamics in unmanipulated naïve T cells, we generated TCR transgenic mice expressing only T cells with a variant of the ovalbumin (OVA) reactive in which the TCR alpha chain is fused to eGFP (Friedman et al. JEM 2010). Direct imaging of TCRs during their interactions with dendritic cells in this context revealed that the most reliable common theme for responses to T cell stimuli was the observation of rapid TCR-GFP internalization—unlike with CD3-GFP, the TCR-GFP appears to persist inside these cells, allowing us to continue to track the receptors. TCRs when observed in these mice, appear to become flexibly arrayed into the synapse, often without forming a centralized ‘cSMAC’ at all but through a process of apparent sequential receptor triggering and internalization. Synapse formed by these cells also often continue to actively move whilst signaling continued motility, largely divorcing motility arrest as a pre-requisite for signaling; a feature of activation of T cells in situ highly suggestedfor cells in lymph nodes by Cahalan and van Andrian’s labs. Using the TCR-GFP T cells directly in lymph nodes, we were then able to study TCR dynamics in vivo through optimization of our 2-photon detection and signal processing. Using the readout of internalized TCR-GFP vesicles as a readout of T cell activation, we once again observed rapid internalization in the absence of either motility arrest or evident cSMAC formation.

This work was extended in Beemiller et al Nature Immunology 2012 in which we demonstrated the concurrence of signaling T cell receptor microclusters on the T cell surface and how this type of signaling can concur in time with ongoing motility as a transient synapse is formed. We also demonstrated how actin movements are utlized to coordinate these two seemingly-disparate activities.  Using a life-act probe, we also demonstrated that formation of a cSMAC requires an active actin depolymerization at the center of the synapse.

The sum of these studies to date suggests that assembly of TCR clusters does not rely on a ‘stable’ cSMAC and can take place against a background of motility. The details of how signaling clusters initially organize or are perpetuated in this setting is yet to be understood.


Other Cytoskeletal Players: 

We have also used image-based screening to isolate gene-products important for T cell function as well as for motility (see Tooley et al. Seminars in Immunology. 2005. In Tooley, Gilden et al. Nature Cell Biology. 2008, our lab was the first to identify the Septin cytoskeleton as an important player in cortical tension and for effective motility.  While septin fibers appear to form a corset around the midbody of T cells and thus reinforce this zone, the septin cytoskeleton also appears to reinforce cortical rigidity and tension.  This results in two distinctive phenotypes: First, a ‘floppiness’ that results in an elongated uropod that does not track the cell body and the second is an ability to generate or maintain excess protrusions.  We hypothesize that this latter feature accounts for our demonstration that septin-null T cells had gained the ability to transit very constrictive barriers, a feature perhaps quite important both for immune tissue surveillance and in cancer metastasis.  When we then made septin knockouts, we found that these proteins are also specifically required for cell division when T cells are no anchored to other cells. Cell division while in contact with an APC was septin-independent.  We hypothesize that specific septin inhibitors would block cell division to soluble cytokines in vivo while sparing responses mediated by synaptic engagements.


Krummel Lab © 2018