INSIGHTS | PERSPECTIVES
704 19 MAY 2017 • VOL 356 ISSUE 6339 sciencemag.org SCIENCE
and that only these few high-affinity cells get
successfully selected (9). Likewise, the short
duration of the B cell–TFH cell interaction in
the GC probably permits a single TFH cell to
encounter many different GC B cells bearing
B cell receptors with various affinities and
thereby instructs TFH cells to select and help
only high-affinity GC B cells.
A possible mechanism underlying these
interactions could be controlling the dwell
time of TFH cells in the GC. Indeed, TFH cells
migrate constantly into and out of the GC
(10, 11), and during memory immune responses, some TFH cells can leave one GC
and migrate to another (12). The molecular
regulations of all these processes remain
largely elusive. By using two-photon microscopy and mice engineered to carry a B
cell–specific ephrin B1 deficiency, Lu et al.
show that GC B cell–expressed ephrin B1
limits the retention of TFH cells in the GC
by means of forward signaling through TFH-
expressed EPHB6. Furthermore, the authors
show that the formation of plasma cells is
defective in mice with ephrin B1–deficient
B cells despite adequate GC generation and
increased accumulation of TFH cells in the
GC (1, 13). Lu et al. clarify this seeming discrepancy by demonstrating that the amount
of TFH cell–produced interleukin-21 (IL-21),
which is an important factor for plasma cell
generation (14), relies on ephrin B1–
mediated forward signaling through EPHB4.
Thus, binding of GC B cell–expressed ephrin B1 to EPHB4 or EPHB6 on TFH cells regulates both the dwell time of TFH cells in the
GC and optimizes plasma cell formation.
Lu et al. provide insights into mechanisms
that orchestrate the organization of, and lymphocyte migration in, GCs. It will be interesting to see whether the EPH-ephrin system
also controls interaction durations between
other pairs of immune cells or immune cells
with stromal cells. j
REFERENCES AND NOTES
1. P. Lu, C. Shih, H. Qi,Science 356, eaai9264 (2017).
2. P. Friedl, Curr. Opin. Cell Biol. 16, 14 (2004).
3. E. B. Pasquale, Nat. Rev. Mol. Cell Biol. 6, 462 (2005).
4. L. Mesin, J. Ersching, G. D. Victora, Immunity 45, 471
5. C. G. Vinuesa et al. , Annu. Rev. Immunol. 34, 335 (2016).
6. C. D. Allen et al. , Science 315, 528 (2007).
7. T. Okada et al. , PLOS Biol. 3, e150 (2005).
8. D. Liu et al., Nature 517, 214 (2015).
9. A. D. Gitlin et al., Nature 509, 637 (2014).
10. H. Qi et al., Nature 455, 764 (2008).
11. Z. Shulman et al., Science341, 673 (2013).
12. D. Suan et al., Immunity 42, 704 (2015).
13. B. J. Laidlaw et al., J. Exp. Med. 214, 639 (2017).
14. D. Zotos et al. , J. Exp. Med. 207, 365 (2010).
R. F. is funded by the European Research Council (advanced
grant 322645, LYMPHATICS-HOMING), Deutsche
Forschungsgemeinschaft (SFB900-B1, SFB738-B5, Fo334/2-
2, Fo334/5-1), and the State of Lower Saxony (N-RENNT;
BIOFABRICATION for NIFE).
for materials modeling
Tracking defect motion in polycrystalline materials
can help model emergent properties
By Robert Suter
X-ray Bragg coherent diffractive im- aging (BCDI) provides an opportu- nity to connect the nanoscale world of crystalline lattice defects with the meso- and macroscale world of emergent materials behaviors. On
page 739 of this issue, Yau et al. (1) demonstrate that BCDI measurements that are
sensitive to individual lattice defects can
be carried out on micrometer-scale crystalline grains in complex environments.
Crystal defect motions and other structural
changes can be tracked. Extending these
measurements and combining them with
existing and developing mesoscale probes
will enable the connection of dynamics on
vastly different length scales.
Polycrystals are composed of many small
crystals (“grains”) held together by grain
boundaries. In a simple case, all grains have
the same crystallographic phase but are
delineated by unit cell orientation discontinuities, whereas in many technological
materials, different grains have different
compositions and phases, with the mixture
tailored to generate specific macroscopic
characteristics (such as hardness or electrical properties). However, all polycrystalline materials have complex multiscale
responses because the properties of single
crystals are anisotropic, and these properties are convolved with the distributions of
orientations and near-neighbor relations
(grain boundary types).
The fundamental mechanisms with
which materials respond to external con-
ditions operate on the nanoscale. For ex-
ample, during thermally induced growth
(coarsening) of grains, atomic rearrange-
ments at boundaries generate boundary
motions, and when mechanical forces in-
duce plastic deformation, line defects (dis-
locations) are generated and move. These
mechanisms are relatively well understood
in single crystals. However, to understand,
much less predict, macroscopic properties
of polycrystals requires modeling thou-
sands or millions of crystals together with
their mutually constraining interactions
(2). In applications where materials fail-
ures can be catastrophic (e.g., airframes
or turbine blades), statistical fluctuations
in the structure will lead to variations in
properties from sample to sample. With
the emergence of nondestructive, three-
dimensional (3D) probes of structure and
its evolution, measurements and models
can be tightly coupled from length scales of
a few tens of nanometers up to millimeters.
Although Yau et al. performed in situ
BCDI measurements during heat treatment
in a polycrystalline environment (at least
in a 2D one), a limitation of this and prior
BCDI measurements (3) is that only a single
Bragg peak from an essentially randomly
chosen grain is studied (see the figure).
The sample is scanned across the incident
beam until a grain is found that generates
scattering onto an area detector oriented at
an appropriate scattering angle, 2u. Once
located, the high-resolution, high–dynamic
range measurement proceeds. The observed
Bragg peak determines a single crystallographic direction or pole but is insensitive to rotations around that pole. Further,
the defect-induced atomic displacements
that are measured are projections of 3D
displacement vectors onto the measured
pole direction. If one started with a grain
with fully known orientation, several Bragg
peaks could be studied and the full 3D displacement field reconstructed.
Furthermore, to connect with emergent
behaviors, it will be necessary to study
clusters of neighboring grains. For example, it would be valuable to obtain BCDI
data in the twinned crystals studied by Yau
et al., but the lattice orientations of those
twins are unknown. Ideally, an orientation
Department of Physics, Carnegie Mellon University,
Pittsburgh, PA 15213, USA. Email: email@example.com
“…to connect with emergent
behaviors, it will be
necessary to study clusters
of neighboring grains.”