DOI: 10.1126/science.1254399, 800 (2014);344 Science

Quentin WilliamsDeep mantle matters

This copy is for your personal, non-commercial use only.

clicking here.colleagues, clients, or customers by , you can order high-quality copies for yourIf you wish to distribute this article to others

  here.following the guidelines

can be obtained byPermission to republish or repurpose articles or portions of articles

  ): May 27, 2014 www.sciencemag.org (this information is current as of

The following resources related to this article are available online at

http://www.sciencemag.org/content/344/6186/800.full.htmlversion of this article at:

including high-resolution figures, can be found in the onlineUpdated information and services,

http://www.sciencemag.org/content/344/6186/800.full.html#relatedfound at:

can berelated to this article A list of selected additional articles on the Science Web sites

http://www.sciencemag.org/content/344/6186/800.full.html#ref-list-1, 6 of which can be accessed free:cites 12 articlesThis article

http://www.sciencemag.org/cgi/collection/geochem_physGeochemistry, Geophysics

subject collections:This article appears in the following

registered trademark of AAAS. is aScience2014 by the American Association for the Advancement of Science; all rights reserved. The title

CopyrightAmerican Association for the Advancement of Science, 1200 New York Avenue NW, Washington, DC 20005. (print ISSN 0036-8075; online ISSN 1095-9203) is published weekly, except the last week in December, by theScience

on

May

27,

201

4w

ww

.sci

ence

mag

.org

Dow

nloa

ded

from

o

n M

ay 2

7, 2

014

ww

w.s

cien

cem

ag.o

rgD

ownl

oade

d fr

om

on

May

27,

201

4w

ww

.sci

ence

mag

.org

Dow

nloa

ded

from

800 23 MAY 2014 • VOL 344 ISSUE 6186 sciencemag.org SCIENCE

INSIGHTS | PERSPECTIVES

ILL

US

TR

AT

ION

: A

DA

PT

ED

FR

OM

E.

GA

RN

ER

O

The lower mantle, lying between ~670-

and ~2890-km depth, comprises most

of the rocky portion of Earth. Convec-

tion processes within this region trans-

fer heat from the iron core upward,

advecting the heat flow that drives

the near-surface tectonic engine. As many

of the largest volcanic events appear to be

correlated with seismic features in the deep

mantle ( 1), the deep lower mantle may also

represent an occasional scourge of our sur-

face environment. In this issue, two sets of

challenging experiments yield new pictures

for how different deep seismic anomalies

might be generated. On page 892, Andrault

et al. ( 2) examine the melting temperature of

oceanic crust (basalt) to core-mantle bound-

ary (CMB) pressures and temperature, and

use that to explain the genesis of areas with

ultralow seismic velocities near the CMB. On

page 877, Zhang et al. ( 3) report the startling

discovery of a new, iron-rich silicate phase

that may be a major component of the lower-

most ~700 km of Earth’s mantle.

Seismic probing has led to a richness of

phenomena being recognized in the inacces-

sible lowermost ~1000 km of Earth’s mantle

(see the figure). Large, blocky structures

beneath Africa and the Pacific with a ~3%

decrease in shear velocity extend up to 1000

km above the CMB (large low-shear-velocity

provinces, LLSVPs) ( 4). Regions with almost

horizontal discontinuities in seismic velocity

~100 to 300 km are present above the CMB

with velocity jumps of 2 to 3% ( 5) (these dis-

continuities define the anomalous D� zone at

the base of the mantle). There are also re-

gions with large reductions in compressional

and shear wave velocity in sporadic places in

the lowermost ~5 to 40 km of the mantle,

with shear velocity decreases as large as 45%

( 6) (ultralow-velocity zones, ULVZs). But all

of these features are volumetrically small

relative to the bulk of the mantle (LLSVPs

are at most a few percent of the lower man-

tle volume, while ULVZs represent a small

fraction of a percent). These features reside

within a lower mantle that has long been rec-

ognized to be dominated by a simple mix of

(Mg,Fe)SiO3-perovskite and (Mg,Fe)O [and

with possibly some exotic behavior associ-

ated with spin transitions in iron ( 7)].

Modeling and experiments at the extreme

pressures and temperatures of the deep

mantle (on the order of 100 GPa and 2000

K) reveal links between variations in phase,

composition, and temperature and the ob-

served seismic anomalies. One explanation

for LLSVPs is that they are associated with

slight iron and (Mg,Fe)SiO3 enrichment

( 8), the seismic discontinuities with a pres-

sure-induced transition from (Mg,Fe)SiO3-

perovskite to a post-perovskite structure

( 9), and the ULVZ with partial melting and

a possible increase in iron content ( 10). But

discerning how these anomalies are gener-

ated is difficult because our direct informa-

tion is confined to a seismic snapshot of the

present-day structure of this region.

The new studies provide process-oriented

constraints on the genesis of ULVZs and

chemical variations within the lowermost

mantle ( 2) and shift the view on the per-

ceived mineralogic simplicity of much of the

lower mantle ( 3). Andrault et al. study the

melting behavior of basalt (ocean crust) to

pressures of the CMB. Basalt is the lowest–

melting temperature rock likely to be present

near the CMB, and is expected to melt when

exposed to the temperatures of Earth’s outer

core ( 11). The picture that emerges is that

subducted plates may carry oceanic crust to

depths near the base of the mantle, where it

can partially melt, forming ULVZs. But An-

drault et al. propose that the melt can only

be retained within the unmelted residue of

the original basalt. Once it migrates into nor-

mal mantle, its chemistry is such that it will

react and form solid magnesian perovskite,

chemically remixing oceanic crust back into

the mantle. As a result, ULVZs should be gen-

erated from, and associated with, subducted

material at depth. Thus, the most abundant

magma at Earth’s surface may also produce

most of the melt near the bottom of Earth’s

rocky mantle—and plate tectonics may play a

fundamental role in the geochemistry of ma-

terial juxtaposed with Earth’s metallic core.

Zhang et al. observe the disproportion-

ation of (Mgx,Fe

1-x)SiO

3-perovskite (where

x within Earth’s mantle is ~0.9) to nearly

pure MgSiO3 and a new, iron-enriched ap-

proximately (Mg0.6

Fe0.4

)SiO3 phase. This “H-

phase” has a hexagonal structure and occurs

at pressures corresponding to depths greater

than ~2200 km. The synthesis of this phase

occurred at high temperatures (2200 K and

above), and this temperature and a low-

Deep mantle matters

Lower mantle

Outer core

ULVZ

CMB

1000 km

LLSVPPV

PPvD”

Disproportionation

reaction

The mantle region. Schematic of seismically characterized features occurring in the lowermost ~1000 km of Earth’s mantle, and experimentally proposed oceanic crustal-associated

ULVZ ( 2) and perovskite disproportionation boundary ( 3). (Note that the ULVZ vertical scale is exaggerated for visibility.) PV, perovskite; PPv, post-perovskite.

By Quentin Williams

Experiments reveal how some deep seismic anomalies near the core-mantle boundary might be generated

GEOPHYSICS

Department of Earth and Planetary Sciences, University of California Santa Cruz, Santa Cruz, CA 95064, USA. E-mail: qwilliam@ucsc.edu

Published by AAAS

23 MAY 2014 • VOL 344 ISSUE 6186 801SCIENCE sciencemag.org

The mammalian immune system both

suppresses and tolerates tumors, so

understanding this complexity should

benefit the development of cancer

therapies. Macrophages are proposed

to play an important role in suppress-

ing the immune response to cancer cells,

but it is not clear where these immune cells

come from or whether there are distinct

populations of macrophages with specific

roles in this setting. On page 921 of this is-

sue, Franklin et al. ( 1) forge a more coher-

ent view of macrophages that are associated

with tumor growth by assessing their ori-

gin, phenotype, and functions in an animal

model of breast cancer.

Tumor progression can be divided into

three phases—initiation, growth, and me-

tastasis (see the figure). The first phase is

characterized by the cell-autonomous accu-

mulation of genetic defects that leads to cell

transformation. This is followed by clonal

growth of transformed cells within the tis-

sue—the primary tumor site ( 2). Metastasis

results from the successful “engraftment”

of circulating tumor cells into secondary

locations where they proliferate after a

dormancy phase in which metastatic cells

remain quiescent ( 3). In both primary and

secondary tumor sites, the stroma, which

includes mesenchymal cells, macrophages,

and extracellular matrix (3), is thought to

play a role in the initial survival and prolif-

eration of transformed cells. However, as a

solid tumor grows and tumor cells acquire

the potential to escape the primary site, the

stroma becomes a more complex environ-

ment, with newly formed blood and lym-

phatic vessels and the recruitment and/or

proliferation of lymphoid and myeloid im-

mune cells ( 4). Immune cells are proposed

to prevent tumor progression via the elimi-

nation of immunogenic tumor cells by T

lymphocytes (CD8 subtype), a phenomenon

known as immunosurveillance (also called

immunoediting). During this process, tu-

mors that display either reduced immuno-

genicity or enhanced immunosuppressive

activity will escape elimination ( 5). Macro-

phages present in the tumor site can acti-

vate the immune response, but are mainly

thought to contribute to immunosuppres-

sion and tumor progression ( 6, 7), particu-

larly in the mammary gland ( 8). A high

density of macrophages in tumors is also

associated with worse overall survival in pa-

tients with gastric, urogenital, and head and

neck cancers, although it seems to be associ-

ated with better overall survival in patients

with colorectal cancer ( 7).

Franklin et al. carefully explore the con-

tribution of macrophages to tumor growth

in mice that develop a mammary cancer

that is genetically driven by the expres-

sion of an oncogene. In investigating the

development and differentiation of mac-

rophages in the normal mammary gland

and during the progression of a mammary

tumor, the authors identify a population of

macrophages that accumulates during tu-

mor growth called tumor-associated mac-

rophages (TAMs). These cells develop from

bone marrow–derived cells with the charac-

teristics of inflammatory monocytes, which

are recruited to the tumor where they differ-

entiate into macrophages and subsequently

proliferate. Franklin et al. observed that

when signaling by the protein Notch is pre-

vented in these TAMs, their differentiation

is blocked. Interestingly, TAMs are distinct

from macrophages present in normal mam-

mary tissue, which develop independently

of Notch signaling. Depletion of TAMs led to

a reduced tumor burden in the animal and

increased the cytotoxic potential of T lym-

phocytes present in the primary tumor site.

Thus, monocyte-derived Notch-dependent

TAMs are critical for tumor growth in this

mammary gland tumor model, at least in

Identifying the infiltrators

By Elisa Gomez Perdiguero and

Frederic Geissmann

Molecular characterization of macrophages reveals distinct types during tumorigenesis

CANCER IMMUNOLOGYstress environment appear to be critical for

synthesizing this phase.

So why has this new transition not been

observed seismically? One possibility is that

the velocity change across the transition may

be too small and/or the boundary may undu-

late dramatically in its depth. Alternatively,

the temperature of the deep mantle may lie

below the temperature of the disproportion-

ation reaction [which would require that the

mantle be a few hundred kelvin cooler than

currently inferred ( 12)—but this would also

imply that disproportionation could have

been important in the hotter past]. Another

option is that the oxidation state of iron in

the mantle may differ from those within the

experiments. The provocative aspects of this

discovery include not just changing the pos-

sible mineralogy of the deeper lower mantle,

but also that two phases of markedly differ-

ent densities are produced. Whether these

phases could undergo partial segregation,

thus enriching or depleting regions in the

H-phase (particularly in an earlier, hotter,

less viscous, and possibly partially molten

mantle), is unknown. If such segregation did

occur, a natural explanation for the genesis

of LLSVPs might exist. Depending on its elas-

ticity, an enrichment of H-phase within these

regions might provide an avenue to explain

their anomalous seismic signature.

Each of these experiments is the direct re-

sult of developments in high-pressure, high-

temperature techniques and the availability

of high-intensity synchrotron sources. Prob-

ing the sensitivity of the pressure and tem-

perature of melting and the phase transition

to variable oxygen fugacities, shifts in major

and minor elemental abundances, and vola-

tile contents holds the prospect of mapping

out the likely chemical behavior of the lower

mantle. In doing so, the current void of infor-

mation on the differentiation processes that

govern the chemical variations, structural

features, and evolution of Earth’s deepest

rocky reaches will be filled. ■

REFERENCES

1. J. Austermann et al., Geophys. J. Int. 197, 1 (2014). 2. D. Andrault et al., Science 344, 892 (2014). 3. L. Zhang et al., Science 344, 877 (2014). 4. S. Ni, D. V. Helmberger, J. Tromp, Geophys. J. Int. 161, 283

(2005). 5. M. Wysession et al., in The Core-Mantle Boundary Region,

M. Gurnis, M. Wysession, E. Knittle, B. Buffett, Eds., Geodyn. Ser. (American Geophysical Union, Washington, DC, 1998), vol. 28, 319–334.

6. M. S. Thorne, E. J. Garnero, G. Jahnke, H. Igel, A. K. McNamara, Earth Planet. Sci. Lett. 364, 59 (2013).

7. J. F. Lin et al., Rev. Geophys. 51, 244 (2013). 8. F. Deschamps, L. Cobden, P. J. Tackley, Earth Planet. Sci.

Lett. 349-350, 198 (2012). 9. M. Murakami et al., Science 304, 855 (2004). 10. S. Rost et al., Nature 435, 666 (2005). 11. S. Anzellini et al., Science 340, 464 (2013). 12. T. Katsura, A. Yoneda, D. Yamazaki, T. Yoshino, E. Ito, Phys.

Earth Planet. Inter. 183, 212 (2010).

10.1126/science.1254399ILL

US

TR

AT

ION

: V

. A

LT

OU

NIA

N/SCIENCE

Published by AAAS