Review: Bioengineering strategies to probe T cell mechanobiology

Abstract

T cells play a major role in adaptive immune response, and T cell dysfunction can lead to the progression of several diseases that are often associated with changes in the mechanical properties of tissues. However, the concept that mechanical forces play a vital role in T cell activation and signaling is relatively new. The endogenous T cell microenvironment is highly complex and dynamic, involving multiple, simultaneous cell-cell and cell-matrix interactions. This native complexity has made it a challenge to isolate the effects of mechanical stimuli on T cell activation. In response, researchers have begun developing engineered platforms that recapitulate key aspects of the native microenvironment to dissect these complex interactions in order to gain a better understanding of T cell mechanotransduction. In this review, we first describe some of the unique characteristics of T cells and the mounting research that has shown they are mechanosensitive. We then detail the specific bioengineering strategies that have been used to date to measure and perturb the mechanical forces at play during T cell activation. In addition, we look at engineering strategies that have been used successfully in mechanotransduction studies for other cell types and describe adaptations that may make them suitable for use with T cells. These engineering strategies can be classified as 2D, so-called 2.5D, or 3D culture systems. In the future, findings from this emerging field will lead to an optimization of culture environments for T cell expansion and the development of new T cell immunotherapies for cancer and other immune diseases.

Full text

Review: Bioengineering strategies to probe T cell
mechanobiology
Adi de la Zerda,
1
Michael J. Kratochvil,
1,2
Nicholas A. Suhar,
1
and
Sarah C. Heilshorn
1,a)
1
Department of Materials Science and Engineering, Stanford University, Stanford,
California 94305, USA
2
Department of Medicine, Division of Infectious Disease, Stanford University, Stanford,
California 94305, USA
(Received 26 September 2017; accepted 29 January 2018; published online 29 March 2018)
T cells play a major role in adaptive immune response, and T cell dysfunction can
lead to the progression of several diseases that are often associated with changes in
the mechanical properties of tissues. However, the concept that mechanical forces
play a vital role in T cell activation and signaling is relatively new. The
endogenous T cell microenvironment is highly complex and dynamic, involving
multiple, simultaneous cell-cell and cell-matrix interactions. This native complex-
ity has made it a challenge to isolate the effects of mechanical stimuli on T cell
activation. In response, researchers have begun developing engineered platforms
that recapitulate key aspects of the native microenvironment to dissect these com-
plex interactions in order to gain a better understanding of T cell mechanotransduc-
tion. In this review, we first describe some of the unique characteristics of T cells
and the mounting research that has shown they are mechanosensitive. We then
detail the specific bioengineering strategies that have been used to date to measure
and perturb the mechanical forces at play during T cell activation. In addition, we
look at engineering strategies that have been used successfully in mechanotrans-
duction studies for other cell types and describe adaptations that may make them
suitable for use with T cells. These engineering strategies can be classified as 2D,
so-called 2.5D, or 3D culture systems. In the future, findings from this emerging
field will lead to an optimization of culture environments for T cell expansion and
the development of new T cell immunotherapies for cancer and other immune
diseases. V
C2018 Author(s). All article content, except where otherwise noted, is
licensed under a Creative Commons Attribution (CC BY) license (http://
creativecommons.org/licenses/by/4.0/).https://doi.org/10.1063/1.5006599
I. INTRODUCTION
In recent years, the field of mechanobiology and how forces influence the behavior of cells
and tissues has become an important area of research. Recent data showing a link between
mechanical signaling and the pathogenesis of several disorders highlight the significance of
understanding how tissue mechanics convert into biochemical signals,
1
an understanding of
which may elucidate a greater knowledge of disease progression. For a number of years,
mechanical degradation of tissues was thought to be a symptom of disease. However, now there
is a growing shift in the field that instead views abnormalities in tissue mechanics and dysfunc-
tional mechanotransduction as not the end result, but rather significant contributors to disease
progression. One example is breast cancer, where it has been shown that an increase in tissue
stiffness promotes metastasis in vitro and in vivo and where there is active research about the
use of T cells with improved activity to inhibit this malignancy.
2
Additionally, several studies
have reported that tissue mechanics are significantly altered in inflamed organs. Inflamed organs
a)
E-mail: Heilshorn@stanford.edu
2473-2877/2018/2(2)/021501/27 V
CAuthor(s) 2018.2, 021501-1
APL BIOENGINEERING 2, 021501 (2018)

can result from either injury, infection, or autoimmune reaction,
3
and since T cells participate
in many of these inflammatory responses, T cell mechanobiology has become an intense area
of research as well.
T cell function in a highly complex and dynamic mechanical microenvironment in which
they undergo cell-cell and cell-matrix interactions, all of which may affect T cell mechanotrans-
duction and the resulting activation responses [Fig. 1(a)]. As T cells circulate throughout the
body to locate antigen presenting cells (APCs), they come into contact with differing microen-
vironments that have varied topography and mechanical stiffness [Fig. 1(b)].
4,5
Simultaneously,
the T cell is processing highly complex interactions with one or more APCs, which also pro-
vide multiple independent mechanical stimuli for any one T cell. When a T cell encounters an
APC, it forms an immunological synapse (IS) that connects the APC’s peptide-major histocom-
patability complex (pMHC) with the T cell receptor (TCR). At the site of the IS, the T cell
changes its morphology to form invadosome-like protrusions that physically push against and
probe the membrane of the APC. The T cell’s ability to exert force on the APC membrane dur-
ing this interaction is critical for T cell activation,
8
as T cells that are unable to exert forces on
the APC have a defective activation response.
9
Another layer of complexity to this interaction
is that the APC’s membrane rigidity dynamically changes in response to cues from inflamma-
tion and the IS,
10,11
while simultaneously the activated T cell’s membrane rigidity also changes
and becomes more compliant.
12
These changes in membrane rigidity may reflect the T cell’s
FIG. 1. Microenvironmental cues that may impact T cell mechanotransduction. (a) Biophysical/biomechanical factors
affecting T cell mechanotransduction. The description begins with the top panel and moves clockwise: T cells encounter a
wide range of microenvironments in the body with a diversity of matrix stiffness (e.g., Young’s modulus, E) values ranging
from 10 to 10
6
Pa and matrix topography that result in cell surface curvature. T cells also encounter differences in APC
stiffness, since the APC membrane rigidity dynamically changes in response to cues from inflammation. The T cell stiffness
itself is also dynamic, with the membrane rigidity changing during immunological synapse (IS) formation and T cell activa-
tion. A single T cell can simultaneously form multiple contacts with several APCs. Before forming a stable IS, the T cell
can sequentially encounter different APCs for brief periods of time, leading to serial contact with cells of differing mechan-
ical properties. (b) Stiffness of common tissues and organs where T cell actions take place. Stiffness is presented in
Young’s modulus in kPa.
5–7
021501-2 de la Zerda et al. APL Bioeng. 2, 021501 (2018)

ability to sense and respond to fluctuating mechanical cues while simultaneously being acti-
vated by the APC. Finally, another dimension to consider is that a single T cell may simulta-
neously interact with multiple APCs
13
as well as sequentially encounter different APCs for brief
periods of time, both of which bring with it a number of other mechanical stimulants that may
affect T cell behavior. As an example of when this may occur, in the case of a pMHC complex
having a weak affinity to the TCR, several APC encounters are necessary in order to reach a
critical activation threshold.
14
These latter behaviors in particular, impose a significant chal-
lenge to researchers trying to dissect the roles of mechanical cues on T cell activation.
Reductionist approaches try to reduce the complexity of T cell interaction with the micro-
environment to enable quantitative biology. Bioengineers employ different strategies to dissect
the cellular responses to mechanical cues by creating natural and synthetic platforms that per-
turb and/or quantify T cell mechanobiology. These strategies have led to the relatively new dis-
covery that T cells are mechanosensitive and the identification of the TCR as a module where
force generation can occur.
15–17
T cells are constantly exerting forces and being submitted to forces, both during their
migration and while interacting with their cognate APC. Recently, a study revealed that T cells
leverage these mechanical forces to aid in cytotoxic activity against target cells.
9
Other key
results revealed that T cells can modify their growth and proliferation based on different sub-
strate stiffness and substrate topologies.
9,18–21
Several studies have also suggested that extracel-
lular mechanical forces can facilitate activation of surface receptors
22
and that pMHC complex
recognition may be mediated by mechanical cues from both the APC and the extracellular
matrix (ECM),
23
however to date, the only mechanoreceptor for T cells that has been experi-
mentally validated is the TCR, and the downstream signaling mechanisms remain unknown.
24
In addition, many findings on how mechanics influence T cell activation are potentially contra-
dictory, which may be attributed to the non-physiological nature of most systems used to date.
Overcoming this gap in knowledge will require the efforts of biologists, engineers, and clini-
cians alike to develop techniques that are physiologically relevant and reductionist to character-
ize the underlying features in T cells responsible for these mechanotransductive pathways.
Studying these mechanotransduction pathways and the mechanics that affect T cell function
within a living host is not just of interest to the scientific community, but signifies a research
area that is fundamental to understanding immune response and a necessary step in developing
novel therapeutic strategies. Studies report that tissue rigidity changes during the course of dis-
ease.
1,3
For example, most people are familiar with the experience of having a physician touch
their lymph nodes to detect the perceived stiffness, as this is often correlated with inflammation
and malignancy. In scientific studies, the stiffness of the lymph nodes has been reported to be
in the range of 120 Pa to 1 kPa, as detected by a variety of methods including shear-wave ultra-
sound elastography and a tactile sensor.
3,5,6
By elucidating how mechanics influence T cell acti-
vation, we may be able to identify how to both encourage T cell activation for fighting infec-
tions and cancers as well as to suppress T cell activation for controlling autoimmune disease.
Local alteration of tissue stiffness by drugs may be used to manipulate our body’s natural
defense system to be more effective and help thwart diseases where T cells are too active or
not active enough. T cell mechanobiology will also have practical application in T cell immu-
notherapy. T cell immunotherapy encompasses taking the patient’s T cells out of the body,
reprogramming them to attack cancer cells, and then expanding them ex vivo before injecting
them back into the patient. Optimizing the mechanics of the ex vivo culture to achieve an ade-
quate T cell expansion is essential for immunotherapy, and especially important in the case of
leukemia patients who have very few T cells available in their blood.
25
In this review, we describe some of the unique characteristics of T cells in Sec. II. Section
III is presented as a tutorial with a specific focus on experimental design choices to activate T
cells in engineered culture systems. We then describe the mounting research that has shown
that T cells are mechanosensitive, beginning with studies employing T cells in 2D culture sys-
tems. This is followed with a description of engineering strategies to perturb and quantify T
cell forces in the so-called 2.5D and 3D culture systems. Finally, we look at potential future
techniques that could be used to study T cell mechanotransduction and how the findings from
021501-3 de la Zerda et al. APL Bioeng. 2, 021501 (2018)

this emerging field may lead to an optimization of culture environments for T cell expansion
and an overall greater immunotherapeutic potential for cultured T cells.
II. INTRODUCTION TO T CELL BIOLOGY
Sections II A–II D are written as a quick introduction for those who are not familiar with T
cell biology. This brief tutorial is not meant to be comprehensive, but rather focuses on topics
that may be critical to consider when evaluating past T cell mechanotransduction studies and/or
designing future mechanotransduction studies. In this section, we discuss the role of T cells in
the body, their life cycle and a description of the processes following antigen recognition. For
the interested reader, many excellent review papers providing greater depth on T cell immunol-
ogy are available elsewhere.
15,26
A. Adaptive immunity and autoimmunity
T cells, also commonly referred to as T lymphocytes, are key players in the adaptive
immune system and are responsible for triggering a host response in the presence of antigen
presenting pathogens and tumor cells. Since these cells are responsible for initiating the
immune cascade, it is critical that they are able to accurately distinguish between self and non-
self for efficient self-defense. In the absence of this distinction, T cells may incite an immune
response against the host resulting in an “autoimmune disease” wherein the body begins to
attack its own cells, as is the case in diseases such as type 1 diabetes (T1D) and multiple scle-
rosis (MS).
B. T cell subsets: CD4
1
and CD8
1
There are two major subsets of T cells: CD4
þ
and CD8
þ
, which are distinguished based on
the type of major histocompatability complex (MHC) that the T cell recognizes. MHC mole-
cules are displayed on the surface of the target cell and can be categorized as either MHC class
I or II. CD8
þ
T cells, also known as cytotoxic T lymphocytes (CTLs), recognize MHC class I,
which is displayed on all nucleated cells in the body. Once a target cell is identified, CTLs
bind to the target cell and induce apoptosis by releasing lytic granules containing the toxic pro-
teins perforin and granzyme, which bore pores in the lipid bilayer of the target cell.
Alternatively, CD4
þ
T cells recognize MHC class II, which is expressed by specialized immune
cells called APCs. In general, CD4
þ
T cells are tasked with activating other cells of the
immune system. Their functions involve helping B cells to produce antibodies, inducing macro-
phages to enhance their microbicidal activity, and recruiting other types of immune cells such
as neutrophils to an inflammation site. Because of these “assistor” functions, they are also
referred to as T helper cells. CD4
þ
T cells can further differentiate into subsets, with the four
most prevalent being type 1 T helper cells (Th
1
), type 2 T helper cells (Th
2
), T follicular helper
cells (T
fh
), and type 17 T helper cells (Th
17
). These subsets also have specialized functions. For
example, Th
1
responds to infections caused by intracellular bacteria, whereas Th
2
responds to
extracellular parasites. The discussion here will focus on Th
1
, whose response to inflammation
is well documented and whose mechanotransduction pathways are the most understood. When
Th
1
cells identify a target cell expressing an antigen cognate to their TCR, they bind to it and
start secreting cell-signaling proteins known as cytokines. Th
1
cells typically express interferon-
gamma (IFN-c), interleukin 2 (IL-2), and tumor necrosis factor (TNF), which orchestrate the
immune cell-mediated response.
C. Life cycle of the T cell
The life cycle of T cells begins in the thymus where they differentiate into either the
CD4
þ
or CD8
þ
subsets. Afterwards, the T cells migrate to the secondary lymphoid tissues
[e.g., lymph nodes (LNs)] where they are activated after encountering their cognate antigen,
expand, and finally differentiate into either an effector subset (i.e., Th
1
,Th
2
,T
fh
,Th
17
, and
effector CD8
þ
) or a memory cell. A T cell identifies a particular APC using the TCRs
021501-4 de la Zerda et al. APL Bioeng. 2, 021501 (2018)

presented on its surface that identifies cognate antigen-peptides coupled to MHC (pMHC).
Immediately after the TCR identifies the pMHC, the T cell binds to the APC and an activation
process is triggered. The activation process leads to formation of a stable contact with the APC
and initiates a cascade of events that includes TCR phosphorylation, cytoskeletal reorganization,
Ca
2þ
influx, and cytokine production. After approximately 3–4 days of contact with the APC,
these effector T cells leave the LNs and travel to the site of infection to further orchestrate the
immune response.
D. Immunological synapse
To understand how physiological T cell activation occurs, we first need to familiarize our-
selves with the TCR module and CD28 costimulator ligands (Fig. 2). Costimulator ligation is
essential for T cell activation, as TCR stimulation without costimulation will lead to cell unre-
sponsiveness, or anergy. The TCR module is a transmembrane complex consisting of CD3 pro-
tein subunits, sometimes denoted as the TCR/CD3 complex. The intracellular component of the
CD3 contains immunoreceptor tyrosine-based activation motifs (ITAMs). Once a TCR is ligated
by a pMHC, the lymphocyte-specific protein tyrosine kinase (LCK) is activated and simulta-
neously initiates two signaling cascades. First, LCK phosphorylates the CD3 ITAMs, which cre-
ate a docking site for Zap70, a protein critical for T cell activation. Zap70 is recruited to the
docking site and activated to phosphorylate the cytoplasmic segment of the adaptor protein
linker for activation of T cells (LAT), which in turn controls signal amplification and diversifi-
cation downstream of the TCR.
27
In the second signaling cascade, LCK phosphorylates the
cytoplasmic tail of the costimulator protein CD28, a critical step for proper functionality of the
CD28 surface receptor.
26,28
Following TCR-pMHC binding, CD28 stimulation through APC receptor CD80 is the sec-
ond signal required for activation. Stimulating the TCR and CD28 will direct the T cell to form
an IS with the APC.
30
In contrast, ligation of CD28 alone will lead to the induction of inhibi-
tory signals in T cells, and TCR binding alone results in either apoptosis or a state of anergy.
31
T cell polarization and IS maturation begins about 5–10 min after the IS has formed. This pro-
cess includes T cell spreading across the APC and the formation of micrometer-scale clusters
of a variety of cell receptors, including TCR and CD28. The TCR/CD3 complexes accumulate
at the center of the IS to form the central supramolecular activation complex (cSMAC), while
microclusters of lymphocyte function-associated antigen 1 (LFA-1) that promotes cell adhesion
enclose it, creating the peripheral SMAC (pSMAC)
32
[Fig. 2(b)]. As these receptor-
microclusters migrate across the IS, their corresponding counterpart receptors on the APC sur-
face [pMHC and intercellular adhesion molecule-1(ICAM-1)] move in a complementary man-
ner. This surface ligand mobility on the APC can be an important experimental design parame-
ter for studying how mechanical cues affect T cell activation.
III. EXPERIMENTAL METHODS TO ACTIVATE T CELLS AND STUDY THEIR RESPONSE
OUTCOMES OVER TIME
Sections III A and III B are written as a brief tutorial for those who are unfamiliar with
in vitro methods of T cell activation, particularly focusing on stimulation approaches and the
resulting time line of cell responses. The design of the experimental setup directly affects the T
cell activation response and needs to be carefully planned depending on the research hypothe-
sis. Thus, this section serves to present key concepts critical for understanding current research
done in the field of T cell mechanobiology.
A. Current methods to activate T cells
The mechanobiology of T cell activation is an increasing area of scientific research. A key
requirement for any experiment in this field is the ability to induce T cell activation in a repro-
ducible manner. Using natural APCs [i.e., macrophages, B cells, or dendritic cells (DCs)] in
combination with their cognate antigen for T cell stimulation is the most physiologically
021501-5 de la Zerda et al. APL Bioeng. 2, 021501 (2018)

FIG. 2. TCR downstream signaling and cytoskeletal reorganization. (a) Structure of the TCR module and early
downstream signaling. Upon TCR recognition of an antigenic peptide loaded onto MHC (pMHC), phosphorylation
of CD3 ITAMs (open blue circles) by the protein tyrosine kinase LCK leads to the recruitment and activation of
ZAP70, which in turn phosphorylates tyrosine residues (filled blue circles) found in the cytoplasmic segment of the
linker for activation of T cells (LAT), amplifying and diversifying the seminal signal. The CD28 costimulator recog-
nizes CD80 or CD86 ligands at the surface of the APC. (b) Cytoskeletal reorganization following TCR stimulation.
Following TCR stimulation, filamentous (F)-actin polymerization is induced at the IS, and the T cells microtubule-
organizing center (MTOC) is polarized. A mature IS has a typical bull’s-eye pattern consisting of concentric rings
of membrane receptors: the inner circle, the central supramolecular activation cluster (cSMAC), and the peripheral
supramolecular activation cluster (pSMAC). This pattern will occur when Th
1
T cells contact B cells, tumor
cells, and artificial APCs (aAPCs), but not when contracting dendritic cells (DCs).
29
At the opposite pole to the IS,
a less well understood protein complex called the Distal-Pole Complex (DPC) is formed. The DPC consists of the
cell-surface receptor CD43 and also involves F-actin polarization to the rear side of the T cells. (c) In vitro T
cell activation. Only two signals are needed to activate T cells artificially in vitro, TCR and CD28 (costimulator)
stimulation. Substrates (e.g., beads, gel surfaces) can be functionalized with stimulatory antibodies for these
receptors.
021501-6 de la Zerda et al. APL Bioeng. 2, 021501 (2018)

relevant approach to emulate native conditions. However, practical challenges of this approach
include the typically low yield of isolated cells using laborious, expensive processes requiring
specialized equipment. Another major challenge with cell-mediated activation is the large vari-
ety of cellular subsets. Because different cell subsets can modify the kinetics and the magnitude
of T cell activation responses, isolated APCs must be carefully analyzed and sorted to achieve
reproducible, quantitative data.
33
As a result, simplified approaches to activate T cells are often
currently used in laboratories [Fig. 2(c)].
Depending on the activation method, T cells may proliferate at different rates, secrete dif-
ferent cytokine repertoires, and release these cytokines with unique temporal profiles.
34–36
Therefore, choosing the appropriate activation method is crucial for the design of the experi-
mental set-up. Here, we will discuss some of the current methods to stimulate na€
ıve T cells and
how they differ in terms of their physiological relevance, required stimulation time, TCR
dependency, and antigen specificity.
1. Phorbol myristate acetate (PMA) and ionomycin (Iono)
PMA and Iono are small organic molecules that diffuse into the cytoplasm through the cell
membrane. When used together they directly activate protein kinase C (PKC) and raise the
intracellular level of Ca
2þ
, which triggers the calcium release required for the nuclear factor of
activated T-cells (NFAT) signaling. Five hours of stimulation with these two chemicals is
enough to achieve complete T cell activation and to induce sustained production of IFN-c, IL-
2, and IL-4.
35,37
However, because this method completely bypasses TCR stimulation and sig-
naling, this method is nonphysiological, not TCR dependent, and not antigen specific.
Furthermore, this method upregulates the Fas ligand, which is involved with cell death in T
cells, and thus is toxic to them over long incubation times.
34,38
While small molecule activation
with PMA is a type of purely chemical activation, all other methods of T cell activation (dis-
cussed later) include both a biochemical and a potential biomechanical component, since
receptor-ligand binding interactions are involved.
2. Artificial APC-mimicking interfaces
Two signals are necessary for in vitro T cell activation by artificial APC (aAPC): first TCR
engagement and then binding of the costimulatory receptor [Fig. 2(c)]. A variety of strategies
have been developed to attempt synapse formation in 2D and 3D utilizing aAPCsurfaces,
including the use of antibodies or pMHC tetramers on 2D or 3D surfaces.
39
For 3D strategies,
the surfaces can be either non-living or cell-based.
40,41
Below is a brief compilation of the key
features, advantages and limitation for each of these aAPC systems, which have been presented
and discussed in greater detail elsewhere.
41,42
The 2D and 3D surfaces of aAPCs can be designed to precisely fine-tune activation signal
strength by modulating several parameters: the TCR-ligand and costimulatory-ligand surface
density, ligand affinity, and the cytokine milieu.
43
A common strategy utilizes monoclonal anti-
bodies (mAbs) and pMHC-tetramers as stimulating ligands for the TCR signaling pathway.
44
Specifically, anti-CD3 and anti-CD28 mAbs are used to artificially mimic TCR-dependent acti-
vation by artificially aggregating the receptors in the membrane and binding the CD3esignaling
subunits. They deliver a much stronger signal than the physiological cognate ligand pMHC and
lead to a robust polyclonal activation response.
40
One reason for this particularly strong activa-
tion response can be attributed to anti-CD28 mAbs lack of ability to bind to the inhibitory
receptor CTLA-4 that negatively regulates T cell activation. Comparatively, the pMHC-tetramer
is antigenic peptide specific, thus more physiologically relevant.
Both of these methods induce cytokine burst-release profiles
35
and can be used for long
incubation periods. Interestingly, anti-CD3 and anti-CD28 mAbs conjugated to the surface of
3D beads induce greater levels of cytokine secretion than the corresponding 2D plate-bound
aAPC.
45
This result was hypothesized to be due to the larger surface area that the geometry of
the bead offers compared to the 2D plate. Geometry is a critical component in designing the
shape of the aAPC to improve T cell activation. The increased aspect ratio of ellipsoidal
021501-7 de la Zerda et al. APL Bioeng. 2, 021501 (2018)

microparticles conjugated with anti-CD28 mAbs and pMHC with particle volume and antigen
content held constant also resulted in enhanced T cell activation over the comparative spherical
aAPC. The long axis of the ellipsoidal aAPC led to increased interaction with CD8
þ
T cells.
46
A main limitation of most 2D and 3D aAPCs is that they are static, and thus do not mimic the
changes in geometry, ligand configuration, and surface stiffness that naturally occur in APCs
during IS formation.
11
Importantly, the surface rigidity of the aAPC has been shown to mediate
the amount of force that T cells can exert.
47
This effectively limits the dynamic mechanical
feedback experienced by the T cells, which is likely a critical component of this mechanotrans-
ductive process.
Cell-based aAPCs are engineered cell lines transfected using a retrovirus or lentivirus to
express the pMHC and the costimulator receptor. Notable examples include the K562 human
erythromyeloid line
48
and the murine NIH/3T3fibroblast line.
49
Unlike most static, rigid
acellular-aAPCs, these cell-based aAPCs can have dynamic responses and mechanical proper-
ties that are more similar to the native APC. However, because of their inherent complexity,
cell-based aAPCs have not yet been used in the context of mechanotransduction studies and are
reviewed elsewhere.
40,41,50
B. Activation response outcomes
The activation response manifests itself in two phases: early and late activation, the time-
line of which is described graphically in Fig. 3(a). Sections III B 1 and III B 2 will discuss fea-
tures and time-points characteristic of the activation response with a focus on those that can be
observed and quantified in engineered systems.
1. Early activation
a. Migration. Antigen-specific T cells are very rare, and dendritic cells (DCs) maintain a
low frequency (approximately 1% of the LN cells), as a result T cell motility plays a key role
in locating antigens from APCs and targeting cells. T cells can modify their motility patterns
based on environmental cues and current level of activation.
51
For example, increasing T cell
motility increases the probability of encountering a target cell and triggering TCR, whereas
decreasing motility allows for the formation of a more stable IS with a given target call. As
FIG. 3. T cell activation. (a) Outcomes and time course of T cell activation. Cytoplasmic calcium concentration will
increase within seconds after IS formation; T cell polarization and IS maturation begins about 5–10 min after the IS has
formed. This process includes T cell spreading across the APC and the formation of micrometer-scale clusters of a variety
of cell receptors, including TCR, CD3, and CD28. Within about 2–4 h, cell surface activation-markers, i.e., CD69, CD44,
and CD25 are upregulated, followed by increased secretion of cytokines, such as IL-2 and INF-c. T cells start proliferating
24–48 h after TCR stimulation. (b) Cultured density of T cells significantly affects T cell activation. T cells were activated
in the presence of IL-2 either with anti-CD3 alone or together with anti-CD28 and quantified after six days in culture to
determine the percentage of T cells exhibiting an activated “blast” morphology. Reproduced with permission from Ma
et al., J. Biomed. Biotechnol. 2010, 386545. Copyright 2010 Hindawi Publishing Corporation.
55
021501-8 de la Zerda et al. APL Bioeng. 2, 021501 (2018)

soon as a T cell identifies a cognate APC, it decelerates from >10 lm/min to <2lm/min and
fully arrests upon forming a stable IS.
52
Stimulation of na€
ıve T cells in the LN is organized
into distinct phases: (1) brief interactions with the APC (<7 min), followed by (2) a stable con-
tact phase with prolonged interaction and arrest (30 min), and (3) a final phase of serial, brief
interactions in which motility is restored and the T cells proceed to the inflammation site (Fig.
1).
53
When finally reaching the target tissue, the T cell motion is described as amoeboid, reach-
ing speeds up to 20 lm/min; this remarkably fast motility is thought to be enabled by weak
adhesion to the matrix.
54
b. Cytoskeletal changes. Many of the processes essential to initiate and sustain T cell activa-
tion response are cytoskeleton-dependent.
30(b)
These include integrin-mediated adhesion, IS for-
mation and maturation, cellular polarization, Ca
2þ
flux, receptor signaling, and downstream
changes in gene expression. The cytoskeleton regulates T cell activation by serving as a plat-
form for the recruitment of molecules that regulate adhesion and signal transduction, such as
Zap70 and LAT.
56
Cytoskeletal reorganization follows TCR stimulation and includes actin
polymerization and accumulation in the IS, reorientation of the microtubule-organizing center
(MTOC) towards the region of cell-cell interaction,
57
and the formation of an actin-rich struc-
ture known as the distal-pole complex on the opposing side of the cell
58
[Fig. 2(b)]. During cell
activation, the actin filaments can be polymerized or depolymerized in a dynamic manner as a
means of regulating the mechanical forces needed to sustain activation and motility.
59
Disruption of the cytoskeleton impacts T cell activation,
13,56,60
and inhibition of actin polymeri-
zation after the IS has formed diminishes the activation response.
61,62
While this is only a brief
summary of cytoskeletal rearrangements during T cell activation, several thorough reviews of
this highly complex interaction can be found elsewhere.
62–64
c. Ca
2þ
flux. Calcium signaling is an early event following IS formation. The Ca
2þ
concen-
tration within the cytoplasm dramatically increases within seconds following initial APC-T cell
contact [Fig. 3(a)].
65
In effect, the influx of Ca
2þ
serves as a “stop signal” and initiates the
steps necessary to reduce T cell motility. This Ca
2þ
increase occurs in parallel with TCR
microstructure formation and is a necessary step in rapid cytoskeletal reorganization
66
to
accommodate the formation of a stable IS.
d. Activation markers and gene expression. CD44 is a surface glycoprotein that is a known
receptor for hyaluronic acid. However, CD44 can also bind to other ligands, such as collagens
and matrix metalloproteinases (MMPs), and is involved in cell-cell interactions, cell adhesion,
and migration. CD69 is a surface glycoprotein that functions as an immunoregulatory molecule
during immune response.
67
Both of these glyocoproteins are upregulated during the brief serial
encounters with APCs in phase one
68
and are rapidly expressed after TCR stimulation, peaking
after approximately 24 h [Fig. 3(a)]. The CD69 activation marker has been correlated with the
strength of the activation response, and its upregulation is a highly sensitive measure of antigen
recognition. However, while CD44 is continually expressed, CD69 expression rapidly declines
following the end of TCR stimulation.
69
While these glycoproteins are useful markers to quan-
tify the T cell activation response, one needs to be careful as they can be upregulated even
when stimulation does not reach the minimum threshold for complete T cell activation.
NFAT and nuclear factor kappa-light-chain-enhancer of activated B cells (NF-kB) are tran-
scription factors that are expressed within half an hour following TCR stimulation. The expres-
sion of these transcription factors is independent of the formation of a stable T cell-APC conju-
gate and are expressed during brief serial APC encounters in phase one.
70
The induction of
NF-kB is crucial for substantial production of the cytokine interferon-c(IFN-c).
71
2. Late activation
a. Cytokine expression: Interferon-c(IFN-c) and interleukin-2 (IL-2). Cytokine secretion is nec-
essary for regulating the body’s inflammatory response. Two main cytokines are secreted when
021501-9 de la Zerda et al. APL Bioeng. 2, 021501 (2018)

Th1 cells and CD8
þ
T cells are activated: IFN-cand IL-2, both of which are expressed intracel-
lularly and then secreted out of the cell.
72
Intracellular expression is rapid, and can be detected
anywhere between 30 min to 2 h following APC-T cell conjugation. However, we will focus on
the secreted form, which can be detected only at later time points (>3–8 h) [Fig. 3(a)]. IFN-c,
named for its ability to interfere with viral infection, is a potent activator of macrophages and
plays a central role in inflammation and autoimmune disease and is produced during the brief
sequential encounters with APCs, about 3–6 h after initial contact. IL-2 is critical for the stimu-
lation of T cell proliferation, differentiation, and survival. Thus, it is also essential for ex vivo
T cell expansion.
73
In comparison to IFN-c, IL-2 production requires a prolonged stable con-
tact
74
and is secreted 2–4 h after a stable IS has formed.
75
b. Activation marker—CD25. CD25, also known as the IL-2 receptor, is an activation marker
expressed on the T cell surface. CD25 expression is already induced during the brief sequential
encounters that occur in phase one. Its expression can be detected 24 h after TCR stimulation
and peaks at 48 h, but it is rapidly lost after 72 h [Fig. 3(a)].
76
c. T cell proliferation. T cell proliferation starts 24–48 h after TCR stimulation, and in cul-
ture systems, the T cell proliferation is significantly affected by the density of T cells and
APCs [Fig. 3(b)]. Low cell density leads to infrequent cell-cell contact, which results in a defi-
ciency of CD3 crosslinking. As a result, na€
ıve T cells die quickly by apoptosis when seeded in
low cell densities (<110
4
cells/ml), but can survive for extended periods when cultured in
2D at high density (>110
6
cells/ml), which promotes higher levels of T cell activation and
proliferation.
55,77
In addition to T cell density, the duration of stimulation is a major factor in
determining the fate of na€
ıve T cells.
78
Na€
ıve T cells require approximately 20 h of sustained
contact to be fully activated and to start proliferating. This time duration includes the first
sequential encounter phase, which is inversely correlated with APC density and the number of
pMHC per APC,
77
and the stable contact phase. Chronic stimulation in vivo or overstimulation
in vitro, usually more than 3 days, carries the risk of inducing cell death.
78,79
IV. CURRENT TOOLS AND FUTURE OPPORTUNITIES TO EXPLORE T CELL
MECHANOBIOLOGY
Mechanosensing is the act of converting mechanical cues from the external environment
and converting them into biochemical signals that affect cell fate and function. As a T cell
migrates through its microenvironment, it undergoes both cell-cell and cell-matrix interactions
that provide these mechanical cues and influence cell fate. However, investigating how these
cues influence cell fate is challenging due to the difficulty of decoupling these signals or con-
trollably studying them together to elucidate meaningful, reproducible data. For example, a few
critical factors that may cross-interact to influence T cell fate include matrix stiffness, the tim-
ing of cell-APC contact, and the matrix structure, such as topography and porosity. These indi-
vidual cues often drive T cell fate and function but are hard to controllably tune within the lab-
oratory either individually or together to gain meaningful conclusions about the causal
relationships of the mechanotransduction process. To overcome these limitations, researchers
have tried to study T cell mechanobiology in artificial cellular environments where they can
control specific biophysical and biochemical parameters independently. In this section, we will
discuss some of the tools and technologies (Fig. 4) that have been developed to study this elu-
sive system and how they offer varying degrees of control over some relevant physiological
parameters such as stiffness, type of activator ligands, the binding moieties, and the seeded cell
densities.
A. Emerging methods to explore T cell mechanobiology using localized forces
Two major observations led to exploration of the TCR as a possible mechanosensor. First,
it was observed that soluble monovalent pMHCs bind to the TCR but fail to induce TCR sig-
naling,
80
whereas, soluble pMHC tetramers showed some degree of success.
81
Second, a T cell
021501-10 de la Zerda et al. APL Bioeng. 2, 021501 (2018)

FIG. 4. Potential bioengineering strategies to study T cell mechanotransduction. (a) Methods using localized forces to probe sin-
gle cells. A micropipette brings a T cell into contact with an antigen presenting cell (APC); the micropipette can pull the T cell
and induce normal or shear forces. The tip of an atomic force microscope (AFM) can be functionalized with either pMHC or
anti-CD3 to deliver a stimulatory signal to an adherent T cell while applying force-loads in specific patterns and time-durations.
Optical tweezers can apply a directional force on an adherent T cell using coated beads, commonly functionalized with antigenic
peptide loaded onto MHC (pMHC). The fluorescence biomembrane force probe (fBFP) strategy includes a functionalized bead
that is attached to the apex of a micropipette-aspirated red blood cell (RBC) and a T cell that is also micropipette-aspirated. The
latter brings the T cell and the bead into contact while precisely controlling the distance between them and the external forced
applied. (b) Magnetic force-based platform. Magnetic fields are applied to cells dosed with magnetic nanoparticles. This tech-
nique enables on-demand exertion of localized force over a population of cells. Reproduced with permission from Tseng et al.,
Nat. Methods 9(11), 1113 (2012). Copyright 2012 Nature Publishing Group.
92
(c) Methods using synthetic polymer substrates
with variable stiffness. T cells are stimulated through antibodies (i.e., anti-CD3, anti-CD28) that are conjugated to the substrate
surface (left side) or through APCs that are seeded together with the T cells on the substrate surface (right side). (d) Micropost
array platform. Polymeric microposts bend in response to cell-applied traction forces (F), which can be estimated based on the
micropillar height (L) and radius (r), pillar mechanical properties (Young’s modulus E), and the amount of bending or deflection
of the pillar (DX). Image provided courtesy of the Mechanobiology Institute, National University of Singapore.
021501-11 de la Zerda et al. APL Bioeng. 2, 021501 (2018)

and an APC perform a “dance” when they interact, during which the T cell pushes and pulls
the APC’s plasma membrane, thus potentially exposing the TCR to significant forces.
82
Upon
observing this forceful interaction, several researchers began speculating that mechanical cues
may trigger TCR signaling by driving conformational changes in the TCR/CD3 complex.
17,83
This led to a series of studies that utilized methods to exert localized forces on T cells to probe
the nature of this interaction.
Kim et al.
84
used optical tweezers (method reviewed elsewhere
85
) to manipulate pMHC-
coated beads to apply a directional force of 50 pN on the T cell membrane [Fig. 4(a)].
Tangential (shear) and perpendicular (normal) forces were applied, but only the tangential force
initiated Ca
2þ
flux, suggesting that the TCR/CD3 complex is an anisotropic force sensor. Later
papers would challenge this result,
17
but a proposed possible explanation is that tangential
movement is better at uncovering shielded TCRs, which are covered by a thick layer of large
glycoproteins, than vertical movement.
86
Uncovering the shielded TCRs would increase the
possibility of a pMHC-coated bead binding to the TCR, and therefore yield the observed Ca
2þ
flux. A follow up study by the same lab used optical tweezers to investigate how applied force-
loads and force-directionality affect the chemical threshold of TCR stimulation.
87
The pMHC
surface concentration of the coated-bead was carefully manipulated and controlled. The results
indicated that in the absence of applied force, the pMHC surface concentration required for
activation is orders of magnitude higher than the physiologically relevant concentration. In con-
trast, an applied force of 10–20 pN in the shear direction can stimulate TCR activation with as
little as two pMHC present.
In complementary work, Li et al. developed an experimental setup that attempted to
increase the probability of TCR-ligand binding and thereby eliminate the confounding effects
of mechanical force on TCR unshielding.
17
In their work, the authors used a micropipette (tech-
nique reviewed elsewhere
88
) to pull and induce shear forces on T cells that were brought in
contact with fibroblast-based aAPCs, which expressed a membrane-tethered, anti-CD28, co-
stimulatory ligand and defined anti-CD3-ligands [Fig. 4(a)]. There were two types of TCR-
ligands, a short one and an elongated one. In principle, the probability of encountering the TCR
is higher for the elongated TCR-ligand because of its extended reach. However, results indi-
cated that in the absence of external forces, the elongated TCR-ligand resulted in poor activa-
tion, whereas the short TCR-ligand led to robust Ca
2þ
flux and T cell proliferation. In contrast,
in the presence of external forces, the elongated TCR-ligand successfully resulted in increased
Ca
2þ
flux for both tension forces and shear forces. Furthermore, these experiments provided an
indication that the process may be Src kinase-dependent. In summary, unlike the earlier report,
this manuscript concluded that the TCR may not be anisotropic.
Liu et al. followed up on this work by developing a newly modified fluorescence biomem-
brane force probe (fBFP) technology to quantitatively investigate the strength of the TCR-
pMHC bond in the presence of external forces.
89
Specifically, they focused on the in situ kinet-
ics of the TCR-pMHC interaction to determine whether force plays a role in antigen recognition
and discrimination, which is critical for adaptive immunity. To achieve this, they expanded on
BFP, a previously developed technology that measures receptor-ligand binding kinetics for a
single-molecule,
90
and modified it to also simultaneously monitor the Ca
2þ
response resulting
from this binding interaction. The experimental setup [Fig. 4(a)] includes a pMHC-biotinylated
glass bead that is manually attached to the apex of a micropipette-aspirated red blood cell
(RBC). Another micropipette aspirates a primary, na€
ıve T cell to control contact between the T
cell and the functionalized bead. While this micropipette-based approach enables nanometer
precision in cell-ligand positioning, it requires a laborious and sensitive process of manual pro-
duction and assembly.
91
The RBC is used as a sensitive force transducer, whose spring constant is calculated based
on the aspiration pressure and a series of radii including the probe micropipette, (R
p
), RBC
(R
0
), and circular contact area between the RBC and probe bead (R
c
). While the T cell and
bead are brought into contact, Ca
2þ
imaging is used to detect early T cell activation, thus
enabling them to establish the connection between the external force, bond characteristics, and
early activation in T cells. The authors discovered that the pMHC–TCR bond is a “catch bond,”
021501-12 de la Zerda et al. APL Bioeng. 2, 021501 (2018)

(i.e., increased force prolongs the bond’s lifetime). In contrast, the antagonist ligand-TCR is a
“slip bond,” (i.e., increased force shortens the bond’s lifetime) which may help elucidate the
mechanism of self-versus non-self-discrimination. In addition, the timing and magnitude of
applied force was also found to be critical for activation, and the latter observation has been
confirmed by other studies.
93
The fBFP technique was also used by Husson et al., who investi-
gated how force-exertion of the T cell changes in response to different stiffness of the aAPC.
47
They used an antibody-coated bead aAPC and indirectly modeled different stiffness by chang-
ing the aspiration pressure of the RBC, which effectively varies the spring constant of the force
probe between 50 and 1000 pN/mm. The study demonstrated that the force exerted by the T
cell during the pulling phase increases with higher force-probe stiffness, which may imply that
a T cell can adjust its force response based on APC mechanics. Later, Pryshchep et al. used the
fBFP technique to apply cyclic force-load patterns with varied intermission time and constant
contact time on the TCR of a na€
ıve T cell using a pMHC biotinylated RBC.
94
The study found
that a 5-s intermission time was sufficient to induce Ca
2þ
flux, while a 10-s intermission time
impaired this ability. For all of these various fBFP studies, the experiments were performed
inside a glass chamber filled with the medium. In the future, it may be interesting to perform
fBFP experiments in the presence of other mechanical environments such as hydrogels or other
mechanically modular environments to better recapitulate the natural system.
Another approach to quantify the forces involved in early T cell activation is the use of a
functionalized AFM tip [Fig. 4(a)]. Hu et al. used this method to deliver an antigenic signal
(pMHC) or anti-CD3 signal to primary and effector T cells to evaluate the role of force in
mediating actin cytoskeleton rearrangement.
59
T cell activation is known to be prevented in T
cells with altered actin cytoskeleton dynamics.
63
AFM was used to measure the forces gener-
ated by the T cell as well as to apply force-loads in specific patterns including cyclic, continu-
ous, and spaced, while Ca
2þ
flux was monitored. As expected, treatment of the T cells with
latrunculin A (LatA), which prevents polymerization of actin filaments, inhibited the ability of
T cells to exert pushing and pulling forces and prevented Ca
2þ
flux upon TCR-engagement.
However, upon application of a cyclic force-load, LatA-treated T cells displayed partial recov-
ery of TCR-induced Ca
2þ
flux. Application of a cyclic force-load alone, without TCR engage-
ment, was not successful in inducing Ca
2þ
flux, supporting the idea that mechanical forces
were transmitted to the T cell through the TCR. Interestingly, continuous and spaced force-
loads did not yield Ca
2þ
-flux recovery in LatA-treated cells upon TCR-engagement. These data
suggest that the dynamics of actin cytoskeleton rearrangement plays a key role in T cell-force
generation and subsequent Ca
2þ
flux induction. The fact that only a cyclic force-load success-
fully initiated Ca
2þ
flux may imply that the interval time between sequential encounters needs
to last less than 10 s to allow for signal integration. This AFM experimental design uses com-
mercially available equipment and thus enables a fairly high throughput analysis of multiple,
single-cell events for robust statistical analysis compared to fBFP. However, this evaluation of
many different single-cell events also raises a new technical concern, that one must standardize
the level of T cell activation for each individual, single-cell experiment. This includes quantifi-
cation and standardization both of the level of pre-activation and stimulation. Here, a popula-
tion of T cells were pre-stimulated in bulk for 2 days and then allowed to rest for 4 days prior
to AFM evaluation. Thus, it is possible that not all cells had similar levels of pre-activation.
Furthermore, each single-cell experiment used a freshly prepared AFM tip without quantifica-
tion of ligand conjugation. As the ligand concentration can influence the magnitude of T cell
activation, it is also possible that different cells were presented with different levels of activa-
tion. Nevertheless, this study demonstrates that controlled AFM-delivered forces can transform
an incapacitated T cell to be operable by replacing the role of the actin cytoskeleton in force-
actuation.
This section discussed methods that can be used to mechanically probe a single cell,
including optical tweezers, micropipette aspiration, fBFP, and AFM. These methods can assist
to examine short and early signaling events after TCR stimulation but are limited in the explo-
ration of late activation signaling, such as cytokine secretion and proliferation, which occurs
during long incubation periods and are influenced by a threshold density of T cells.
021501-13 de la Zerda et al. APL Bioeng. 2, 021501 (2018)

Furthermore, it may be challenging to keep a single T cell viable for long time periods while
being probed with these techniques. That being said, there might be a way to circumvent this
challenge by adding different combinations of cytokines to the cell medium that will facilitate
delivery of survival cues to the T cell.
One potential future opportunity to overcome some of the limitations of the mechanical
probing strategies used thus far in T cell research is techniques based on magnetic tweezers.
Recently, Tseng et al. developed a novel strategy to control localized, on-demand, time-varying
exertion of force over a population of cells. In this method, they used individually patterned
magnetic nanoparticle-dosed cells and applied magnetic fields to achieve localized and spatially
resolved forces, which can be maintained for extended periods of time on the cell membrane
[Fig. 4(b)].
92
Finally, most of these experimental setups have used an aAPC in the form of a functional-
ized glass bead or a silicon AFM probe, both of which have stiffness in the GPa range, which
is non-physiological. As an alternative approach, researchers have begun to use 2D polymeric
substrates, which are the focus of Sec. IV B, to evaluate T cell responses to physiological stiff-
ness changes over long activation time periods.
B. Emerging methods to explore T cell mechanobiology using 2D substrates
Judokusumo et al. formed 2D, polyacrylamide (PAA) gel substrates with different material
stiffness (2–200 kPa) by modulating the crosslinker density.
19
The substrates were coated with
anti-CD3 and anti-CD28 antibodies using streptavidin-biotin conjugation, and na€
ıve murine
CD4
þ
T cell activation was monitored by quantifying IL-2 secretion [Fig. 4(c)]. Interestingly,
there was no difference in IL-2 secretion for substrates with moduli lower than 10 kPa, possibly
implying that a minimal substrate rigidity is required for effective mechanotransduction. To
study whether this response is actin cytoskeleton-mediated, they treated the cells with a myosin
inhibitor, blebbistatin. Blebbistatin-treated samples showed reduced IL-2 secretion on samples
with Young’s moduli higher than 10 kPa. They further suggested that this mechanotransduction
effect may be mediated by phosphorylation of the Zap70 protein and the Src family kinase pro-
teins (SFK), which are known to be downstream from TCR stimulation.
Motivated by immunotherapy applications, O’Connor et al. investigated the effect of sub-
strate rigidity on the expansion of human CD4
þ
and CD8
þ
T cells using poly(dimethylsiloxane)
(PDMS) [Fig. 4(c)].
18
They produced substrates with Young’s moduli ranging from 100 kPa
to above 2 MPa by manipulating the crosslink density, and they stimulated the cells through the
adsorption of anti-CD3 and anti-CD28 ligands. Their work showed that IL-2 and IFN-cproduc-
tion were significantly higher on more compliant gels (100 kPa). Additionally, they evaluated
proliferation over a 15-day incubation period and found that the log phase period of cell dou-
bling extended to 15 days on the more compliant gels, whereas on the stiffer gels this phase
continued only for 10 days. These data are difficult to compare with the findings of
Judokusumo et al., since the modulus values for the intermediate stiffness PDMS substrates
were not characterized and the range of stiffness only partially overlap between the two studies.
In addition, the two reports use different polymer chemistries, which may result in different
nano- and/or micro-structure of the polymer substrates and may have different levels of back-
ground cell toxicity.
95,96
In addition, O’Connor et al. coated their substrates through passive
adsorption, which may not be stable over long incubation times. While their data demonstrated
stable binding of the stimulatory ligands for the first 2 days, further validation is required to
support longer experimental times. Nonetheless, this paper highlights the importance of
mechanical properties of the microenvironment as a regulator of T cell function.
The stiffness range explored in the previous papers was markedly higher than the lymphoid
native tissue and APC stiffness, which typically are lower than 1.5 kPa.
11
Hui et al. explored
the stiffness range of 1–5 kPa using PDMS substrates treated with poly-d-lysine (PDL) to pro-
mote cell adhesion and covalent attachment of anti-CD3 ligands.
20
They quantified the activa-
tion response of an immortalized cell line, Jurkat T cells, through immunoblotting, specifically
looking at tyrosine phosphorylation. The results suggested that increased substrate stiffness
021501-14 de la Zerda et al. APL Bioeng. 2, 021501 (2018)

leads to an increase in transient tyrosine phosphorylation, which is known to contribute to early
activation events. A limitation of this study is that it lacked presentation of a costimulatory
ligand, such as anti-CD28. Regardless, this paper indicates that contrary to what had been pre-
viously reported, the mechanics of compliant substrates may also affect T cell
mechanotransduction.
Most recently, Saitakis et al. investigated the physiological range of stiffness (0.5–100 kPa)
using PAA gel substrates.
97
The substrates were coated with anti-CD3, anti-CD28, and intercel-
lular adhesion molecule-1/FC-chimeric molecules (ICAM-1) using the streptavidin-biotin conju-
gation method employed previously.
19
The study used pre-activated human CD4
þ
T cells (day
6) that were seeded on gel substrates with variable stiffness. The results indicated that T cells
were more activated on stiffer substrates than on softer substrates and that some effector func-
tions were sensitive to the whole range of stiffness examined (e.g., migration, gene expression,
cytokine expression, and secretion), whereas other functions (e.g., cell cycle progression and
metabolism) were only affected at later (72 h), but not early (24 h), time points. Interestingly,
stiffness-induced changes in gene expression were only observed when anti-CD3 stimulatory
ligands were present, which suggests that the TCR/CD3 complex may be the main mechano-
sensing module. In addition, cytokine gene expression and secretion were enhanced with
increasing substrate stiffness with and without ICAM-1 present. This study was the first to eval-
uate the role of matrix stiffness within the context of an integrin ligand, ICAM-1. This is
important as other studies have reported that integrins mediate the mechanotransductive process
in other cell types.
98
While most systems have investigated the effects of substrate stiffness by using ligand-
functionalized synthetic materials, Basu et al. employed natural materials to study these interac-
tions under more physiologically relevant conditions. To achieve this, they made two modifica-
tions to the PAA system from before: an adsorbed fibronectin coating was put on the PAA sur-
face (stiffness range of 12 kPa and 50 kPa) and antigenic peptide ovalbumin (OVA)-loaded
APCs were used to physiologically stimulate CD8
þ
T cells.
9
The use of true APCs on top of
substrates with different stiffness enabled them to decouple the matrix mechanical cues from
the cell-cell interactions involved in IS formation [Fig. 4(c)]. Results indicate that CD8
þ
-medi-
ated killing, specifically the ability of T cells to bore through the membrane of the APCs, is
enhanced on stiffer substrates. Therefore, this paper was the first to link T cell mechanotrans-
duction to a physiologically functional outcome of T cell activation. One limitation of the
experimental protocol is that the fibronectin was coated onto the PAA using physical adsorption
rather than chemical conjugation. The cells pushing and pulling against the surface may cause
the fibronectin to detach from the substrate, thus remodeling the fibronectin protein layer and
causing the microenvironment’s biochemical cues to vary across substrates. In addition, the uni-
formity and concentration of the fibronectin coating was not quantified, and this may be altered
by the PAA crosslink density.
96
T cells express receptors for fibronectin (e.g., VLA-4 and
VLA-5) that are known to assist in TCR activation, but their signaling pathways and functional-
ity is not fully understood.
99
It is an open question whether these receptors contribute to mecha-
notransductive effects that mediate CD8
þ
T cell killing.
In this same manuscript, the authors also used a micropillar method to spatially resolve the
forces exerted by cells on the surfaces to which they are adhered. Micropillars of defined geom-
etry and fabricated from elastic materials (e.g., PDMS) have a known spring constant.
100
Observing the micropillars’ deflection during interaction with cells provides a means of measur-
ing the cell traction forces. Basu et al. used micropillars to verify that T cells exert mechanical
forces onto doomed target cells, resulting in more efficient killing.
9
In a spatiotemporally coor-
dinated attack, CD8
þ
T cells exert force onto a targeted cell’s membrane along with a signifi-
cant release of perforin. This combination of mechanical force with perforin release aids in
pore formation in the targeted cell by essentially stretching out the cell membrane and thus
priming it for pore formation. Activated CD8
þ
T cells exerted more force than na€
ıve T cells
and in an asymmetric pattern, suggesting a spatially targeted interaction with a target cell.
These newly formed pores allow a large influx of cytotoxic proteins into the cell, thus resulting
in more rapid cell death. Locations of the highest force exertion were coupled with a local
021501-15 de la Zerda et al. APL Bioeng. 2, 021501 (2018)

increase in cytolytic protein secretion. This study showed that T cells leverage mechanical
forces to aid in cytotoxic activity against target cells.
Bashour et al. used micropillars functionalized with anti-CD3 and anti-CD28 to demon-
strate that T cells exhibit non-integrin-based mechanosensing through these activating and co-
stimulating signals.
101
The interactions between a T cell and the pillars were classified into four
distinct phases: (1) contact, (2) spreading, (3) transient, and (4) contractile phases. The greatest
traction forces were observed during the transient and contractile phases, with forces reaching
approximately 100 pN on pillars presenting both anti-CD3 and anti-CD28. Upon stimulation
with pillars displaying only anti-CD3, T cells generated about half the traction force as
observed in the anti-CD3/anti-CD28 combined condition. Cells seeded on anti-CD3-only pillars
with anti-CD28 present in the culture medium exerted approximately the same traction forces
as the anti-CD3/anti-CD28 combined condition, suggesting that TCR-CD3 interaction is where
force generation occurs, while CD28 increases the force through biochemical signaling. T cells
did not significantly adhere to anti-CD28-only decorated pillars. This study concluded that T
cells exert traction forces through TCR-CD3 interactions with associated signaling in the Src
kinase family.
Both Basu et al. and Bashour et al. used micropillars with a single defined rigidity and
geometry. In the future, researchers can employ the micropost array platform developed by Fu
et al., in which PDMS pillar rigidity is altered through changing their composition or height
[Fig. 4(d)].
100
This controlled system can be adopted to investigate T cell force generation in
response to substrates of different rigidity. That said, studying T cell traction forces using the
micropillar approach may introduce other complexities, since T cell force-generation may be
limited by the pillar spacing. Previous work (discussed in more detail below) has shown that T
cells are exquisitely sensitive to patterns of stimulating ligands; thus, it is unclear if forces
induced by micropillars (which inherently present stimulating ligands as an arrayed pattern)
would be similar to forces induced by homogeneous, 2D substrates.
102
One possible approach
to overcome this challenge in the future is traction force microscopy (method reviewed else-
where
103,104
). This approach measures the forces generated by adherent cells on flat, elastomeric
substrates. By imaging the displacement of fluorescent beads embedded near the top of the sub-
strate, one can estimate the traction forces exerted on the substrate. A potential additional
advantage of traction force microscopy is that it can be used to resolve forces applied both in
the horizontal and vertical directions.
104,105
When preparing functionalized substrates for mechanotransduction studies, one important
variable that has yet to be quantitatively evaluated is the spatial organization of the activating
ligands. In complementary work, spatial organization of signaling factors has been shown to
have important consequences for T cell activation.
106
For example, Shen et al. explored how
spatial distribution of anti-CD28 and anti-CD3 cues influenced T cell activation on rigid sub-
strates using microcontact printing.
107
Cell behavior on microscale, circular patterns containing
anti-CD3 co-localized with anti-CD28 was compared to that on circular patterns of segregated
“islands” of anti-CD3 interspersed with “islands” of anti-CD28. Murine na€
ıve T cells (CD4
þ
)
were observed to have greater IL-2 secretion on segregated patterns compared to co-localized
patterns, which was correlated with NF-kB translocation and PKB/Akt signaling.
In another micropatterning study, Jung et al. demonstrated that the asymmetric division of
activated T cells could be controlled by the microscale presentation of anti-CD3 and anti-CD28
with an ICAM-1 background on PDMS substrates.
108
Asymmetric division of activated T cells
is believed to be a mechanism by which a single T cell can give rise to a diversity of effector
or memory T cells. In this work, the size (4 and 10 lm) and distance (15 and 20 lm) between
activation sites of anti-CD3 and anti-CD28 were varied. A larger distance between activation
sites led to higher instances of asymmetric division of CD4þmurine T cells upon activation.
Furthermore, they used time-lapse microscopy to image the adhesion or migration of the two
resulting daughter cells. They observed that when one daughter cell remained anchored to the
pattern, the TCR remained polarized during cytokinesis and asymmetric division occurred. This
work highlights the influence that microscale presentation of antigens to T cells has on the sub-
sequent T cell expansion behavior.
021501-16 de la Zerda et al. APL Bioeng. 2, 021501 (2018)

In addition to microscale patterning, nanoscale patterning of stimulatory molecules has also
been explored. The Spatz group utilized block-copolymer micelle nanolithography to create
nanopatterns of immobilized anti-CD3.
109,110
Briefly, in Matic et al., gold nanoparticles
(AuNPs) were ordered on glass substrates in a nanoarray, passivated with a polyethylene glycol
(PEG) layer, and then conjugated with anti-CD3.
102(b)
The steric constraints of the PEG passiv-
ation resulted in a near 1:1 ratio of antibodies per AuNP, thus providing exquisite control over
nanoscale patterning. They found that anti-CD3 alone could induce CD69 expression on T cells
when the interparticle distance was less than 100 nm. In general, higher activation was achieved
with higher density arrays, plateauing at 316 AuNPs/lm
2
. Although the anti-CD3 nanoarray did
not induce measurable IL-2 production, addition of soluble anti-CD28 as a costimulator did effi-
ciently lead to increased IL-2 production.
In another report, Spatz and coworkers further explored the response of T cells to the density
and specific nanopatterning of pMHC.
109
Deeg et al. determined that about 100 pMHC molecules
per lm
2
was the threshold for eliciting cell adhesion and IL-2 production, which corresponds to a
ligand spacing of about 115 nm. For pMHC spacing above 150 nm, a significant number of cells did
not adhere. They then evaluated surfaces that included both nano-patterning and micro-patterning.
When comparing substrates with extended, continuous nanopatterning to substrates with micro-
nanopatterning (i.e., micrometer-scale regions with dense nanometer-scale patterning), the micro-
nanopatterned substrates showed less adhesion and activation than the extended nanopatterns even
while the nanoscale particle density was held constant. This finding suggests that the global ligand
density is of primary importance in T cell adhesion and activation over the nanoscale ligand presenta-
tion. In the future, an interesting follow-up to these studies would be to micro- and nano-pattern flexi-
ble substrates to evaluate how rigidity and stimulatory ligand patterning co-influence T cell response.
To summarize, regardless of the substrate chemistry, the stimulator identity, matrix coating
method (chemical coupling or physical adsorption), or even the type of T cell, all of the studies
discussed in this section indicate that T cells are mechanosensitive (Table I). While a broad
range of mechanical stiffness has been probed, most of the work to date has focused on use of
culture systems that are much stiffer than that of LNs where na€
ıve T cells get activated in vivo.
The fact that trends in the results are not necessarily consistent highlights the need for experi-
mental designs with fully controlled and quantified parameters. This will become increasingly
important as this nascent field begins to turn its attention to elucidating the specific mechanisms
of how mechanical cues, like substrate stiffness, are translated into chemical signals inducing
activation. To address this limitation, it may be necessary to incorporate new types of biomate-
rials in which the stiffness, ligand density, and viscoelasticity can be independently tuned.
C. Emerging methods to explore T cell mechanobiology using 2.5D substrates
For the purpose of this review, 2.5D substrates are defined as those that have topographical
features that can modify the cell membrane curvature but still have one free cell surface, result-
ing in different apical and basal cell properties. Thus, these systems are distinct from 2D sub-
strates where the basal surface of the cell membrane is typically assumed to be flat.
Furthermore, 2.5D systems are distinct from 3D systems where cells are completely encapsu-
lated within a surrounding material [Fig. 5(a)]. By this definition, micropillar systems could be
categorized as either 2D or 2.5D, depending on the size and spacing of the pillars and the
resulting basal cell membrane geometry [Fig. 5(b)]. Additional materials commonly used that
could be considered 2.5D may include topographically varied surfaces, some fibrous or curved
substrates, and certain synthetic supported lipid bilayers (SLBs) that induce the cell membrane
curvature. T cell studies using these materials have provided insights into the behavior of T
cells, including understanding the ramifications of antigen display patterns and measuring the
traction forces exerted by T cells. In this section, we will describe these various materials with
a focus on the important findings these studies have contributed to T cell biology.
Kwon et al. developed nanotopographies of zigzag patterns with varied turning angles and
side lengths on polyurethane acrylate surfaces to study the impact of topography on T cell
migration.
110
The patterns were fabricated using UV-assisted capillary force lithography with
021501-17 de la Zerda et al. APL Bioeng. 2, 021501 (2018)

TABLE I. Comparison of experimental methods and results of T cell activation in 2D cultures with varying stiffness. Experimental methods and results are not necessarily consistent between studies.
There is variability between (from left to right): T cell identity; density of T cells, which is known to impact T cell fate;
55
substrate chemistry; method of stimulatory ligand presentation; substrate stiff-
ness range; T cell function(s) measured; and reported response outcomes.
Cell type
Seeded cell
density (cells/cm
2
)
Substrate
chemistry Activators Stiffness range (E) T cell function measured Activation outcomes References
Mouse naive CD4
þ
3.1 10
5
PAA-gels
containing
streptavidin
Biotinylated anti-CD3,
anti-CD28
10–200 kPa IL-2 production,
phosphorylation of
SFK and Zap 70
*No change in cell response
below 10 kPa *Activation
increased with stiffness
Judokusumo et al.
Human naive CD4
þ
and CD8
þ
0.8–1 10
6
cell/ml *Culture
plate is not specified
PDMS Physical adsorption of
anti-CD3, anti-CD28
100 kPa–2 MPa IL-2 and IFN-cproduction,
cell proliferation
*Activation decreased with
stiffness
O’Connor et al.
Jurkat (immortalized
cell line)
Not reported PAA treated with
hydrazine hydrate
Only anti-CD3 1, 5 kPa Phosphorylation of Zap70,
Lat, and SLP76
* Early signaling enhanced
with stiffness *Late signaling
decreased with stiffness
Hui et al.
Mouse CD8
þ
3.1 10
4
PAA coated
with adsorbed
fibronectin
pMHCþAPC (B16
melanoma cells)
þpeptide antigen
12, 50 kPa Target cell killing CD8
þ
-mediated killing
increased with stiffness
Basu et al.
Human primary CD4
þ
5.7 10
5
PAA-gels containing
streptavidin
Biotinylated ICAM-1,
biotinylated anti-CD3
and anti-CD28
0.5, 6.4, 100 kPa Migration, morphology,
metabolism, cell-cycle-related
genes, IFN-cand
TFN-aproduction,
CD25 and CD69 surface
marker expression
*Generally activation
increased with stiffness.
*Some functions had a
low stiffness
threshold for activation
(e.g., cytokine signaling)
*Some functions were
sensitive to higher stiffness
regimes (e.g., metabolism
and cell-cycle genes)
Saitakis et al.
021501-18 de la Zerda et al. APL Bioeng. 2, 021501 (2018)

feature lengths from 15 to 60 lm and turn angles of 45,90
, and 135. The surfaces of the pat-
terned substrates were coated with ICAM-1 to render the materials adhesive for T cells. They
observed that sharp turn angles of 45significantly decreased mean cell velocity, but the feature
side lengths had little impact on mobility. Following these findings, the researchers investigated
the importance of lamellipodia on motility using the nanopatterned surfaces and an actin-
branching inhibitor. Using the Arp2/3 inhibitor CK-636, T cell formation of lamellipodia on the
leading cell edge was greatly inhibited. This loss of lamellipodia correlated with a significant
decrease in the cell mean velocity, as well as an increase in T cells becoming “trapped” at the
turning points. These results demonstrate that lamellipodia are integral for rapid movement of T
cells on complicated topographical environments.
FIG. 5. Classification and ambiguity of 2D, so-called 2.5D, and 3D culture systems. (a) Common examples of dimensional-
ity in cell culture systems. Illustration of common 2D, 2.5D, and 3D culture systems. 2D culture systems are those that are
perceived by the cell to be flat substrates; so-called 2.5D culture systems have topographical features that induce changes
in the membrane curvature of the cell; 3D culture systems include those that completely and continuously encapsulate the
cell to provide matrix contact on all sides. (b) Examples of cell culture systems with ambiguous dimensionality. Top panel,
micropillar systems could be classified as either 2D or 2.5D, depending on the size and spacing of the pillars with respect
to the basal cell membrane geometry. Middle panel, encapsulation systems can be classified as either 2.5D or 3D depending
on the pore size diameter (d) of the material. Bottom panel, cell encapsulation within fibrous materials can be classified as
either 2.5D or 3D depending on the spacing between fibers (F
S
).
021501-19 de la Zerda et al. APL Bioeng. 2, 021501 (2018)

Hu et al. recently developed a 96-well plate platform termed the Integrated
Mechanobiology Platform (IMP) for the high-throughput screening of cellular response to topo-
graphical features.
21
The plate bottoms for this platform are formed by casting PDMS onto
molds previously patterned with various topographies via electron-beam lithography. These bot-
toms are then bonded onto bottomless 96-well plates; thus, each topography is separated into
individual wells. This platform enables use of a variety of cell analysis techniques including
enzyme-linked immunosorbent assay (ELISA) and flow cytometery, which are not compatible
with standard microarray chips. Furthermore, the PDMS-coated well bottoms are also compati-
ble with several microscopy techniques including confocal microscopy. Proof-of-concept studies
explored T cell responses to various topographies coated with anti-CD3 and anti-CD28. On
grid-patterned topographies, T cell activation (as measured through IL-2 production), spreading,
and proliferation were significantly decreased on patterns with 100 nm wide trenches compared
to cells on patterns with 200 or 300 nm wide trenches. To demonstrate potential applications
for drug screening, they demonstrated that the myosin inhibitor blebbistatin in combination
with a 200 nm trench led to an enhancement of IL-2 production. This platform may be useful
for quickly identifying topographical patterns that influence T cell biology.
Nano and micro technologies enable the control of a myriad of physical features, including
topographical geometry, surface roughness, and curvature, which may be key in designing con-
trolled mechanical microenvironments. One area that has not been as widely studied, but is
recently becoming of interest, is the link between the membrane curvature and mechanotrans-
duction through the Bin/Amphiphysin/Rvs (BAR) domain proteins, which are found in many
types of cells including T cells.
111
Curvature changes of the lipid membrane are involved in a
number of cellular processes, including facilitating filopodia formation and cell division.
Similarly, the BAR proteins are also known to play a role in a wide range of cellular processes,
such as cortical actin structure regulation and endocytosis. Galic et al., reported that induction
of curvature in the plasma membrane of migrating 3T3 cells by nano-cones triggers the recruit-
ment of isolated BAR domains to the sites of the local membrane curvature (Fig. 6).
112
An
alternative technique to alter the local membrane curvature is the use of supported lipid bilayers
(SLBs) on coated-curvature-controlled platforms, where different curvature radii are fabricated
using colloidal assembly.
113
SLBs are made of a phospholipid bilayer created through the spon-
taneous adsorption or fusion of phospholipid vesicles onto the supporting surface. Coating with
SLBs is beneficial as it results in a biocompatible surface that enables ligand fluidity as well as
controlled composition of stimulatory ligand density and rigidity. SLBs were used as early as
1984 as aAPCs to elicit an immune response from T cells
114
and have shown great utility in
probing T cell activation. Additionally, artificial diffusion barriers can be installed, thus provid-
ing “corrals” of bilayers with potentially unique characteristics compared to the rest of the sur-
face. Mossman et al. used SLBs with nanometer-scale geometric constraints to frustrate normal
IS formation.
115
In this work, they found that the signaling activity depends on the radial posi-
tion of T cell receptors. Combining these types of T cell activation studies with measurements
FIG. 6. Nanocone platform. Left, scanning electron microscope (SEM) side image of the nanocones. Right, transmission
electron microscopy (TEM) image of NIH 3T3 fibroblast cells (colored red), in which the nanocones are inducing a curva-
ture in the plasma membrane of the cells. Reproduced with permission from Galic et al., Nat. Cell Biol. 14(8), 874 (2012).
Copyright 2012 Nature Publishing Group.
112
021501-20 de la Zerda et al. APL Bioeng. 2, 021501 (2018)

of membrane curvature and local signaling of mechanotransduction proteins may yield new
insights into how mechanical forces potentiate T cell activation.
D. Emerging methods to explore T cell mechanobiology in 3D scaffolds
While there is a wealth of knowledge about maintaining cell cultures in 2D, these models
are inherently limited when trying to mimic the native 3D architecture. Many cells behave dif-
ferently in 2D culture systems compared to 3D systems.
116
Thus, studying T cell mechanobiol-
ogy in a 3D context may provide insights that otherwise could not, or would not as accurately,
be observed. While no studies of T cell mechanotransduction within 3D microenvironments
have yet been reported, there have been several elegant strategies developed to incorporate T
cells into 3D systems. Since most of these systems use polymeric materials with tunable
mechanical properties, these platforms could be feasibly extended into studies of 3D T cell
mechanotransduction. Several approaches can be used to study mechanical forces in 3D micro-
environments, including particle tracking microscopy in 3D, diffusing wave spectroscopy, and
dynamic light scattering (methods reviewed elsewhere
117
).
To date, the study of T cells within 3D engineered platforms has largely focused on the
design of synthetic lymphoid organoids (SLOs). The broader goal of this field is to leverage
our native immune system as a therapeutic by artificially providing cues that drive the accumu-
lation of adaptive immune cells, e.g., T cells, towards a tumor or site of infection. Since T cells
are maintained, activated, and differentiated in the secondary LNs, SLOs are being developed
as in vitro and in vivo systems to better understand what critical features can be leveraged to
enhance immunotherapies.
118
Scaffold materials explored for these applications include both
synthetic and isolated native materials including alginate, agarose, polyamide, polyethylene gly-
col (PEG), and collagen. Typically, these include embedded stromal cells, which are known to
secrete various ECM proteins into the surrounding microenvironment. As a result, these culture
systems lack precise control over the mechanical cues sensed by the T cells. Nevertheless, these
studies demonstrate the feasibility of embedding T cells in an in vitro 3D culture system, and
several of these platforms may be useful for further investigation of how mechanics affect T
cell activation.
FIG. 7. T cells encapsulated within various matrices. (a) T cell in a confined/continuous 3D collagen matrix gel. A scanning
electron micrograph of a T cell encapsulated within a collagen gel matrix. Reproduced with permission from Schor et al.,J.
Cell Biol. 96(4), 1089 (1983). Copyright 1983 Rockefeller University Press.
119
(b) T cells encapsulated in an inverted colloi-
dal crystal (ICC) scaffold. Confocal fluorescence microscopy images of thin sections cut through a 1 mm-thick ICC scaffold.
T cells are labeled with carboxymethylfluorescein diacetate (CMFDA) (green); PEG ICC (red) has 80-lm void diameters.
Scale bar: 41 lm. Reproduced with permission from Stachowiak and Irvine, J. Biomed. Mater. Res., Part A 85(3), 815 (2008).
Copyright 2007 Wiley Periodicals, Inc.
120
(c) Matrix viscoelasticity. The left panel demonstrates that many native tissues are
viscoelastic, and thus exhibit time-dependent stress relaxation. Reproduced with permission from Chaudhuri et al.,Nat.
Mater. 15(3), 326 (2016). Copyright 2015 Nature Publishing Group.
121
The right panel schematic shows how stress relaxation
can occur at a molecular level due to rearrangements of the polymer network crosslinks over time.
021501-21 de la Zerda et al. APL Bioeng. 2, 021501 (2018)

Schor et al., was the first to encapsulate na€
ıve T cells in a 3D confined/continuous hydrogel
[Fig. 7(a)].
119
In their studies, collagen matrices were used to study the kinetics of lymphocyte
penetration and to determine the T cell distribution throughout the collagen matrix. This matrix
was specifically selected as it is a constituent of the tissue stroma through which lymphocytes
migrate in vivo. Lymphocytes were observed to migrate in a “random walk” fashion, and lym-
phocytes seeded onto the surface of the collagen gel were observed to penetrate into the gel
matrix. This surprising observation of cell infiltration into the collagen gel did not appear to
involve collagenolytic activity and is hypothesized to be a consequence of the uniquely small
size (7 lm) and fast speed (10 lm/min) of T cells. Gunzer et al. expanded upon this paper and
used DCs to activate na€
ıve T cells in a 3D collagen gel.
14
This work provided the first indica-
tion that T cell activation can occur in a 3D in vitro model and provided insights into a T cell’s
migration pattern after encountering antigen-loaded DCs. Both papers used physically cross-
linked collagen and observed T cell migration rates similar to in vivo rates of 10–11 lm/min.
Several other biopolymers are present in the T cell’s native microenvironment, including
hyaluronan and laminin. However, use of naturally derived ECM polymers in in vitro studies
have the limitations of batch-to-batch variability, undergoing microarchitectural remodeling
through protease degradation, and having a limited range of tunability in terms of mechanical
and biochemical properties. Recently, a variety of naturally derived ECM polymers and syn-
thetic polymers have been chemically modified to create semisynthetic hydrogel systems that
address many of these limitations.
122
Due to their control over matrix mechanical properties,
these materials may be well-suited for future studies of T cell mechanotransduction in 3D.
However, since most synthetic polymer scaffolds typically have pore sizes that are much
smaller than those found in natural-polymer scaffolds, typically cells seeded onto the top sur-
face of these gels will not be able to infiltrate into the scaffold. Thus, the cells are usually pre-
mixed with the gel precursors in order to attain high levels of homogeneous cell-loading. This
fabrication process can expose the cells to potential damage from crosslinking reagents or tran-
sient exposure to non-physiological environmental conditions, such as low pH or temperature,
to induce gelation of the scaffold. In addition, the small pore sizes in these synthetic systems
typically require that the scaffold must degrade or remodel over time in order to enable migra-
tion of encapsulated cells.
123
An alternative approach to promote cell migration into 3D scaffolds is the use of inverted
colloidal crystals (ICCs), also referred to as inverse opal hydrogels [Fig. 7(b)]. In the ICC
method, a sacrificial porogen is used to create a macro-porous scaffold; thereby providing an
isotropic 3D environment, with long range order, uniform interconnectivity, and tunable pore
size, typically within the range of 20–500 lm. The ICC structure can achieve a high porosity of
74%, which facilitates oxygen and nutrient diffusion and the homogeneous distribution of
cells throughout the scaffold.
124,125
T cells have been reported to be sensitive to oxygen trans-
port within 3D collagen gels, thus, this is another important consideration when designing a 3D
platform for study of T cells.
126
Macro-porous systems may be categorized as either 2.5D or
3D platforms depending on the pore size. When the scaffold pore size is much larger than the
dimension of a cell, the cell effectively encounters a 2.5D local environment and can utilize the
associated void space to efficiently migrate [Fig. 5(b)]. In contrast, if the pore size is about the
size of a single cell or cell cluster, then the cell may become entrapped within an effectively
3D local environment [Fig. 5(b)]. Stachowiak and Irvine et al. developed a composite of an
ICC-patterned PEG hydrogel infused with fibrillar collagen as a lymphoid tissue model to study
immune cell migration.
120,127
The highly motile T cell migratory patterns observed within this
scaffold were similar to those found in native secondary lymphoid organs. Because a wide
range of different materials with tunable mechanical properties could be used to fabricate ICC
scaffolds, this technique may be an attractive option to begin moving T cell mechanotransduc-
tion studies into 3D.
125
So far, only methods for T cell encapsulation in vitro have been reviewed. Bridging to the
in vivo setting, Monette et al. recently demonstrated that T cells can be encapsulated in a ther-
mogel and then injected into a rat to allow for localized delivery.
128
Specifically, the goal was
to deliver tumor-expanded CD8
þ
T cells to a tumor mass for cancer immunotherapy. The
021501-22 de la Zerda et al. APL Bioeng. 2, 021501 (2018)

injectable and biodegradable scaffold is chitosan-based, and it allows for T cell growth, prolif-
eration, and survival in vitro as well as in vivo for periods of up to 3 weeks.
129
Its mechanical
properties are tunable and can reach more physiologically relevant levels (0.1–5 kPa), thus it
can effectively mimic the stiffness experienced by cells within the LNs and other soft tissues.
These results together with the fact that thermogels have been used to study mechanotransduc-
tion in other cell types,
130
indicate that this chitosan-based thermogel formulation may be prom-
ising in studying 3D T cell mechanobiology.
When designing a material platform to mimic the mechanical properties of native tissue, it
may also be critical to consider the time-dependent mechanics of the ECM. Native tissues are
viscoelastic, i.e., they possess both time-independent elastic stiffness and time-dependent vis-
cosity, which enables the matrix to undergo molecular-level remodeling and stress relaxation
over time [Fig. 7(c)]. However, most in vitro studies of cell mechanobiology to date have
focused only on the cellular effects of elastic stiffness. Recent work by Chaudhuri et al. and
Cooper-White et al. have demonstrated that cells can be exquisitely sensitive to the time-
dependent mechanical behavior of the matrix.
121,131
These effects are thought to be due to local
clustering of cell-adhesive ligands that can occur more readily on matrices that have faster rates
of stress relaxation. It remains to be seen if similar mechanisms may be important for T cells,
which are relatively low adherent compared to many other cell types.
132
To summarize, there is a significant trade-off between designing a 3D system that accu-
rately mimics the physiological complexity of the native microenvironment and yet is simplified
enough to maintain a feasible level of reproducibility. The study of 3D T cell mechanobiology
is nascent, and there are many experimental variables that need to be explored in order to con-
duct a reliable and biologically relevant experiment.
E. Emerging methods to explore T cell mechanobiology ex vivo and in vivo: Lysyl
oxidase (LOX) and b-aminopropionitrile (bAPN)
To date, research in T cell mechanotransduction has been exclusively performed in vitro.
However, to gain insights that are truly physiologically relevant, it may be necessary to develop
experimental protocols that work with in vivo or ex vivo systems. Here we discuss one potential
opportunity to pursue these studies using the lysyl oxidase (LOX) family of proteins (Fig. 8).
Many diseases that are linked to ECM mechanical changes have also been shown to have
abnormal levels of LOX. While most healthy tissues have a low level of LOX expression
(which is secreted by fibroblasts, smooth muscle cells, osteoblasts, and vascular endothelium),
various diseases including cancer are associated with either upregulation or downregulation of
LOX.
133,134
FIG. 8. Potential strategy to study mechanotransduction in vivo using lysyl oxidase (LOX). Left, schematic of collagen
fibrils (beige) with a typical, basal-level of lysyl oxidase (LOX, red). Top right, increased levels of LOX lead to increased
covalent crosslinking of adjacent collagen fibers, leading to matrix stiffening. Bottom right, depletion or inhibition of LOX
disrupts collagen crosslinking and prevents stiffening.
021501-23 de la Zerda et al. APL Bioeng. 2, 021501 (2018)

LOX is an enzyme family that crosslinks collagen and elastin through oxidative deamination
of lysine residues to form a semialdehyde, which subsequently enables covalent crosslinking to
adjacent collagen fibers and ECM stiffening.
134–136
On the other hand, inhibition of LOX, which
can be achieved using blocking antibodies, specific RNA interference, or b-aminopropionitrile
(bAPN), disrupts this stiffening.
137
bAPN is a small, irreversible, competitive inhibitor that targets
the catalytic domain of LOX.
138
Importantly, inhibition of LOX with bAPN in vivo can alter the
matrix mechanics without changing the matrix composition, fiber density, and fiber organization,
opening the door to mechanistic studies.
137
Manipulation of LOX activity in vivo and ex vivo has
been used to probe the role of tissue mechanics in the progression of cancer.
135
For example,
Levental et al. demonstrated that upregulation of LOX resulted in stiffer tissue and greater tumor
invasion, while inhibiting LOX resulted in more compliant tissue and reduction in metastatic
spread. Furthermore, LOX upregulation correlated with increased integrin clustering and subse-
quent enhancement of growth factor signaling. As CD8
þ
T cells have been implicated in cancer
progression, one interesting avenue may be to evaluate the effects of this LOX/bAPN treatment
on local T cell responses.
V. CONCLUSION
Through the course of this review, we have highlighted some of the key studies that pro-
vide the foundation for understanding T cell mechanobiology, with a focus on the engineering
techniques that enabled these studies. Continuing to improve our understanding of T cell
mechanobiology will require close collaborations between researchers with both engineering
and immunology backgrounds. As has been pointed out throughout this review, in many instan-
ces, the techniques for more sophisticated mechanobiology experiments have been demonstrated
already with other cell types, but are yet to be applied to questions fundamental to T cell biol-
ogy. In other instances, new engineering techniques may need to be developed to both quantita-
tively perturb and measure mechanical forces imposed on, generated by, and found within T
cells. Importantly, attention to using materials and forces that recapitulate physiological condi-
tions will be vital to understanding T cell behaviors in both healthy and diseased tissues.
Improved understanding of T cell mechanobiology will be valuable for both fundamental
and applied immunology. Although currently underappreciated, the ability of mechanical forces
to dramatically alter T cell activation may be exploited in the future to control T cell behavior
in a variety of immunotherapies. These might include enhancement of T cell activation to better
fight off infection or suppression of T cell activation to better regulate autoimmunity. In addi-
tion, as engineered T cell therapies become more prevalent, T cell mechanobiology may also
offer new opportunities to efficiently expand and program T cells ex vivo. Thus, by merging the
disciplines of bioengineering, mechanobiology, and T cell immunology, new fundamental and
translational discoveries will be made.
ACKNOWLEDGMENTS
The authors acknowledge funding support from the National Science Foundation (NSF)
Graduate Research Fellowship Program and Stanford Bio-X (A.dlZ.), the Stanford ChEM-H
Postdoctoral Training Program in Quantitative Mechanobiology (M.J.K.), the Stanford Graduate
Training Program in Biotechnology National Institutes of Health (NIH) T32 GM008412 (N.A.S.),
and NSF DMR 1508006 and NIH U19 AI116484 (S.C.H.). The authors thank Professor Paul
Bollyky (Stanford Department of Medicine, Infectious Diseases) for the helpful discussions.
1
D. E. Ingber, Ann. Med. 35(8), 564 (2003).
2
M. J. Paszek, N. Zahir, K. R. Johnson, J. N. Lakins, G. I. Rozenberg, A. Gefen, C. A. Reinhart-King, S. S. Margulies, M.
Dembo, D. Boettiger, D. A. Hammer, and V. M. Weaver, Cancer Cell 8(3), 241 (2005); M. Janakiram, Y. M. Abadi, J.
A. Sparano, and X. Zang, Discovery Med. 14(77), 229 (2012).
3
M. L. Previtera, Cell. Mol. Bioeng. 7(3), 473 (2014).
4
I. Levental, P. C. Georges, and P. A. Janmey, Soft Matter 3(3), 299 (2007).
5
K. Miyaji, A. Furuse, J. Nakajima, T. Kohno, T. Ohtsuka, K. Yagyu, T. Oka, and S. Omata, Cancer 80(10), 1920 (1997).
6
K. Arda, N. Ciledag, E. Aktas, B. K. Arıbas, and K. K€
ose, Am. J. Roentgenol. 197(3), 532 (2011).
021501-24 de la Zerda et al. APL Bioeng. 2, 021501 (2018)

7
J. Weickenmeier, R. de Rooij, S. Budday, P. Steinmann, T. C. Ovaert, and E. Kuhl, Acta Biomater. 42(Supplement C),
265 (2016); S. Mueller and L. Sandrin, Hepatic Med. 2, 49 (2010); N. Sachot, E. Engel, and O. Castano, Curr. Org.
Chem. 18(18), 2299 (2014); N. Nagy, A. de la Zerda, G. Kaber, P. Y. Johnson, K. H. Hu, M. J. Kratochvil, K. Yadava,
W. Zhao, Y. Cui, G. Navarro, J. P. Annes, T. N. Wight, S. C. Heilshorn, P. L. Bollyky, and M. J. Butte, J. Biol. Chem.
293, 567 (2017).
8
A. Trautmann and S. Valitutti, Curr. Opin. Immunol. 15(3), 249 (2003).
9
R. Basu, B. M. Whitlock, J. Husson, A. Le Floc’h, W. Jin, A. Oyler-Yaniv, F. Dotiwala, G. Giannone, C. Hivroz, N.
Biais, J. Lieberman, L. C. Kam, and M. Huse, Cell 165(1), 100 (2016).
10
Z. Wan, S. Zhang, Y. Fan, K. Liu, F. Du, A. M. Davey, H. Zhang, W. Han, C. Xiong, and W. Liu, J. Immunol. 190(9),
4661 (2013); E. Natkanski, W. Y. Lee, B. Mistry, A. Casal, J. E. Molloy, and P. Tolar, Science 340(6140), 1587 (2013).
11
N. Bufi, M. Saitakis, S. Dogniaux, O. Buschinger, A. Bohineust, A. Richert, M. Maurin, C. Hivroz, and A. Asnacios,
Biophys. J. 108(9), 2181 (2015).
12
T. J. Thauland, K. H. Hu, M. A. Bruce, and M. J. Butte, Sci. Signaling 10(469), eaah3737 (2017).
13
S. Valitutti, M. Dessing, K. Aktories, H. Gallati, and A. Lanzavecchia, J. Exp. Med. 181(2), 577 (1995).
14
M. Gunzer, A. Schafer, S. Borgmann, S. Grabbe, K. S. Zanker, E. B. Brocker, E. Kampgen, and P. Friedl, Immunity
13(3), 323 (2000).
15
W. Chen and C. Zhu, Immunol. Rev. 256(1), 160 (2013).
16
Z. Ma and T. H. Finkel, Trends Immunol. 31(1), 1 (2010).
17
Y. C. Li, B. M. Chen, P. C. Wu, T. L. Cheng, L. S. Kao, M. H. Tao, A. Lieber, and S. R. Roffler, J. Immunol. 184(11),
5959 (2010).
18
R. S. O’Connor, X. Hao, K. Shen, K. Bashour, T. Akimova, W. W. Hancock, L. C. Kam, and M. C. Milone, J. Immunol.
189(3), 1330 (2012).
19
E. Judokusumo, E. Tabdanov, S. Kumari, M. L. Dustin, and L. C. Kam, Biophys. J. 102(2), L5 (2012).
20
K. L. Hui, L. Balagopalan, L. E. Samelson, and A. Upadhyaya, Mol. Biol. Cell 26(4), 685 (2015).
21
J. Q. Hu, A. A. Gondarenko, A. P. Dang, K. T. Bashour, R. S. O’Connor, S. Lee, A. Liapis, S. Ghassemi, M. C. Milone,
M. P. Sheetz, M. L. Dustin, L. C. Kam, and J. C. Hone, Nano Lett. 16(4), 2198 (2016).
22
G. Giannone and M. P. Sheetz, Trends Cell Biol. 16(4), 213 (2006).
23
V. Vogel and M. P. Sheetz, Curr. Opin. Cell Biol. 21(1), 38 (2009).
24
G. R. Fedorchak, A. Kaminski, and J. Lammerding, Prog. Biophys. Mol. Biol. 115(2-3), 76 (2014).
25
A. M. Kruisbeek, E. Shevach, and A. M. Thornton, Current Protocols in Immunology (John Wiley & Sons, Inc., 2001);
B. L. Levine, W. B. Bernstein, M. Connors, N. Craighead, T. Lindsten, C. B. Thompson, and C. H. June, J. Immunol.
159(12), 5921 (1997).
26
B. Malissen and P. Bongrand, Annu. Rev. Immunol. 33, 539 (2015).
27
L. Balagopalan, R. L. Kortum, N. P. Coussens, V. A. Barr, and L. E. Samelson, J. Biol. Chem. 290(44), 26422 (2015).
28
J. S. Boomer and J. M. Green, Cold Spring Harbor Perspect. Biol. 2(8), a002436 (2010); K. D. Carey, T. J. Dillon, J. M.
Schmitt, A. M. Baird, A. D. Holdorf, D. B. Straus, A. S. Shaw, and P. J. Stork, Mol. Cell. Biol. 20(22), 8409 (2000).
29
B. Alarc
on, D. Mestre, and N. Mart
ınez-Mart
ın, Immunology 133(4), 420 (2011).
30
(a) T. Saito, T. Yokosuka, and A. Hashimoto-Tane, FEBS Lett. 584(24), 4865 (2010); (b) S. Kumari, S. Curado, V.
Mayya, and M. L. Dustin, Biochim. Biophys. Acta 1838(2), 546 (2014).
31
N. Mueller, E. Avota, L. Collenburg, H. Grassme, and S. Schneider-Schaulies, PLoS Pathog. 10(12), e1004574 (2014);
M.-L. Alegre, K. A. Frauwirth, and C. B. Thompson, Nat. Rev. Immunol. 1(3), 220 (2001); N. Beyersdorf, T. Kerkau,
and T. H€
unig, Immunotargets Ther. 4, 111 (2015).
32
J. Delon and R. N. Germain, Curr. Biol.: CB 10(24), R923 (2000).
33
M. Oelke, M. V. Maus, D. Didiano, C. H. June, A. Mackensen, and J. P. Schneck, Nat. Med. 9(5), 619 (2003).
34
I. Olsen and L. M. Sollid, J. Immunol. Methods 390(1-2), 106 (2013).
35
Q. Han, N. Bagheri, E. M. Bradshaw, D. A. Hafler, D. A. Lauffenburger, and J. C. Love, Proc. Natl. Acad. Sci. U. S. A.
109(5), 1607 (2012).
36
J. Baran, D. Kowalczyk, M. Ozog, and M. Zembala, Clin. Diagn. Lab. Immunol. 8(2), 303 (2001); E. Donnadieu, G.
Bismuth, and A. Trautmann, Pfl€
ugers Arch. 429(4), 546 (1995).
37
T. Chatila, L. Silverman, R. Miller, and R. Geha, J. Immunol. 143(4), 1283 (1989).
38
T. Yahata, N. Abe, C. Yahata, Y. Ohmi, A. Ohta, K. Iwakabe, S. Habu, H. Yagita, H. Kitamura, N. Matsuki, M. Nakui,
M. Sato, and T. Nishimura, Eur. J. Immunol. 29(3), 727 (1999).
39
A. Trickett and Y. L. Kwan, J. Immunol. Methods 275(1-2), 251 (2003).
40
J. V. Kim, J. B. Latouche, I. Riviere, and M. Sadelain, Nat. Biotechnol. 22(4), 403 (2004).
41
C. J. Turtle and S. R. Riddell, Cancer J. 16(4), 374 (2010).
42
K. Perica, A. K. Kosmides, and J. P. Schneck, Biochim. Biophys. Acta (BBA) - Mol. Cell Res. 1853(4), 781 (2015); I.
Platzman, J.-W. Janiesch, J. Matic´, and J. P. Spatz, Israel J. Chem. 53(9-10), 655 (2013); J. van der Weijden, L. E.
Paulis, M. Verdoes, J. C. M. van Hest, and C. G. Figdor, Chem. Sci. 5(9), 3355 (2014); Nanomedicine 8(7), 1173
(2013); S. Adutler-Lieber, I. Zaretsky, I. Platzman, J. Deeg, N. Friedman, J. P. Spatz, and B. Geiger, J. Autoimmun.
54(Supplement C), 100 (2014). R. H. Fang and L. Zhang, Annu. Rev. Chem. Biomol. Eng. 7, 305 (2016).
43
P. S. Andersen, C. Menne, R. A. Mariuzza, C. Geisler, and K. Karjalainen, J. Biol. Chem 276(52), 49125 (2001).
44
M. Oelke and J. P. Schneck, Immunol. Res. 47(1), 248 (2010).
45
F. Ito, A. Carr, H. Svensson, J. Yu, A. E. Chang, and Q. Li, J. Immunother. 26(3), 222 (2003).
46
J. C. Sunshine, K. Perica, J. P. Schneck, and J. J. Green, Biomaterials 35(1), 269 (2014).
47
J. Husson, K. Chemin, A. Bohineust, C. Hivroz, and N. Henry, PloS One 6(5), e19680 (2011).
48
K. M. Friedman, L. E. Devillier, S. A. Feldman, S. A. Rosenberg, and M. E. Dudley, J. Immunother. 34(9), 651 (2011).
49
G. A. Papanicolaou, J. B. Latouche, C. Tan, J. Dupont, J. Stiles, E. G. Pamer, and M. Sadelain, Blood 102(7), 2498 (2003).
50
S. Sasawatari, T. Tadaki, M. Isogai, M. Takahara, M. Nieda, and K. Kakimi, Immunol. Cell Biol. 84(6), 512 (2006); S.
Kondo, A. Demachi-Okamura, T. Hirosawa, H. Maki, M. Fujita, Y. Uemura, Y. Akatsuka, E. Yamamoto, K. Shibata, K.
Ino, F. Kikkawa, and K. Kuzushima, Hum. Immunol. 74(9), 1103 (2013).
51
M. D. Cahalan and I. Parker, Semin. Immunol. 17(6), 442 (2005).
021501-25 de la Zerda et al. APL Bioeng. 2, 021501 (2018)

52
G. Shakhar, R. L. Lindquist, D. Skokos, D. Dudziak, J. H. Huang, M. C. Nussenzweig, and M. L. Dustin, Nat. Immunol.
6(7), 707 (2005).
53
S. D. Katzman, W. E. O’Gorman, A. V. Villarino, E. Gallo, R. S. Friedman, M. F. Krummel, G. P. Nolan, and A. K.
Abbas, Proc. Natl. Acad. Sci. U. S. A. 107(42), 18085 (2010); S. Hugues, L. Fetler, L. Bonifaz, J. Helft, F. Amblard, and
S. Amigorena, Nat. Immunol. 5(12), 1235 (2004); T. R. Mempel, S. E. Henrickson, and U. H. Von Andrian, Nature
427(6970), 154 (2004); M. J. Miller, O. Safrina, I. Parker, and M. D. Cahalan, J. Exp. Med. 200(7), 847 (2004).
54
J. Lee, M. Leonard, T. Oliver, A. Ishihara, and K. Jacobson, J. Cell Biol. 127(6 Pt 2), 1957 (1994); L. Dupr
e, R.
Houmadi, C. Tang, and J. Rey-Barroso, Front. Immunol. 6(586), 1704 (2015).
55
Q. Ma, Y. Wang, A. S. Lo, E. M. Gomes, and R. P. Junghans, J. Biomed. Biotechnol. 2010, 386545.
56
I. Tskvitaria-Fuller, A. L. Rozelle, H. L. Yin, and C. Wulfing, J. Immunol. 171(5), 2287 (2003); M. F. Krummel, M. D.
Sjaastad, C. Wulfing, and M. M. Davis, Science 289(5483), 1349 (2000).
57
A. Kupfer and G. Dennert, J. Immunol. 133(5), 2762 (1984).
58
P. Cullinan, A. I. Sperling, and J. K. Burkhardt, Immunol. Rev. 189, 111 (2002).
59
K. H. Hu and M. J. Butte, J. Cell Biol. 213(5), 535 (2016).
60
K. Badour, J. Zhang, and K. A. Siminovitch, Immunol. Rev. 192, 98 (2003).
61
M. Vicente-Manzanares and F. Sanchez-Madrid, Nat. Rev. Immunol. 4(2), 110 (2004).
62
D. D. Billadeau, J. C. Nolz, and T. S. Gomez, Nat. Rev. Immunol. 7(2), 131 (2007).
63
T. S. Gomez and D. D. Billadeau, Adv. Immunol. 97, 1 (2008).
64
P. Beemiller and M. F. Krummel, Immunol. Rev. 256(1), 148 (2013); J. K. Burkhardt, E. Carrizosa, and M. H. Shaffer,
Annu. Rev. Immunol. 26, 233 (2008); Y. Yu, A. A. Smoligovets, and J. T. Groves, J. Cell Sci. 126(Pt 5), 1049 (2013).
65
J. Delon, N. Bercovici, G. Raposo, R. Liblau, and A. Trautmann, J. Exp. Med. 188(8), 1473 (1998); J. Delon, N.
Bercovici, R. Liblau, and A. Trautmann, Eur. J. Immunol. 28(2), 716 (1998).
66
N. Joseph, B. Reicher, and M. Barda-Saad, Biochim. Biophys. Acta 1838(2), 557 (2014); S. Feske, Nat. Rev. Immunol.
7(9), 690 (2007).
67
R. Gonz
alez-Amaro, J. R. Cort
es, F. S
anchez-Madrid, and P. Mart
ın, Trends Mol. Med. 19(10), 625 (2013).
68
P. E. Simms and T. M. Ellis, Clin. Diagn. Lab. Immunol. 3(3), 301 (1996); K. Tkach and G. Altan-Bonnet, Curr. Opin.
Immunol. 25(1), 120 (2013).
69
S. F. Ziegler, F. Ramsdell, and M. R. Alderson, Stem Cells 12(5), 456 (1994).
70
N. Blanchard, M. Decraene, K. Yang, F. Miro-Mur, S. Amigorena, and C. Hivroz, J. Immunol. 173(5), 3062 (2004).
71
A. Sica, L. Dorman, V. Viggiano, M. Cippitelli, P. Ghosh, N. Rice, and H. A. Young, J. Biol. Chem. 272(48), 30412 (1997).
72
N. Chaffey, Ann. Bot. 91(3), 401 (2003).
73
K. A. Smith, S. Gillis, P. E. Baker, D. McKenzie, and F. W. Ruscetti, Ann. N. Y. Acad. Sci. 332, 423 (1979).
74
V. Hurez, A. Saparov, A. Tousson, M. J. Fuller, T. Kubo, J. Oliver, B. T. Weaver, and C. T. Weaver, J. Exp. Med.
198(1), 123 (2003).
75
A. Weiss, R. Shields, M. Newton, B. Manger, and J. Imboden, J. Immunol. 138(7), 2169 (1987).
76
B. Fazekas de St Groth, A. L. Smith, and C. A. Higgins, Immunol. Cell Biol. 82(3), 260 (2004).
77
S. E. Henrickson, T. R. Mempel, I. B. Mazo, B. Liu, M. N. Artyomov, H. Zheng, A. Peixoto, M. P. Flynn, B. Senman,
T. Junt, H. C. Wong, A. K. Chakraborty, and U. H. von Andrian, Nat. Immunol. 9(3), 282 (2008).
78
G. Lezzi, K. Karjalainen, and A. Lanzavecchia, Immunity 8(1), 89 (1998).
79
J. K. Wang, S. T. Ju, and A. Marshak-Rothstein, Eur. J. Immunol. 30(3), 931 (2000).
80
J. J. Boniface, J. D. Rabinowitz, C. Wulfing, J. Hampl, Z. Reich, J. D. Altman, R. M. Kantor, C. Beeson, H. M.
McConnell, and M. M. Davis, Immunity 9(4), 459 (1998); S. Casares, C. S. Zong, D. L. Radu, A. Miller, C. A. Bona,
and T. D. Brumeanu, J. Exp. Med. 190(4), 543 (1999).
81
J. R. Cochran, T. O. Cameron, and L. J. Stern, Immunity 12(3), 241 (2000).
82
B. C. Chen, W. R. Legant, K. Wang, L. Shao, D. E. Milkie, M. W. Davidson, C. Janetopoulos, X. S. Wu, J. A. Hammer
III, Z. Liu, B. P. English, Y. Mimori-Kiyosue, D. P. Romero, A. T. Ritter, J. Lippincott-Schwartz, L. Fritz-Laylin, R. D.
Mullins, D. M. Mitchell, J. N. Bembenek, A. C. Reymann, R. Bohme, S. W. Grill, J. T. Wang, G. Seydoux, U. S. Tulu,
D. P. Kiehart, and E. Betzig, Science 346(6208), 1257998 (2014).
83
Z. Ma, P. A. Janmey, and T. H. Finkel, FASEB J. 22(4), 1002 (2008); A. Lanzavecchia, G. Iezzi, and A. Viola, Cell
96(1), 1 (1999).
84
S. T. Kim, K. Takeuchi, Z. Y. Sun, M. Touma, C. E. Castro, A. Fahmy, M. J. Lang, G. Wagner, and E. L. Reinherz,
J. Biol. Chem. 284(45), 31028 (2009).
85
Y. Wu, K. Liu, K. Song, and S. Pan, Appl. Biochem. Biotechnol. 165(2), 485 (2011).
86
P. A. van der Merwe and S. J. Davis, Annu. Rev. Immunol. 21, 659 (2003).
87
Y. Feng, K. N. Brazin, E. Kobayashi, R. J. Mallis, E. L. Reinherz, and M. J. Lang, Proc.Natl.Acad.Sci.U.S.A.114(39),
E8204 (2017).
88
L. M. Lee and A. P. Liu, J. Nanotechnol. Eng. Med. 5(4), 0408011 (2014); M. Unal, Y. Alapan, H. Jia, A. G. Varga, K.
Angelino, M. Aslan, I. Sayin, C. Han, Y. Jiang, Z. Zhang, and U. A. Gurkan, Nanobiomedicine 1, 5 (2014).
89
B. Liu, W. Chen, B. D. Evavold, and C. Zhu, Cell 157(2), 357 (2014); Y. Chen, B. Liu, L. Ju, J. Hong, Q. Ji, W. Chen,
and C. Zhu, J. Vis. Exp. 2015(102), e52975.
90
D. Smit, C. Fouquet, M. Doulazmi, F. Pincet, A. Trembleau, and M. Zapotocky, BMC Biophys. 10, 2 (2017); E. Evans,
K. Ritchie, and R. Merkel, Biophys. J. 68(6), 2580 (1995); W. Chen, E. A. Evans, R. P. McEver, and C. Zhu, Biophys. J.
94(2), 694 (2008).
91
C. Gourier, A. Jegou, J. Husson, and F. Pincet, Cell. Mol. Bioeng. 1(4), 263 (2008).
92
P. Tseng, J. W. Judy, and D. Di Carlo, Nat. Methods 9(11), 1113 (2012).
93
Y. Liu, L. Blanchfield, V. P. Ma, R. Andargachew, K. Galior, Z. Liu, B. Evavold, and K. Salaita, Proc. Natl. Acad. Sci.
U. S. A. 113(20), 5610 (2016).
94
S. Pryshchep, V. I. Zarnitsyna, J. Hong, B. D. Evavold, and C. Zhu, J. Immunol. 193(1), 68 (2014).
95
M. G. Haugh and S. C. Heilshorn, Curr. Opin. Solid State Mater. 20(4), 171 (2016); W. L. Murphy, T. C. McDevitt, and
A. J. Engler, Nat. Mater. 13(6), 547 (2014).
96
B. Trappmann, J. E. Gautrot, J. T. Connelly, D. G. Strange, Y. Li, M. L. Oyen, M. A. Cohen Stuart, H. Boehm, B. Li, V.
Vogel, J. P. Spatz, F. M. Watt, and W. T. Huck, Nat. Mater. 11(7), 642 (2012).
021501-26 de la Zerda et al. APL Bioeng. 2, 021501 (2018)

97
M. Saitakis, S. Dogniaux, C. Goudot, N. Bufi, S. Asnacios, M. Maurin, C. Randriamampita, A. Asnacios, and C. Hivroz,
eLife 6, e23190 (2017).
98
J. C. Friedland, M. H. Lee, and D. Boettiger, Science 323(5914), 642 (2009).
99
H. Hosoi, H. Ikeda, N. Imai, C. Amaike, L. Wang, Y. Orito, M. Yamane, Hiroaki Ueno, M. Ideno, I. Nukaya, T. Enoki,
J. Mineno, K. Takesako, S. Hirano, and H. Shiku, Eur. J. Immunol. 44(6), 1747 (2014); Y. Shimizu, G. A. van Seventer,
K. J. Horgan, and S. Shaw, J. Immunol. 145(1), 59 (1990).
100
J. Fu, Y. K. Wang, M. T. Yang, R. A. Desai, X. Yu, Z. Liu, and C. S. Chen, Nat. Methods 7(9), 733 (2010).
101
K. T. Bashour, A. Gondarenko, H. Chen, K. Shen, X. Liu, H. Huse, J. C. Hone, and L. C. Kam, Proc. Natl. Acad. Sci.
U.S.A 111(6), 2241 (2014).
102
(a) T. C. Laurent and J. R. Fraser, FASEB J. 6(7), 2397 (1992); (b) J. Matic, J. Deeg, A. Scheffold, I. Goldstein, and J. P.
Spatz, Nano Lett. 13(11), 5090 (2013); (c) J. Deeg, M. Axmann, J. Matic, A. Liapis, D. Depoil, J. Afrose, S. Curado, M.
L. Dustin, and J. P. Spatz, ibid. 13(11), 5619 (2013).
103
W. R. Legant, J. S. Miller, B. L. Blakely, D. M. Cohen, G. M. Genin, and C. S. Chen, Nat. Methods 7(12), 969 (2010);
S. Munevar, Y. Wang, and M. Dembo, Biophys. J. 80(4), 1744 (2001); J. H. Wang and J. S. Lin, Biomech. Model.
Mechanobiol. 6(6), 361 (2007).
104
C. Franck, S. A. Maskarinec, D. A. Tirrell, and G. Ravichandran, PloS One 6(3), e17833 (2011).
105
W. R. Legant, C. K. Choi, J. S. Miller, L. Shao, L. Gao, E. Betzig, and C. S. Chen, Proc. Natl. Acad. Sci. U. S. A.
110(3), 881 (2013).
106
J. Doh and D. J. Irvine, Proc. Natl. Acad. Sci. U. S. A. 103(15), 5700 (2006).
107
K. Shen, V. K. Thomas, M. L. Dustin, and L. C. Kam, Proc. Natl. Acad. Sci. U. S. A. 105(22), 7791 (2008).
108
H. R. Jung, K. H. Song, J. T. Chang, and J. Doh, PloS One 9(3), e91926 (2014).
109
J. Deeg, M. Axmann, J. Matic, A. Liapis, D. Depoil, J. Afrose, S. Curado, M. L. Dustin, and J. P. Spatz, Nano Lett.
13(11), 5619 (2013).
110
K. W. Kwon, H. Park, and J. Doh, PloS One 8(9), e73960 (2013).
111
V. Vogel and M. Sheetz, Nat. Rev. Mol. Cell Biol. 7(4), 265 (2006); S. Liu, X. Xiong, X. Zhao, X. Yang, and H. Wang,
J. Hematol. Oncol. 8(1), 47 (2015).
112
M.Galic,S.Jeong,F.C.Tsai,L.M.Joubert,Y.I.Wu,K.M.Hahn,Y.Cui,andT.Meyer,Nat. Cell Biol. 14(8), 874 (2012).
113
M. Sundh, S. Svedhem, and D. S. Sutherland, J. Phys. Chem. B 115(24), 7838 (2011).
114
A. A. Brian and H. M. McConnell, Proc. Natl. Acad. Sci. U. S. A. 81(19), 6159 (1984).
115
K. D. Mossman, G. Campi, J. T. Groves, and M. L. Dustin, Science 310(5751), 1191 (2005).
116
B. M. Baker and C. S. Chen, J. Cell Sci. 125(Pt 13), 3015 (2012).
117
D. Wirtz, Annu. Rev. Biophys. 3, 301 (2009); A. Mukhopadhyay and S. Granick, Curr. Opin. Colloid Interface Sci. 6(5),
423 (2001); R. Xu, Particuology 18(Supplement C), 11 (2015).
118
J. K. Tan and T. Watanabe, Adv. Immunol. 105, 131 (2010); S. Suematsu and T. Watanabe, Nat. Biotechnol. 22(12),
1539 (2004); N. Okamoto, R. Chihara, C. Shimizu, S. Nishimoto, and T. Watanabe, J. Clin Invest. 117(4), 997 (2007);
D. J. Irvine, A. N. Stachowiak, and Y. Hori, Semin. Immunol. 20(2), 137 (2008); C. Giese, C. D. Demmler, R. Ammer,
S. Hartmann, A. Lubitz, L. Miller, R. Muller, and U. Marx, Artif. Organs 30(10), 803 (2006).
119
S. L. Schor, T. D. Allen, and B. Winn, J. Cell Biol. 96(4), 1089 (1983).
120
A. N. Stachowiak and D. J. Irvine, J. Biomed. Mater. Res. Part A 85A(3), 815 (2008).
121
O. Chaudhuri, L. Gu, D. Klumpers, M. Darnell, S. A. Bencherif, J. C. Weaver, N. Huebsch, H. P. Lee, E. Lippens, G. N.
Duda, and D. J. Mooney, Nat. Mater. 15(3), 326 (2016).
122
I. L. Kim, R. L. Mauck, and J. A. Burdick, Biomaterials 32(34), 8771 (2011); D. Loessner, C. Meinert, E. Kaemmerer,
L. C. Martine, K. Yue, P. A. Levett, T. J. Klein, F. P. W. Melchels, A. Khademhosseini, and D. W. Hutmacher, Nat.
Protoc. 11(4), 727 (2016).
123
A. M. Douglas, A. A. Fragkopoulos, M. K. Gaines, L. A. Lyon, A. Fernandez-Nieves, and T. H. Barker, Proc. Natl.
Acad. Sci. 114(5), 885 (2017); M. P. Lutolf, J. L. Lauer-Fields, H. G. Schmoekel, A. T. Metters, F. E. Weber, G. B.
Fields, and J. A. Hubbell, Proc. Natl. Acad. Sci. U. S. A. 100(9), 5413 (2003).
124
S. W. Choi, Y. Zhang, and Y. Xia, Langmuir: ACS J. Surf. Colloids 26(24), 19001 (2010).
125
Y. S. Zhang, S.-W. Choi, and Y. Xia, Soft Matter 9(41), 9747 (2013).
126
A. Pavesi, A. T. Tan, S. Koh, A. Chia, M. Colombo, E. Antonecchia, C. Miccolis, E. Ceccarello, G. Adriani, M. T.
Raimondi, R. D. Kamm, and A. Bertoletti, JCI Insight 2(12), e89762 (2017).
127
A. N. Stachowiak, A. Bershteyn, E. Tzatzalos, and D. J. Irvine, Adv. Mater. 17(4), 399 (2005).
128
A. Monette, C. Ceccaldi, E. Assaad, S. Lerouge, and R. Lapointe, Biomaterials 75, 237 (2016).
129
E. Assaad, M. Maire, and S. Lerouge, Carbohydr. Polym. 130, 87 (2015).
130
M. H. Park, H. J. Moon, J. H. Park, U. P. Shinde, D. Y. Ko, and B. Jeong, Macromol. Biosci. 15(4), 464 (2015).
131
O. Chaudhuri, L. Gu, M. Darnell, D. Klumpers, S. A. Bencherif, J. C. Weaver, N. Huebsch, and D. J. Mooney, Nat.
Commun. 6, 6364 (2015); A. R. Cameron, J. E. Frith, and J. J. Cooper-White, Biomaterials 32(26), 5979 (2011); D. D.
McKinnon, D. W. Domaille, J. N. Cha, and K. S. Anseth, Adv. Mater. 26(6), 865 (2014); A. R. Cameron, J. E. Frith, G.
A. Gomez, A. S. Yap, and J. J. Cooper-White, Biomaterials 35(6), 1857 (2014).
132
P. Friedl, K. S. Zanker, and E. B. Brocker, Microsc. Res. Tech. 43(5), 369 (1998).
133
M. W. Pickup, J. K. Mouw, and V. M. Weaver, EMBO Rep. 15(12), 1243 (2014); C. C. DuFort, M. J. Paszek, and V. M.
Weaver, Nat. Rev. Mol. Cell Biol. 12(5), 308 (2011); H. E. Barker, T. R. Cox, and J. T. Erler, Nat. Rev. Cancer 12(8),
540 (2012).
134
T. R. Cox and J. T. Erler, Dis. Models Mech. 4(2), 165 (2011).
135
K. R. Levental, H. Yu, L. Kass, J. N. Lakins, M. Egeblad, J. T. Erler, S. F. Fong, K. Csiszar, A. Giaccia, W. Weninger,
M. Yamauchi, D. L. Gasser, and V. M. Weaver, Cell 139(5), 891 (2009).
136
H. M. Kagan and W. Li, J. Cell. Biochem. 88(4), 660 (2003); Y. M. Kim, E. C. Kim, and Y. Kim, Mol. Biol. Rep. 38(1),
145 (2011).
137
J. E. Marturano, J. D. Arena, Z. A. Schiller, I. Georgakoudi, and C. K. Kuo, Proc. Natl. Acad. Sci. U. S. A. 110(16),
6370 (2013).
138
S. S. Tang, P. C. Trackman, and H. M. Kagan, J. Biol. Chem. 258(7), 4331 (1983).
021501-27 de la Zerda et al. APL Bioeng. 2, 021501 (2018)