标题: Mesenchymal stem cells within tumour stroma promote breast cancer metastasis [打印本页] 作者: gkshl 时间: 2012-5-17 17:15 标题: Mesenchymal stem cells within tumour stroma promote breast cancer metastasis
Mesenchymal stem cells have been recently described to localize to breast carcinomas, where they integrate into the 4 @/ `2 W- x% e0 K( Y; itumour-associated stroma. However, the involvement of mesenchymal stem cells (or their derivatives) in tumour . Q) a% d( h4 ^" A& d! cpathophysiology has not been addressed. Here, we demonstrate that bone-marrow-derived humanmesenchymal stem cells,8 a: T ?+ a4 b1 D+ H
when mixed with otherwise weakly metastatic human breast carcinoma cells, cause the cancer cells to increase their$ U' c" h/ [( o- h5 w
metastatic potency greatly when this cell mixture is introduced into a subcutaneous site and allowed to form a tumour , i' Q) H+ D8 z# v4 p9 p$ e; Fxenograft. The breast cancer cells stimulate de novo secretion of the chemokine CCL5 (also called RANTES) from3 ~+ w' R9 u% D- [( X$ B t
mesenchymal stem cells, which then acts in a paracrine fashion on the cancer cells to enhance their motility, invasion and) H* @- P: |0 @( j1 I2 W) I
metastasis. This enhanced metastatic ability is reversible and is dependent on CCL5 signalling through the chemokine# t3 X+ T9 @. _7 g- J' X1 a
receptor CCR5. Collectively, these data demonstrate that the tumour microenvironment facilitates metastatic spread by8 R) |; f! B7 K m0 i) B, z5 Z' `
eliciting reversible changes in the phenotype of cancer cells.# {8 [2 i" H# O, _* n
The origins of the invasive and metastatic phenotypes of carcinoma) u& a) o" X N2 B( V6 y) @) c s
cells have been the subjects of intense investigation. Whereas some$ n, i. Z# Z) C# m
current models depict these phenotypes as cell-autonomous alterations 1 t z2 s# o1 f+ U) v% Nspecified by the genomes of cancer cells, alternative views propose/ a6 F5 {- M1 |, `! s
that metastatic traits are acquired through exposure of epithelial+ V- m% y3 G& U1 k
cancer cells to paracrine signals that they receive from mesenchymal) D4 b; T K" A
cell types within the tumour-associated stroma. Although several b5 l! `7 m( T4 Y3 zlines of evidence demonstrate the contributions of stromal cells to/ @$ @/ a t5 T' r+ H/ R) [
primary tumour growth1, direct experimental demonstration of the _& G! E* C6 Ninfluence of these various cells on the metastatic abilities of cancer ' m) o$ Q" s/ P9 m p. B2 G# A$ Zcells has been difficult to obtain. This is due, in part, to the complexity 9 l# A; h! |1 ^' q7 Q( D Q* p6 A9 }6 Mof the mesenchymal cell types that are recruited into the stroma, and) R& C. l w$ A0 u- h
to the elusive nature of the putative paracrine signals that are 3 `8 @: O; j9 t# T) ~% T. pexchanged between the mesenchymal and epithelial compartments & p; m) l- B: f/ Z+ J) Rof a tumour. Recent reports proposed that the bone-marrow-derived: s5 y) n- H" Z% k: @
mesenchymal stem cell (MSC) is a cell type that is recruited in large 7 z3 ~" r- ]* e1 S: F* Dnumbers to the stroma of developing tumours2. To characterize better' Q) e6 s+ X) K& P
the role of this stromal cell in tumorigenesis, we set out to determine5 i2 v$ p3 i+ T& r$ o/ T2 ^2 v1 ]
whether MSCs could supply contextual signals that serve to% Q% L0 s8 ]8 _6 N0 S
promote cancer metastasis.! A# g/ Y% l$ s" j- ^& M
Mesenchymal stem cells are pluripotent progenitor cells that contribute 9 E" Q6 t* K7 }4 {; o( j' hto the maintenance and regeneration of a variety of connective) B, |% c- ?/ o) ]
tissues, including bone, adipose, cartilage and muscle3. Although / a( X; P+ N" k# \MSCs reside predominantly in the bone marrow, they are also distributed 7 ^4 `* }) ]$ z; u6 F. _) Q( _( M# ?throughout many other tissues, where they are thought to$ a' [4 h# p0 c7 `
serve as local sources of dormant stem cells4,5. The contributions of ! X! t% j7 E: O# mMSCs to tissue formation become apparent only in cases of tissue6 @6 G2 ]% A. f' J! L7 ~ `% V
remodelling after injury or chronic inflammation. These conditions* ?5 l4 M' T' f. V
are typically accompanied by the release of specific endocrinal signals 9 _: R7 _" t, M% ~. V( Ifrom the injured or inflamed tissue that are then transmitted to the % u8 w4 M3 o; h: s8 e" ^bone marrow, leading to the mobilization of multi-potent MSCs and $ \8 E' g6 H9 P: h! N! Ftheir subsequent recruitment to the damage site6. For example, MSCs ' S( |& p0 y7 K1 Xhave been shown to contribute to the formation of fibrous scars after 9 {7 W/ Y8 v. E0 X0 finjury7.2 k) Y' Z7 e' r. L0 J9 {
The formation of breast carcinomas is often accompanied by a , ~- z$ F7 u7 z' p) B4 ^well-orchestrated desmoplastic reaction, which involves the recruitment 9 C9 @4 C5 w/ k, h3 A) lof a variety of stromal cells with both pro- and anti-tumorigenic : w* s0 g: d) E) ?) U" A- Lactivities1. Such response closely resembles wound healing and scar + P% E" Y# [. P9 \+ p: o' A; Yformation, and entails the constant deposition of growth factors, 3 h1 ~3 T* Y$ g% Tcytokines and matrix-remodelling proteins that render the tumour ' G- j+ g; z. f6 k- o3 [0 b3 @- @, C6 `site a ‘wound that never heals’8. This suggests that, similar to sites of, s) \1 S1 m" p$ X% t9 D
injury, actively growing tumours recruit MSCs through the release of3 S3 O6 A( |& a8 b5 X
various endocrine and paracrine signals. Indeed, as we have found, 0 B' w/ I4 F- ?* Jmouse stroma prepared from developing human MCF7/Ras or , A3 Y7 |# E l z9 W d6 QMDA-MB-231 breast cancer xenografts is rich in cells with an ability % v) L9 D- O8 x: B- ]: mto generate fibroblastoid colony-forming units (CFU-F) in vitro + H$ Q$ Y- u4 ]( I( Y1 W, T0 l(Supplementary Fig. 1a), a hallmark of MSCs3. The absence of such0 l& i/ @( i3 |
colonies from control Matrigel plugs or from neighbouring tissues0 x2 ^- ]' n& {3 X d3 U
(negative control; Supplementary Fig. 1a) suggested that endogenous1 f0 ` }2 l" ] }- c( d& [
murine MSCs localize specifically to sites of neoplasia.6 z; w% U' R' e* Z w: Z6 _
To investigate whether human breast cancer cells also have the9 ]; @% @' `2 q7 C- D
ability to attract human MSCs, we established a transwell assay' Z u* Z T" x5 M# q3 u0 R
in which bone-marrow-derived human MSCs were allowed to o9 b) `) u/ u+ Vmigrate towards media derived from MCF7/Ras or MDA-MB-231) j& y3 m0 V6 f; Z5 M
cultures. We found that human MSCs migrated much more avidly 8 Y8 ?, _! J- B& d9 X(,11-fold more) towards media derived from these cancer cells - q9 t8 \1 j" ]% V' Othan towards control media (Supplementary Fig. 1b). More importantly,1 r( a$ N9 A, Y7 u
green fluorescent protein (GFP)-labelled human MSCs * D# D% H g' o; Kinfused into the venous circulation of mice bearing MCF7/Ras$ M" K( U \/ h, v! d. v0 }
or MDA-MB-231 human breast cancer xenografts localized specifically 0 n! p' j7 ?4 F* n) ?* T) Tto the developing tumours, with no observable accumulation ) h% Z6 T6 ~- }6 {* h: J3 L" pin other tissues, such as the kidneys (Supplementary Fig. 1c), liver9 L, S. {$ c% Y, R' d1 V% W% q
and spleen (data not shown). Such findings indicated that MSCs: B' F" \. I2 e4 v6 l; p& Q9 _
are specifically recruited by subcutaneous breast xenografts, and corroborated H8 q& Y0 R3 [& Y% O4 Lrecent studies that described the localization of systemically ' y) ^6 J- I m& yinfused MSCs to other types of malignancy, such as gliomas9,10, & V& r0 Y7 h! L/ B1 c, Q" Qcolon carcinomas11,12, ovarian carcinomas13, Kaposi’s sarcomas14 and 2 o$ I: Q/ |. g( k3 lmelanomas15.MSCs enhance breast cancer metastasis1 h' [& Y9 h& q/ T$ k$ l7 k+ H9 N6 J+ o
To investigate the functional consequences of the heterotypic interactions; z1 z0 e8 X7 [0 V
between MSCs and mammary carcinoma cells, we established ( P$ _( w! i4 Z: d, aa xenograft model in which GFP-labelled MCF7/Ras, MDA-MB-231,# {4 @, |* O1 N- x3 S2 i- P: N o$ ]
MDA-MB-435 and HMLER (see Methods) human breast cancer ]1 g: i' ]0 J. |9 y5 fcells (BCCs) were mixed with bone-marrow-derived human MSCs # n% n4 y+ m$ x; R/ j' e(hereafter referred to as MSCs) and injected subcutaneously into+ `2 e& K* Z- |
immunocompromised mice. The growth kinetics of the MSCcontaining ( s% I6 P# [! w' {" K/ d6 z: [1 D# Rtumours (BCCs plus MSCs) were compared to those of 4 X( s7 p" }7 ~: ]2 E# RBCCs injected alone (BCCs) over the subsequent 8–12 weeks, after + H# t$ _7 w0 L; Ywhich the histopathology of the resulting tumours was studied.1 Z' R: L" { r6 {8 i
We found that MSCs accelerated the growth of MCF7/Ras% F1 }0 T& G! P4 u. D- N
tumours without affecting the kinetics of MDA-MB-231-, MDAMB-8 D4 ~/ k! ^' \' s5 j7 a
435- or HMLER-containing tumours (Fig. 1a). More importantly, ( x, R- Z. F; M2 u; _# F6 z$ P8 Xwhereas mice carrying tumours composed only of BCCs2 D: b3 x; N* h/ w1 g7 @ c
exhibited few microscopic metastases in the lungs (Fig. 1b, d), mice 0 D# h0 F" ]6 v; N/ cbearing the mixed MCF7/Ras1MSC, MDA-MB-2311MSC, MDAMB- ! }8 o# w9 v) K0 Q; ~4351MSC and HMLER1MSC tumours displayed a marked 6 |; @! s' o4 D0 A. Aincrease in the numbers of micro- and macroscopic lung metastases & B j8 x0 Y6 H(Fig. 1b, d). Normalized counts of the metastatic nodules in the lungs% k3 s8 Y5 g3 P! C: Z4 a( |- o
of BCC1MSC-bearing mice compared to their BCC-control littermates % u1 g0 Y( F: g) D$ M+ ]& jrevealed two-, three-, four- and sevenfold enhancements in 8 _; h8 O _% V; Z+ Pthe overall numbers of detectable HMLER, MDA-MB-435, MCF7/8 D7 t) I! r) e
Ras and MDA-MB-231 metastatic deposits, respectively (Fig. 1c). 7 R3 M. h1 \+ CFurthermore, in contrast to the MDA-MB-231-bearing mice, the) m. \; f3 e4 p8 R, y6 a4 r) [: Q
MDA-MB-2311MSC-bearing mice showed metastases to various * ~. l& i; I4 K1 A) ~other tissues, including the mammary glands (Supplementary6 M$ O7 r- j! q0 E( j4 @
Table 1). Although all four of the tested cell lines exhibited enhanced ; U' Y7 n/ }: ?3 Jmetastatic potential after admixture of MSCs, we chose to focus, x3 k1 X- C! f* E0 X$ @
further analysis on the MDA-MB-231 tumour model, because it 8 ~% q3 W: |) f. J7 zdisplayed the greatest relative increase in MSC-induced metastasis- ], v$ g1 z- K( J1 F
without any concomitant effect on either tumour cell proliferation) F! J& `# l* ^7 ]
(as revealed by Ki67 staining; Supplementary Fig. 2) or overall primary # e- z( g) k6 L7 h4 h0 ytumour growth kinetics.- {1 Q$ p# S9 Y) U( b
We note that admixture of other types of mesenchymal cells— 4 w) X" `! x0 F+ f( Sspecifically WI-38 or BJ human fibroblasts (Supplementary Fig. 3$ f1 ?) t) h) x
and data not shown)—to MDA-MB-231 cancer cells before injection6 u4 ? W l! s7 w! u
into host mice did not result in either enhanced growth kinetics & L3 v( z( u6 z8 `) O! o(Supplementary Fig. 3a, b) or increased numbers of lung metastases1 _4 z+ B* Z6 \& }+ W" I, K5 z7 S( M
(Supplementary Fig. 3c, d). Taken together, these observations indicated + J* c- \) F5 y; `+ |that the metastasis-enhancing powers were a specific property$ a% |# _7 y+ |( y
of admixed MSCs or derivatives thereof.- k% k% o$ a! U" ^
Reversible metastasis$ y% S9 ]: Y8 c. v
Implantation of MSCs either contralaterally to MDA-MB-231 cells or ' R) W; i M! D4 sin nearby separate sites of injection did not affect the metastatic X9 y9 F5 a- `7 g/ ^! mpotential of the resulting primary tumours (data not shown), indicating3 T! a. I& A0 U7 u
that MSCs could enhance cancer metastasis only when they. n2 J1 s# s$ d; [3 P% J: x
were in close proximity to the engrafted BCCs. This influence might ( O9 |0 r w! m0 jbe ascribed to various effects that MSCs exert on the commingled- a0 P) b5 o: m2 V K
carcinoma cells. Thus, the MSCs might favour the outgrowth of rare 4 z; ~5 O- a. Wvariants within the MDA-MB-231 cell populations that exhibit 0 w) W. i% H6 Q! C: o2 I# Funusually high metastatic powers. Alternatively, the MSCs might ( F2 [ a! J) M6 x6 Q8 q. ?cause otherwise weakly metastatic MDA-MB-231 cells to acquire/ G/ p" O& @; e) Q9 l: ~
enhanced metastatic abilities. This latter mechanism suggests the* i" J3 [. }, I( n8 Q* ^1 R
possibility that the acquisition of the metastatic phenotype might, N* t4 U3 d$ Z" W
be reversible, in that carcinoma cells might revert to a lower metastatic2 g1 c: u2 \7 C
state once they were no longer in close contact with MSCs. 2 H/ f, a# T8 E) KTo resolve between these two mechanisms, explants of MDA-MB- ' e; X1 p$ [' I231 cells were prepared from BCC plus MSC primary tumours (Texplants)2 N& D7 i$ u$ `( F6 m) I+ n
as well as from their derived lung metastases (L-explants),/ A- `5 D5 } u% ]! H+ U' }
expanded in vitro, cleared from contaminating stromal components," e; U8 x) b8 J5 V; l
and then re-injected into subcutaneous sites in host mice in order 6 t6 g+ q- x6 G- W( g: Z! W4 Fto evaluate their respective metastatic powers (Fig. 2a). Although& ]) x- ?' h, V! |$ c8 C
the growth rate of the resulting L-explant primary tumours was % m; Y( g( t0 o& }' Xmarginally enhanced compared to their T-explant counterparts' _( N" r6 W d. K* [- V: D
(Fig. 2b, c), these L-explant cells were no more metastatic than the/ _/ ]/ J' u4 T% l {' N
parental T-explant cancer cells (Fig. 2d). This suggested that the) j; j# G$ E& [) H/ o
a 1 j4 t5 s0 |1 C2 Mc d7 M' R7 c0 a' ?) f8 z2 z
Days after injection 5 I( f2 n1 s5 A BTumour volume (mm3) 5 c8 y. T& j/ N: _. [10 17 24 31 38 46 72 80 89 13 19 26 33 40 47 54 61 68 75 81 10 19 24 31 38 46 3 b2 ~7 X- z1 H, |8 B" o! Bb " _2 X# V! i! P+ b$ ^6 S' ?! u0 K1 mm 1 mm" X3 X, j# K4 e z' d' N0 g
1 mm 1 mm W7 a1 a" e9 U300 μm ! M" k" H3 K7 ], _- FHMLER HMLER+MSC0 N# V+ P! n9 U" N. D, u- w
MDA-MB-231 MDA-MB-231+MSC * |* Y/ x% t) ^' u$ a. EMDA-MB-435 MDA-MB-435+MSC 7 {3 f' A$ i" ~! l4 uMCF7/Ras MCF7/Ras+MSC ! A- I, \- ]( l; v100 μm7 R! H4 f% L* ?- a2 N
300 μm0 b: x* z1 D+ A' W. y
100 μm. G, s* u0 w6 Z& N; U2 V, Y) E
MDA-MB-231 2 T, j: b5 L; e5 l+ ]3 RMDA-MB-435 ! t5 `4 ?8 i3 c0 j, k: W0 Q* g8 jMCF7/Ras. k6 N% b1 n' }
** ( T" m& r% _- l n7 @Metastasis index (fold) " W+ E0 N1 ~$ \1 @8 a, c0 8 i4 |' F: L( Z- a1 9 u( l9 }0 P8 r: F6 n6 d, O n1 K2 ) i9 n% \" [9 Z" l% F- z P31 u0 I5 z S7 D& n, \! g
4 ) v: Q' G9 B% l; ?( J! L5 8 o" }( K3 l2 ?$ e6 ! D- ]2 ^3 h3 a& z71 K! f/ P' ^# |' o" L
8( f/ O) Y' Z" Y
9 , S4 w# `. p/ ]* ]; ^*9 D+ |7 E- y. t) u) T- d9 Z$ m
**# w6 W) `7 L* Q1 u j
**( k2 C* R x) s8 b O
MSC – + – + – + – +) H) `: {! T! u A; j. t
* + o- K4 e/ V* ]7 l ^7 s( ]2 \& q& |- mHMLER " X# |+ Q3 @9 x0 h# K0 s' n6 ?700 1,200. C% X1 }. w; o
1,000 - e+ R# |; @7 ~2 b4 m% r" o$ j800 ; p4 n. s( S b# X Q600 * d) ?8 v& o# M! k: S2 w400! o8 a+ s, a4 O5 e& O
2003 O5 w/ R8 I+ o% K5 ?1 |" t6 z
0 1 Q$ p* q% N' D @3 K: Z! g800 - ~. C3 @9 s6 m; i) ]% Y700 + [/ z5 l$ u9 H& `600 ) ]( Q Q C6 v# O9 Z5008 r: \+ p5 w2 M, E
400 & L4 e9 k) J1 {9 |* n1 R300( w2 W2 y3 p' `0 \, s
200 6 n! V. g6 M% X5 e6 Z100) d: }& M4 S$ i7 @+ ?+ o4 c2 H# e
0 % a& G: d% V& r$ P2 E2 T3,5002 P) n5 y" ]4 j( `6 B
00 Z+ j X5 E# w
500 4 X, y; y' v [3 W9 J# U2 s. t1,0005 U* r" w3 H, o w! Y9 S0 Y
1,500 % a. ~1 v: F& K' K) x* P2,000 j) k: f7 W2 Q" _4 J% I2,500, c% O0 ]8 m+ @: E) y9 g6 T1 s; C
600 3,000 - Q& a9 h! v# A- {5 Q1 `500( R2 ?1 K' V7 f. a* b
400 & q1 c) k3 d' A2 Q5 f300 8 v/ r( Q9 T+ V$ t200" S6 |8 W2 U' K9 U5 ^
100 5 w: F: o0 e- u4 G' ~* J, O- u$ G0 J02 Q! m- R( o& z6 ?2 `2 x
0 21 31 38 49 56 63 70 78) {5 Z6 T& @- ?. f$ O) |
MCF7/Ras alone. h4 H& g0 {: M! P/ i! ^3 A
MCF7/Ras+MSC, |# S z5 b3 y- V/ I
MDA-MB-231 alone ; L+ q8 d4 d w% M1 j: zMDA-MB-231+MSC ( [- n# K ?8 MMDA-MB-435 alone , O9 r3 [7 Q e% @: J/ CMDA-MB-435+MSC7 T$ q o) y) H8 f7 _1 {1 |
HMLER alone) O: }+ I6 x) o6 @7 q$ D% O
HMLER+MSC ^' G. M, H9 A9 I
HMLER HMLER+MSC 8 l4 y" M# {3 }MCF7/Ras MCF7/Ras+MSC8 l. \' o% t R
MDA-MB-231 MDA-MB-231+MSC ) d. o( n/ j! p0 xMDA-MB-435 MDA-MB-435+MSC 2 U& W. u/ i1 e0 O pFigure 1 | MSCs promote breast cancer metastasis. a, Tumour volume 5 i8 U3 q) Y! T! \ ?- Wmeasurements (mean6s.e.m.) of 500,000 GFP-labelled BCCs injected. E% N: W w5 I% l% @& y
subcutaneously into nude mice with or without 1.53106 MSCs. 8 y8 Q3 @* ?1 B7 }, dRepresentative data from multiple experiments are shown. Diamonds, BCCs. ?6 G% U4 g8 e0 V+ s
alone, n55–7 mice per group; squares, BCCs plus MSCs, n55–8 mice per3 g2 G/ d, C3 e1 O' ]( `/ Z. f
group. b, Representative bright-field/fluorescence images of lungs of mice2 m" f( {4 I. }* z
bearing the indicated tumours. Cancer colonies are in green. MCF7/Rasbearing. M* p& I- C7 B
mice were killed at approximately day 150 to allow these tumours to; E {$ r9 l3 P& A$ y# o k
grow to comparable sizes to their MCF7/Ras1MSC counterparts. c, The# \7 W/ p# O0 r. h4 z7 s
lung metastasis indices pooled within each cohort of mice in a are expressed6 _2 x% b# v5 {. L
as fold increase (6s.e.m.) over controls. Data shown are representative of # H3 M c6 P# W2 kmultiple repeats. Asterisk, P,0.01, double asterisk, P,0.05 using onetailed3 P! C, h* Q2 }) {& I
Student’s t-test. d, Representative haematoxylin-and-eosin-stained 1 ^2 E6 m& C! \6 S, i) P( @sections of lungs of mice bearing the indicated tumours. Metastases are ~/ G6 {6 M% Gdelineated by a dashed line.MSC-induced metastatic powers reflected a reversibly induced trait/ l, F3 E' q* i9 R5 J2 h
of the MDA-MB-231 cells, and that the ability of these cells to metastasize; h; b) V/ k" E; d( K
to the lungs was a consequence of their ‘education’ by MSCs in 7 z4 W' D/ d* H9 _2 G- Z% K- Nthe primary tumour rather than the selection of rare variants of 1 m1 v# {& w1 N0 y9 g% N, c- xMDA-MB-231 cells that display elevated metastatic potency in a ! Q4 z9 H6 ]% t z. l" {stable fashion. & B2 L0 V7 q7 w4 s- E1 j5 T9 `The effects that the MSCs exerted on the BCCs might have7 [1 f9 _7 Z1 z
occurred within the site of primary tumour formation. Alternatively, 1 O% E: T/ `% e, @7 Kthe MSCs might have accompanied the metastasizing BCCs ) P' p D5 E6 Y+ M! B6 r4 }to sites of metastasis formation. To distinguish between these two$ @+ E: l9 Z! r
possibilities, we admixed ds-red-labelled MSCs to GFP-labelled" Z% m1 h0 R/ s6 X
MDA-MB-231 cells and implanted the mixture subcutaneously in N n! u5 x2 p$ a- z# f
host mice. We found that the tumour-derived lung metastases contained$ i$ [) x, D" P ?4 p
green-labelled MDA-MB-231 cells but no detectable redlabelled , }0 H7 m8 L' f6 LMSCs (or their derivatives; Supplementary Fig. 4a) when: L% U, k: z: C3 t, C! \$ I
scored 4, 5 or 6 weeks after primary tumour implantation. The $ F* i0 X# ?7 T+ ~+ Jabsence of red-labelled MSCs from the lung metastatic sites cannot( E- ~. e, A0 H# M7 n5 E! f/ Q
be ascribed to an inhospitable lung parenchyma, as MSCs that lodge$ w, U8 P0 R! D' \9 F' C; i" W& s
in the lungs of recipient animals after tail-vein infusion survive in that2 }6 l! E. P- u# |$ e. d. h5 l! o4 \
environment for ,6 weeks after injection (Supplementary Fig. 4b). Y) C: e* P4 Q1 AHence, it appeared that the admixed MSCs do not migrate in large: H$ k2 O1 Z3 K6 `) Y
numbers to the sites of metastasis, and that they exerted their prometastatic* F* f7 \% t- h8 w
effects on BCCs in the context of primary tumours. 4 F7 t" M1 q# ~2 I) Y) Q: l8 jCCL5 in MSC-induced metastasis 1 p6 c; w4 \+ v& t6 m2 GThe aforementioned observations indicate that MSCs supply locally 4 E* O; D# H0 t; E) [6 U: @acting paracrine cues that induce BCCs within primary tumours to , X( Q, w7 l5 L( hmetastasize. To understand this crosstalk better, in vitro co-cultures( O" V7 ]; a0 i
of MDA-MB-231 breast cancer cells and MSCs were established and 5 W; K5 c% z7 T0 Y$ C" }0 Ptheir conditioned media were screened for the levels of various cytokines, 8 v- z/ b6 x- i4 S- Z' ochemokines and growth factors using the Luminex-based Bio-, h C( c8 O: ?& m& p4 f
Plex suspension array system (Fig. 3a). In some cases, the resulting5 {' Y+ q+ j4 V9 B I. F
a # A ]) H3 F4 wb c d C1 S0 L; M. L6 B0 Y. PDays after injection " s% s7 m9 q: u$ O' W0 I+ A, S: FTumour volume8 }3 T* m* b' Z8 |% o4 u7 G
(× 100 mm3) % x2 f' |3 I/ T% M2 c9 |03 K4 d- y# Y: m
5 0 p: ~+ ]& w- U3 {, C$ x10 ' V7 M3 ]3 l1 n* s, V0 S15 ) w& `9 m g4 G3 \' i y20 6 T+ D7 x! r' i, W k i25 ! e# I+ [, M3 j4 o14 17 21 24 28 31 34 38 42 45 48 66 71 771 N( x' k3 K# {: j+ ~1 m8 @( x) F
Primary tumour explants ) X4 u, a4 g8 e" y6 {# D! \Lung explants7 x* g5 ^* U! p, G ?1 W; E% m
MDA-MB-231/ \0 t$ G" Z; [/ c' P b
MSC) j# e6 c T! \. W5 ^" Y
+ 2 a# N( s$ K. ?5 PLung explants ; @2 G' o, K+ [* J/ i$ P+ ]Primary tumour. n! H W9 B7 Y7 [7 A$ y7 o. G3 }
explants 3 i& r9 V1 z( v, v4 ?$ ]Antibiotic, ?! V7 p% ^, ^6 n
BCC selection( c7 s6 ^- {' {6 }0 m6 `& E
T-explant 4 M( a3 y+ `" X0 \" H7 i# rL-explant : u3 z4 D* {9 `3 e" _( dT-explant 0 O/ o/ n6 |! I+ P& g$ vL-explant( ^6 i2 @. Y" v& n
Tumour mass (g)8 q+ Z; Y, W' B& D/ T& k
3.0% T8 k+ x" a' _! Q* n5 l
2.0: U7 g `! ~1 ]2 U
1.0* ]4 S) S2 W7 G( A
0 09 T) `8 f1 t( N' l
0.4+ w' d) ]& A' d+ k
0.8 4 F3 H0 T0 G7 T, D7 t8 b1.21 j5 E4 Q0 x! C2 L8 I8 M" G
1.6 4 {* M' l5 c! [& K4 ~0 _" k2.0' Y. h$ r$ C4 O! I/ m8 h4 }0 l1 |
#/ c9 G! N+ f6 B4 H. ^0 Z9 D/ z9 R
Metastasis index (fold)+ G: d3 F1 [/ Y4 H* q9 t! Y
##, C* N: @! i& [ t( C
Figure 2 | MSC-induced increase in the metastasis of MDA-MB-231 cells" n. G. A3 @6 e5 {# q
involves reversible mechanisms. a, BCCs were recovered from lung or ) @6 [" x9 ?- d' V2 l6 \+ |. z3 Nprimary tumour tissues, cleared of stromal contaminants by culture in ' Z$ L; c" Q' N$ `blasticidin-containing media (5 mgml21), and re-injected as primary9 l( S5 a6 k( C* I
subcutaneous tumours in recipient animals. b, Tumour growth % _& C) M5 |) Y9 X) \(means6s.e.m.) of 500,000GFP-labelled lung-derived (L-explant) or primary ) Y& d9 {) k: A. \0 o" k, Rtumour-derived (T-explant) MDA-MB-231 cells inoculated subcutaneously. - G; B' [2 x; u3 f: u5 dData shown are representative ofmultiple independent experiments in which) I P3 {9 }: x3 o
four different paired batches of L-explant and T-explant cultures were assayed9 F T) v: N9 `* J" m
in parallel. MDA-MB-231-T-explant (n58 mice); MDA-MB-231-L-explant ! w) M1 O. Q+ S, p& ^(n510mice). c,Masses (means6s.e.m.) of tumours in b.Hash,P.0.4 using 9 L) w" N$ I. n8 k# S& `9 vone-tailed Student’s t-test and indicates no statistical significance. d, Lung& Y) {# S' j5 V- j: d
metastasis index of mice in c. Doublehash, P.0.3 using one-tailed Student’s4 W! M: i" u" i
t-test and indicates no statistical significance.) u! @5 L# `6 y/ t% n/ x* X8 ^
a/ ~, `+ K1 ]2 H( r N% B
CCL4 1 W; o* ]' c$ i* L9 ubFGF ; d% t1 R# M6 `2 l/ k$ r& E. uVEGF ( T! C( f% `5 C3 _ jIFN-γ 2 H, i, M3 Z u5 K5 `# @3 PTNF-α1 _* g p' ^/ _0 V* m4 k5 i @, F
G-CSF" h$ s# Z: } ^* u+ T- [
GM-CSF1 @8 e& I# x3 e
CCL3+ d! g% O& n" V/ d Y
CCL5+ A. b2 v) ~5 a8 l8 Q" `* `
MMP10 c, \7 y0 U& T
MMP3 3 [# C- S2 ~; K* a3 c. v0 kMMP9& P2 i+ s# }/ N. V: K
MMP13" q% a. ^9 H! w4 T$ F/ ?+ |6 x0 d
IL-1α# Q4 c6 Y6 U7 _, T
IL-1β3 |- v8 J1 b+ o. o
IL-4 8 H+ V" ?4 `. A5 y9 r+ |% I" PIL-5 ( y1 C$ A; F( T* K2 bIL-6 0 e# `7 ]& X! z8 J4 u* AIL-7+ V& X8 U, N# |( Q# d" Z8 c
IL-8 7 m5 N" i' i l$ b" d8 k) f; CIL-10 . T- V6 `9 j! s7 _" \! UIL-12" L9 z) T% B# z" H9 e- Q
IL-13 # O( ~( C0 D+ A) J! d' d4 O2 tIL-17/ t+ ~9 D/ y1 `1 \( O$ H$ ]' `1 T
IL-2 $ ~* `! R! @1 }8 PTGF-β$ A# o* F$ h2 f8 u' `0 _& [% v, {
Fold induction 6 ]- ~. f; Q- ?4 e0 @3 JMSC alone* f# l, Q' R& y) o# A) Q3 w
MDA alone+ u* q3 m7 ~8 ?9 ]3 a2 \. x
MDA+MSC (2:1 ratio) 0 z6 a3 w. k9 ^# P1 * * * * * * * * * & g% J! E0 U1 K0 `# j' i# h# I3& s) v) d0 k2 y1 K! F
5 3 l b; q# j# I3 F4 W% X3 M7* k# l2 v6 ?8 S: z* g
9 + d4 [: U& l" q, y& d; s119 O4 s- q7 B( U
13- y% F9 G) I, Z O; `+ F$ x1 T
60 0 `9 |8 q; D- I ab c $ Z$ g+ ~5 m- o/ x# r# }0 B0.4 μm / s' h- ~5 {! q( o2.0; a& E9 Z. o/ M0 n
1.8( I. ^& P, m W: f
1.65 I7 H$ J6 J- l- L$ c" ?
1.4 : e0 L* T6 ^1 j7 X, J# B1.21 N |0 I5 v* W* f# t
1.03 W+ K+ D+ H: x$ _/ [
0.81 X# S* c# t8 J) m9 I- E
0.6 : _4 S2 b7 _7 n' H" q8 Z0.4 ' p2 m& [: \/ [3 ?& y8 @0.2 4 R& p4 M) Y) g# J# q0 0. T3 w9 V4 h' J* V: ~5 U( r5 f
5& d8 t) o4 Y" K' H
102 V8 _$ z, ^+ x% n& U; w9 ?% M
15 2 r, b1 V( i9 Z* L+ V* E/ S8 s20& n# J U: v1 Z" [
253 V$ A. o# t+ ^3 _7 l, {/ n
304 n% K( ~& @2 `5 g8 c5 d
359 p( I# o* \* m, t/ c
40 + e* k8 E& a( D" ^0 k. l8 `45 ' t. y: m: S' N9 i2 w+ ~50 3 t1 f$ k+ B8 G: f& gd1 d2 d3 d4. Y$ n$ v/ z/ w2 M7 d, c
MDA alone; X3 }* h- e- B" W
MSC alone" ?* d6 T8 A6 F6 u: d7 D
MDA+MSC - C9 S2 T) ~$ LCCL5 levels (pg ml–1)4 i9 k# v/ ]. p6 c6 V' L: K
Co-culture- O8 L c- L3 }. g. d$ {8 b1 G `
Fold CCL5 induction) U3 P5 P2 t* e. ^7 ~( p# t/ @
MDA alone! p( B* g+ M7 J8 h6 k
MSC alone+ z- A; j. ^2 r% X# }5 F
MDA+MSC $ O9 |& ]6 ]3 ?& Vd * j1 ~% }2 I$ S" m I4 MCCL5 (A.U. × 100)7 Q# }. P9 ^0 r7 e" A; m. C- c
30# y7 l$ a$ c; a) f% K; [
25 , k' L6 |5 A+ s! @" y/ J20% K) G+ e$ o# M& g
15; R x5 Y# d: R5 {0 x8 e
10 1 ^ @; n/ y7 ], W5 3 L8 ?+ H9 N5 I0 A O" p9 ^! G0 2 P. T, l( i2 j3 P5 k# f' o" _+) j2 d5 ]* m: A' ^& e" @
MSC.c % K' f0 E) f8 }1 ^. ?) }9 |+ 3 U' B, G& O$ c+ FMSC.17 Z: K7 `/ |/ {: P
+0 M" y0 x. [: q2 w
MSC.5 # f, e0 y9 F0 x' b6 ^$ cMDA $ Y" I5 v- r/ \& YMSC% Z5 j+ Z$ T, F1 Z: H! ^' |6 v
MDA.1 & U$ V9 q9 \+ `( PMDA.c / z) ~# l) h2 O7 @MDA.57 q& u/ G4 v$ a+ ?
TC-MSC7 L5 V, M* T' w
MSC (from MDA tumour)# a$ k* k0 {" s: ?) w5 A$ N
BCC (from MDA tumour) N- V7 i3 R. |. {7 w5 z4 w% e
Control (MDA/CCL5)" W: E6 h7 W7 J8 u& o1 x3 F" @3 G3 W
CCL5 $ a d8 f1 M; k6 RGAPDH ' N) \* T0 Y, N* ae% ?; }. D- S( J1 W
Figure 3 | The interaction of BCCs with MSCs causes a rise in the levels of7 n5 c5 f* P+ l" z) B8 q( m0 x
CCL5. a, MDA-MB-231, MSCs, or MDA-MB-2311MSCs were cultured in # v, _3 f! K3 i5 t+ icompletemedia for 3 days. The levels of various factors in the cell-free culture/ v+ |$ P* ~8 q
supernatants were measured by xMAP Bio-Plex cytokine arrays at day 3, and 3 t+ E) y m9 H! Dwere normalized to the levels observed in the media of BCCs cultured alone. 2 A) O( j, H# H& K, r, L xData are expressed as fold induction6s.d. of triplicates. Asterisk indicates4 h4 m9 l' ^; Y4 Z; ~+ B5 _! n
undetectable levels. b, CCL5 ELISA on the media of MDA-MB-231,MSCs, or ; z/ h$ T5 R7 [2 nMDA-MB-2311MSCcultures (1:3MDA:MSCs) at the indicated time points. 7 Y( i0 ]0 n* M$ ^Data points representmeans6s.d. of quadruplicates. c,BCCs were separated * w" s7 U. ~9 {( q# I f9 X9 N" j% Kfrom co-cultured MSCs by a 0.4-mmmembrane. CCL5 levels were probed by8 H& {$ m l2 P# ]1 P! ~# m2 f
ELISA on the culture supernatants. Data are expressed as fold induction over ' a% [8 S5 g( h& Wlevels seen inMDA-MB-231 culture supernatants (mean6s.d. of triplicates).- A& }; Z8 L+ N1 {) y! C4 x
d, CCL5 ELISA on the supernatants of MSC-siluc (MSC.c), MSC-siCCL5.1) F: |( O6 j1 `/ ]/ y+ y1 e: G. v
(MSC.1) and MSC-siCCL5.5 (MSC.5) co-cultured with MDA-MB-231-siluc + X" k0 K" Z+ U( \1 v9 d(MDA.c), MDA-MB-231-siCCL5.1 (MDA.1), or MDA-MB-231-siCCL5.5 & H' i$ O2 Z. H(MDA.5). Data are expressed as means6s.d. of triplicates in arbitrary units9 ^1 m/ N' c# B3 a: F( w% f. p
(A.U.). e, RT–PCR analyses of CCL5 in MSCs and BCCs sorted from n- L6 t: W( {3 g6 o
GFP–MSC1MDA-MB-231 tumours (3:1 ratio) 4 weeks after tumour6 ?, B! W: O9 q; F; O' T4 U
implantation. Tissue-cultured MSCs (TC-MSC) and MDA-MB-231/CCL5, u7 z& F/ q+ k! m
cells were used as controls. GAPDH was used for equal loading.levels of certain released factors (for example, interferon-c or - o# `; a' @# R0 W' x' J0 _- gtumour-necrosis factor-a) reflected the additive contributions of ! z3 C6 d7 s8 [9 kthe two cell types when cultured on their own. Notably, the levels* h6 b6 K4 g& ]- ~% x
of only one cytokine, CCL5, reflected a synergistic interaction2 Y' m& V& Q0 [: ^
between the MSCs and BCCs, as it accumulated to levels ,60-fold $ q+ h: G+ I) j* K# Vhigher than those produced by pure BCC cultures (Fig. 3a). This' ~. i+ C7 `8 F2 v
cooperative induction of CCL5 was proportional to the numbers of) l* \* H3 d) S4 T8 f5 D8 F) Z
MSCs mixed with the BCCs (Supplementary Fig. 5a), and was apparent7 e* Y0 u6 F, W
as early as the third day of co-culture (Fig. 3b). Moreover, this @3 l: l! e6 C6 y! Q1 ginduction required close physical contact between MSCs and cancer4 [8 @/ @% e3 F. P; C9 j8 W
cells, because it failed to occur when the two cell populations were 5 ]# b# t; f: O+ Jseparated by a permeable membrane (Fig. 3c). ; S& W# n' @5 n- Z& v: Q' h. M: OWe undertook to determine the source of the CCL5 produced @4 E3 k. e+ s/ B" N8 zunder conditions of co-culture. To do so, we stably reduced the 8 t: Y8 q }/ ?- T# l# M4 Qexpression of CCL5 in MDA-MB-231 cells by.80% using short " X" J6 X0 ~# w/ i. hhairpin (sh)RNA (variant siCCL5.1; Supplementary Fig. 6). Importantly, 8 L5 O$ c" w" l5 K: V. |however, subsequent co-culture of these MDA-MB-231.1 cells* S7 y5 C# j0 E$ D/ j) ]4 ?
with MSCs continued to allow accumulation of CCL5 in the culture 9 L0 I! }1 C8 \ K7 d7 j1 D Qsupernatants to levels that were comparable to those observed in the / D9 f; Q9 q7 E0 pco-cultures of MSCs and control cancer cells (Fig. 3d). This suggested ( U6 J5 r% [ j% I5 ~ y1 z) tthat the source of CCL5 was the admixed neighbouring MSCs. & j' Z) n6 r) o/ s! A L, ~Indeed, inhibition of CCL5 protein expression in MSCs using the 0 P9 [1 M8 t2 U3 a& Z: xsame shRNA hairpin vector (MSC.1; Fig. 3d) resulted in more than 9 [! l$ f+ I$ \/ ?- A4 K. j75% reduction of CCL5 protein levels in the co-cultures, indicating2 n& L5 r4 Q# k7 A' X j0 y
that the MSCs were the major source of the CCL5 observed on coculture & C5 V/ i; n7 t# Nof the two cell types. In support of this conclusion, analysis of 9 Z- [/ X! K6 |9 FCCL5 levels in the media of MSCs or MDA-MB-231 cells separated - o4 \2 n3 V/ @) |. Ifrom one another after 3 days of co-culture indicated a strong induction# r6 V, c( A: O6 c0 z
of CCL5 in the culture of MSCs, but not that of BCCs (Supplementary( p, E0 v+ z6 _ c* g
Fig. 5b). Finally, polymerase chain reaction with reverse * V, v! D- \& L6 c0 Q3 Dtranscription (RT–PCR) analysis of the RNA prepared from these coculture-) ]( y7 _. s; w7 C7 E
derived MSCs (Supplementary Fig. 5c), as well as from the 4 f+ {4 R, v: u9 ~6 \3 aMSCs isolated from MDA-MB-2311MSC tumours ,4 weeks after ) y4 I" \; I" ttumour implantation (Fig. 3e), indicated a strong accumulation of5 ?) t! l5 ]/ v" Y+ B$ k
CCL5 messenger RNA, suggesting that an active signal transduction " j: y, {: r; tpathway is triggered in MSCs by the nearby BCCs.. @2 W1 Y; w8 C& j
A series of observations has linked CCL5 signalling and cancer. For ! T$ t7 b6 K5 n0 jexample, CCL5 levels in the plasma of breast cancer patients have; L" ^) D1 R- A6 Z9 z* E
been correlated with the severity of the disease, and localized CCL5 ) L M$ A& L& V# i8 T* Zprotein expression was found to be elevated in invasive tumours' k6 x- L6 r- z) E- i4 }
when compared to in situ ductal tumours or benign lesions16,17. 5 a2 n: ?; s) |; SHowever, the precise contributions of CCL5 to cancer development 0 a/ D& u# _, f+ iand progression are poorly understood. To investigate further the5 O" }, K5 P8 j" v
possible causal role of CCL5 in cancer cell metastasis, we overexpressed : p5 E# q2 ~: e) g) y+ ^% ethis chemokine in the MDA-MB-231 BCCs (Supplementary 0 ?3 V6 j& o3 n" m! d: M UFig. 7a) and analysed its effects on cancer cell growth and ! x! }/ q6 a7 Y2 T# Ntumorigenesis. The overexpressed CCL5 did not confer any proliferative: j/ \9 P2 e- R- D0 I
advantage on cultured cancer cells when compared with8 E) z* F6 y2 M$ r) X: C; ?
those lacking such overexpression (Supplementary Fig. 7b), and had ; y/ f8 f2 v4 P4 Q& g$ Y4 h, Nno effect on the ability of BCCs either to grow in an anchorageindependent; o3 F9 w$ N: r* n9 x# K" F
fashion in vitro (Supplementary Fig. 7c), or to form$ D0 ^9 X" K! q: f5 n5 }- ?9 u- }# F9 v
primary subcutaneous tumours in immunocompromised mice (Fig. & I H" E7 x8 B/ r E; p5 t4a). However, these tumours exhibited a ,5-fold enhancement in ) {+ z9 p: O2 c; r3 ~: Jtheir metastatic potential when compared with control tumours . P2 G2 S1 L5 r: ~' R' u6 z( flacking ectopic CCL5 (Fig. 4a). Similarly, overexpression of CCL5$ Y2 k7 X2 o7 {# u# {: V& R+ [
in WI-38 fibroblasts sufficed to enable these cells to promote the - B' c0 Q3 Y* y' ]metastasis of admixed MDA-MB-231 BCCs (Fig. 4b), indicating that / ?* B" r* F# {! s- dthe actions ofCCL5 are responsible formuch, if not all, of the observed $ ^/ @* ^9 n" W: o R N4 z$ kMSC-induced metastasis by the BCCs. 9 p. S4 J$ j, \6 n' u UCCL5 promotes lung colonization! T/ G1 T m: K r% }' j
Previous reports have described an important role for CCL5 as a ' _5 m, \+ A' u3 U1 ?chemoattractant for stromal cells, such as macrophages, that express / z- k p, V0 s& p. Kone of the receptors for CCL5, CCR5 (refs 18, 19). Furthermore, $ U. n* ]8 z, S: RCCL5 expression has been associated with increased tumour neovascularization,- k, f! ~: H" `5 y
suggesting that endothelial cells, which express a variety 8 ]8 M0 H7 |0 H+ S4 Gof chemokine receptors, may also be attracted by CCL5 to sites of 0 k+ K: E( @) a# dtumour formation, thereby enhancing tumour angiogenesis20. Such : x/ s" q5 @2 Q9 ]% Q vobservations suggest that CCL5 may contribute to breast cancer1 r5 H; J1 v9 E6 J6 K
metastasis through the recruitment of a number of stromal cell types 2 z! [; R* j5 e$ @: Q3 yto sites of primary tumour growth. 3 ` `0 z. V6 [2 m4 aHowever, immunohistochemical analyses indicated that the s, v- J7 j4 j; v' `5 X8 C
MDA-MB-231 control and CCL5-overexpressing MDA-MB-231! _5 p5 f4 M$ v0 j( _& `
(MDA-MB-231/CCL5) tumours exhibited comparable numbers of ) _+ S5 p1 a( V1 ?& mtumour-infiltrating macrophages and had similar vessel densities (as3 v7 H6 D4 p9 n, ?
evident by F4/80 and MECA-32 staining for macrophages and * \8 i+ X* T" cendothelial cells, respectively; Supplementary Fig. 8). In addition,( z, w- V T- t( Q
we found that ectopicCCL5 expression did not cause an accumulation / T' c" `5 c8 F1 J8 z% Nof other stromal cells, such as SMA-positive cells, in the examined & U' g4 n' M5 @3 btumours (Supplementary Fig. 8a). Together, these data indicated that 6 y! Z- R3 H: h( Pthe observed CCL5-induced metastasis could not be attributed to 7 F) X2 k! G2 z! Osignificant effects on the numbers of the major constituents of the5 l+ _7 h" \- o& u" S. h4 K
stroma or to the vascularity of these tumour xenografts. # I6 m6 q7 D) F( r* V6 iInvasion and metastatic dissemination of carcinoma cells are often' K- a5 ~% i/ h, m
facilitated by their transdifferentiation through the process termed0 e) E2 A5 `, N! P1 Y
d f3 r/ q/ R4 H! L7 c
Bcl-XL0 I9 U$ ~, |- W, w
Bcl-28 _4 ~6 E1 J& ~3 {! w7 V
β-Actin 4 {4 |5 M4 a0 `2 p( QS473-Akt + m8 |- V# _2 s7 A( \4 z0 gMotility ! E5 g. w3 j! n+ G5 {; v. l0.5%→10% " y) x3 M6 ?4 G" XLY – – + +) k/ c0 }* u) R6 c: k
g& \, P+ z A/ c/ w# f) {( j
Extravasated clusters (mean)3 v& _' g4 K! |# U! Y& K
Migration (fold)$ L8 v1 G+ Z8 r0 @* x* D
*$ ?5 Q; L5 W) `( e8 Y9 ~1 q
e8 U' G+ t8 T0 M2 @% |
0( [/ S- @2 r6 G7 Z8 m3 E8 P! c
MDA/vector ! y* H, [5 k1 N) R9 jMDA/CCL5& M6 F% M# c t4 s# W" t
*** 4 y* `4 \: ]1 _# O5 xInvasion % k, z1 D0 |) i0 ^10%→10% # J/ v1 E7 | V* gInvasion c, |1 F$ |, f3 }1 k( v
0.5%→10%, i( Q5 z) ]( \" |% i. d
Migration (fold)% x$ s$ A4 R3 D7 ]! V( {
** 7 Z# q$ e1 ?0 A. E7 p0. ~% s1 c- }1 L( B* o- T6 A
0.5 2 E o4 q w p1 H6 R1.0( Z8 n: C3 F' u+ H3 m
1.5$ h H! k% @2 N' A$ k
2.06 v+ f+ V j2 k: S
2.5 1 z0 B/ a" J4 j3.0 0 u/ s+ W% n/ a# q& x9 l; A6 c3.5( k$ f1 \; d% s- U1 V2 Z9 T2 f
*# d2 L$ i3 P4 z$ u# {
MDA/vector# S+ s* y9 [0 ~, @4 U) A
MDA/CCL5 `2 t1 o' h; `6 S3 S( N/ o4 S" A: ^
a b# y1 G1 z" }; P$ ?2 o2 @
* ( e- R) n; k3 O( V3 n* ]( ~- R- M* rNodules per lung (mean) ; O g" j6 D' v1 k6 k# p& B) lc2 O3 d1 J3 v% w/ i
30 , d. T6 j7 u+ { ]( A25 8 c2 g; O+ n% b+ c0 G0 @2 U* A" u: A20 - W1 Z9 h0 J# ]1 w4 k7 G6 ]2 F15& D9 M* r& G8 k2 ^
109 {7 r1 ]9 P2 v5 s$ i6 G8 m
5, N. Q; @3 u( C$ V
07 D: o7 b/ ^9 ~4 [" y( Y6 f
Tumour mass (mg)2 E0 U+ I, R8 |, O
304 u3 v' Q9 _) I) K, _) ~4 f
25 & J w% o- U2 G+ J20$ P9 G# Q& h/ o7 x6 I" O. h2 V
15 9 u2 e7 z8 v/ A# y$ Z3 [10 & A. v' r3 ]' n5# f' v0 n2 f+ s+ u$ t. `
0 0/ w( s: c3 U1 ]2 p# J
50 , y g( o" i6 \+ K) o100 8 Z$ X/ Y+ h6 ~3 K150+ U# Y! Q0 d4 D1 ?: y
200 : t9 _% `) v3 `9 ?! |250 ' r. O3 W; {9 K1 {. _. @02 x2 N3 z3 _: x+ g& X3 [
509 ?, ?3 ~ f Q- z% H
1007 J, U8 q9 C) z* _4 ?2 k, h
150; \7 ]3 M# z. O" k
200 0 p5 B1 n( W% @% s250# w8 P& w4 o( |/ n7 t; v
300 1 p: @# O* P7 d# x3505 D5 i9 c+ p; z9 n( } O6 v
Tumour mass (mg) X$ J& l: S) ~Ctrl 6 ]" C9 ]& P2 q3 t5 v0 ACCL50 k& ?0 _1 s' L# D3 Z* }
Ctrl5 q' c, R. X+ E% U; S6 m j
CCL5 2 \$ B* F; i6 h0 X) \Ctrl ! S0 ?0 B* }7 d9 [$ ~CCL5 ; M! K. u0 N! r$ B& t/ iMetastasis index (fold)( b6 E. O! P3 G c) J2 F9 V! E( g+ y
Metastasis index (fold) 3 r! @. `! z! W0 f0$ N$ N$ I& W, }3 r2 \
1. p1 z; `& P. x/ U2 a
2 4 c. r$ E3 I. ^% Q0 _32 G+ Z: E: h) _9 i
4 7 {2 B. y- e. z) `5 q8 N5 0 {' I) M8 k k+ v# Z* 6 , x# N+ @8 Y# ^6 I% ?0 h* |0 5 l' t5 P# i8 g' W3 d. ]1 ! R3 {# y: ^- y$ k. t20 T- H. U( i F9 J! ]
37 p8 q8 ]1 J R- h
4 $ I: ~% F! w% a% v, E) S) s5 |5 . e4 X5 V8 m. c4 y6 8 L6 @$ Q1 g1 ]1 k7/ o7 M+ |+ d0 D. M6 }
* 82 I# B, s# c! P. W; Z5 B
MDA/vector : X2 x/ o; t" jMDA/CCL5 ! m- j. S. i* ~, S3 Q9 ]2 GMDA+WI-38/vector 5 Z, }( j M9 SMDA+WI-38/CCL5- T a6 Z, I% [2 W+ z
1 ! c- Z0 X% f3 E( u$ r2 . s* Y; F' J8 Q c2 H7 P+ L3 0 a; j* F6 u$ O2 P" `& U% l- F45 H" q' U& f3 } E* Y3 L1 f
* + M( ?0 P5 C0 ~# A" O5 jFigure 4 | CCL5 enhances breast cancer cell migration, invasion and, q0 [) _% B/ ~0 z$ l
metastasis. a, A total of 500,000 MDA-MB-231/vector (ctrl) or MDA-MB- R$ d, u" B7 I- e
231/CCL5 cells were injected subcutaneously in NOD/SCID mice. Tumour) z( S% {' C- c# Y5 L
masses (mean6s.e.m., n56 each group) were taken at 10 weeks. Lung9 ]' F M2 v, L8 i0 Z9 k
metastasis indices are expressed as fold increase (6s.e.m.) over controls. Data : L1 R# C2 S# \9 `shown are representative of multiple repeats. Asterisk, P,0.01 in one-tailed 4 x" ]# f0 B: d* Q: M! sStudent’s t-test. b, A total of 500,000 MDA-MB-231 cells were admixed to $ r/ J+ M" }! X5 M9 Q$ N250,000WI-38 fibroblast controls (WI-38/vector) or WI-38 fibroblasts ; l9 ~+ P5 C9 W& V( y8 O: Roverexpressing CCL5 (WI-38/CCL5) and were injected subcutaneously in 0 ~) ? x' a$ `+ pNOD/SCID mice. Tumours (n55 per group) were excised and weighed at' i+ p# j4 @( v1 ]" S
12weeks. Masses shown represent mean6s.e.m. Lung metastasis indices are . p# a2 X8 D! O5 N4 `: Mexpressed as fold increase (6s.e.m.) over controls. Asterisk, P,0.01 in onetailed 8 ?1 s, \/ K9 {4 EStudent’s t-test. c, A total of 800,000 indicated BCCs were introduced ! |2 B7 b' e1 o8 C3 kinto the circulation of NOD/SCID hosts. GFP-positive cancer colonies in the4 g% M% ^' l' d) K
lungs were counted 6.5weeks later. Bars representmeans6s.e.m. (MDA-MB- 0 A- p4 ]- b: o$ t6 ~( Z" N8 |231 controls, n516 mice; MDA-MB-231/CCL5, n518 mice). Asterisk, ' ^- c7 W: H' I& Q! cP,0.01 in one-tailed Student’s t-test. d, Western blot analysis of lysates of6 R4 C T+ l2 L# D' Z3 k3 a
MDA-MB-231 control or MDA-MB-231/CCL5 cells. b-Actin was used as a/ ?, P& g+ _8 B+ n
loading control. e, Transwell migration orMatrigel invasion assays on 50,0007 P9 u! [" v3 I2 N
MDA-MB-231 control orMDA-MB-231/CCL5 cells.Data are representative of 5 B0 m% y. o8 bmultiple independent experiments and are expressed asmeans6s.d. Asterisk, R$ H& m% J9 y! ?# h* _P,0.05; double asterisk, P,0.05; triple asterisk, P,0.01 in one-tailed 1 U _6 p& {4 hStudent’s t-test. f, One million GFP-labelled BCCs were injected into the tail6 B; s/ i- R% k0 Y3 ?3 H" K' S5 S
vein of NOD/SCID mice. Lungs were processed 48 h later and examined for + X$ f' v: H0 K k" v3 S3 l0 B( Yextravasated cells. Bars represent means6s.e.m. (MDA-MB-231 cells, n57+ [# }. x& h) W
mice; MDA-MB-231/CCL5, n510 mice). Asterisk, P,0.01 in one-tailed 0 N; e7 \9 I4 K$ }" MStudent’s t-test. g, Transwellmigrationassays on50,000MDA-MB-231 control Z5 K& t# q0 {1 z. X- g3 Y
orMDA-MB-231/CCL5 cells plated with or without the phosphatidylinositol-+ i5 Q# ]7 w0 ]) Z$ y8 e
3-OH kinase inhibitor LY290042 (0.5 mM); representative experiment shown; / n r C/ ^1 i! L6 I$ D$ Oasterisk, P,0.01 in one-tailed Student’s t-test.the epithelial-to-mesenchymal transition (EMT), in which cells shed 6 } D6 g# ?3 T5 btheir epithelial characteristics and acquire instead a series of mesenchymal * c5 b6 c- q0 @. E. {% bmarkers that enable their invasiveness and intravasation21.0 H, N8 a* W9 Q) T
Despite their lack of E-cadherin and their expression of detectable levels 7 a5 |. q4 t, c+ V, u9 S4 q Kof mesenchymal markers such as fibronectin (data not shown), the9 [, O/ n8 i2 _6 B
MDA-MB-231 cells studied here exist in an intermediary phenotypic 9 P; Y/ k- F- K9 L, C' ~state of ‘partial EMT’, as they retain a distinctive epithelialmorphology ; ^0 g [; ?2 h% Fin vitro and are still responsive to EMT-inducing stimuli in culture. In ; t T5 I7 N- Q0 ?fact, we observed that ectopic CCL5 expression did not cause MDAMB-; O, |) S1 l& S. N
231 cells to undergo themorphological changes usually associated6 B ]0 |* b6 a. V( |7 [
with an EMT(Supplementary Fig. 9a), did not cause rearrangement of# L2 y9 j% d: T! L, p! h
their actin cytoskeleton (Supplementary Fig. 9b), and had no impact on 5 O& R) j2 u. W' v3 G( kthe expression of mesenchymal markers closely associated with the ) ~- a( @9 G8 w" s! |) g( MEMT process, namely vimentin, N-cadherin (Supplementary Fig. 9c)5 Y3 {0 I9 G( F# A7 F: j# p' {
and fibronectin (data not shown). These data suggested thatCCL5 does" O. |2 F6 U8 Q' J5 L. O
not directly promote the EMT programme of MDA-MB-231 cells.$ g. }+ R. V4 I3 I, j4 U) J/ y: b
We proceeded to explore an alternative possibility: that CCL5* Q2 Y0 d, s0 X- J
expression affected some of the later, critical steps of the invasion– ; V. [4 _5 [: {% Gmetastasis cascade, namely the lodging of cancer cells in secondary" |, N1 x# d1 T: k! y6 d2 c
organs and the subsequent step of colonization. For that purpose, * n; A& Z7 y" e2 g( ~2 {- p! o9 rMDA-MB-231/CCL5 cells were injected intravenously into host# J& s- z, I3 }. t& [0 U0 k# K
mice, and the lungs of these hosts were examined ,6 weeks later" f& _1 j$ U3 ?5 Q
using fluorescence microscopy. These experiments revealed that. v+ ~( R( F& O" I
CCL5-overexpressing cells indeed had a significant ,1.8-fold 0 B5 K% K: q# q9 E h1 c( Vadvantage over their control counterparts in colonizing the lungs & c: f, {0 i0 }' V6 Q7 \1 w(Fig. 4c), suggesting that CCL5 exposure has effects on later steps " N( }2 O2 S7 cof the invasion–metastasis cascade. We note once again that this% X4 U& ^. ^ @3 {
enhanced tissue-colonizing ability was not due to CCL5’s effects on ! ^5 j$ s: r$ C5 E- E1 f2 hcellular proliferation measured either in vitro (Supplementary Fig.5 g/ f) J, p5 H6 n9 x3 T+ d
7b) or in vivo (Supplementary Fig. 7g, Ki67 staining). . @; G0 D& ~) S9 u. T! JBecause improved colonization can be due to enhanced cellular I0 P3 A/ d5 Z( ^; A
survival, we tested whether CCL5 protects against apoptosis.1 _5 X0 x, n# j4 h* o3 k
Notably, we found that MDA-MB-231/CCL5 cells exhibited higher& @+ h+ v% [( l
levels of the Ser 473-phosphorylated, activated form of Akt, but5 F# j/ x" Q+ N3 ^4 y, m: x4 t- I/ A
exhibited no difference in the levels of other pro-survival proteins,' J4 Z" s/ @4 _8 T! g3 j3 {
such as Bcl-XL or Bcl-2 (Fig. 4d), or a reduction in the levels of # N. K, d1 i% Vpro-apoptotic molecules such as BAX or BAD (data not shown). 2 _3 X- G( _6 U( aMoreover, we found that overexpression of CCL5 had no effect on, }% r7 Q! e5 L8 c h& H
the ability of MDA-MB-231 cells to withstand serum deprivation 3 S) _/ k$ N7 Z. g(Supplementary Fig. 7b), loss of substrate anchorage (Supplementary % I0 z5 A5 e" f6 F# C0 uFig. 7d), or hyperoxia (data not shown). We also observed that 5 x' \4 H( a& i$ v- Nectopic CCL5 expression did not protect MDA-MB-231 cells from 1 H- C( g: f) o% `* Y7 Pdoxorubicin-induced apoptosis monitored using western blots for 8 c8 B6 D0 i) C/ ycleaved caspase-3 (CC3) and cleaved PARP (as markers of apoptosis; - R7 u8 S" P% e* p3 `6 eSupplementary Fig. 7e), or TdT-mediated dUTP nick end labelling 3 f8 M: E( U4 }) ?$ F(TUNEL) assays (Supplementary Fig. 7f). Finally, immunohistochemical 9 _$ |4 G4 e! aanalyses on control and CCL5-overexpressing tumours ~" q' j" i& c' f" r8 n
revealed only minor differences in the levels of apoptotic CC3-* t9 B o3 x# `4 K( E" V1 d( `
positive cancer cells among the examined groups (Supplementary5 \* l' D3 X, N) r8 X" Z. U. y
Fig. 7g, h). Together, these observations suggested that CCL5 does7 c4 K% h6 K- ^" r; p: v
not exert any detectable pro-survival functions in vitro or in vivo, and; _) k- b# H: L2 `* _8 ^8 e
that the observed enhancement of lung colonization was not a consequence + B7 L% `; F* u8 o# [of significant anti-apoptotic activities of CCL5.* e9 y+ m2 V | ?
Akt serves as a key relay switch for upstream signals that promote " |7 K' t( F' }4 hboth cell survival as well as cellular motility22. Because CCL5-induced, U/ m' {4 }( z5 y$ q6 _# R" c; r
Akt phosphorylation did not correlate with enhanced protection % [( X, v4 A' w0 c7 uagainst apoptosis, we tested whether the CCL5-enhanced lung colonization " \! L, O9 q& ~could be due to an increased ability of MDA-MB-231/ & J( b4 I _0 q) n# q5 V+ s+ _: kCCL5 cells to invade from the microvasculature into the lung * H! y9 X% ]) k/ c, k5 ~8 `8 v4 iparenchyma through the process of extravasation. Indeed, ectopic& l* T9 Z p! N1 I- K# K: [
expression of CCL5 enhanced the motility of MDA-MB-231 cells 6 @& t7 \* f0 V' Ythrough permeable Boyden chamber membranes by,1.5-fold as well% v6 M% B9 m: V. t1 ~
as the invasion of these cells through Matrigel layers by,1.6 or,2.5-9 Y7 K" C" Z9 k% x5 U1 o1 V
fold in either high or low serum conditions, respectively (Fig. 4e).5 e5 m5 c# k' _7 H3 P, |* }
Notably, when we flushed the lungs of mice 48 h after BCC tail-vein : Q, l, A/ [/ x) binjection—in order to remove most cells that remained within the. b, f E) D2 v0 w5 c$ u- ~# P" G
microvasculature of the lungs and thus had not extravasated—we9 ]9 J" p C$ B0 C! x4 \
found twice as many deposits in the MDA-MB-231/CCL5-injected" k: H/ K1 S6 F( {& g: M( L
group than their control-injected littermates (Fig. 4f). This indicates a5 Z$ G6 h8 A/ ^& B% t w
clear effect of CCL5 on cancer cell extravasation. 0 R: [/ c0 q* F4 o8 Z8 z6 YFinally, we investigated the role of Akt in mediating the actions of0 F0 T- d8 e5 M7 _
CCL5 on cellular motility by using the phosphatidylinositol-3-OH . m: |8 t9 f" _' p1 }# { Lb silacZ( p2 h" c0 \0 @4 u! O' O
silacZ" y; W" y% b: Z! s/ D
si8091 Z* z5 C2 z7 j% h# I. v6 y! ~
si8092 I% ?' h0 O& }( `8 I! V- B
si186 0 K- ^1 m0 p0 c! n7 Tsi186 $ k, f3 U9 J( |2 ~CCR57 a4 f6 L$ {3 e4 G+ E9 U
β-Actin 4 C* e! E( j) j' b0 v) ]a d4 E& Q6 \% U% |, n
M8 ?& Q7 N4 e% |
+0 D8 c1 n6 k) F# Y: k
MDA+MSC * X4 g* u' O- B) A s+ IgG 8 ?$ J y* d. b& h0 ]+Anti-CCL5 Ab 4 K. f3 [* O7 h, ]: H6 BMDA- O1 T, B. N$ r
+Anti-CCL5 Ab 2 w( [3 l: A% S4 O. Y+ @! RMDA+MSC: X8 R5 R' E2 T& G( @
MDA$ z0 t) g% ?. l/ V e" N
IgG0 n4 B( D8 {3 U7 t. N: e
< 2 M* g6 X6 H, u$ n/ l< ; o- P9 A* k. y<5 a5 o5 N& T8 v% n/ x, z+ F, |0 P
</ t) ^ V2 |7 Z$ `4 K3 I
<- A+ Q2 k' a. I
DAPI . B s( c# d& R$ i1 FCCR5 # }$ _# U' {* g; hMSCs MDA MDA+MSC, Y; t3 H* E6 o4 l
DAPI DAPI0 ]/ L9 f, J! K5 [" I
CCR5 CCR51 s0 N! y& ?: t' _+ M
MSCs MDA MDA+MSC 2 ]2 J* a; }0 y# ?- ]4 ~; ?1 w- b*. z6 @: ^3 g3 ^% d1 f* {
c ; D* s7 f4 ]" g! \! D6 i0 rMetastasis index (fold) ! L6 X* R* I" E: eMetastasis index (fold) # Q( l( r* `& t) R*) _. s& g3 M: ~ k/ }- t
Control siCCR5 5 C5 l# |" E- q/ x- X1 A$ VMSC – + + +8 k% G7 U8 R/ W2 G/ T
0! T1 ~7 w, y7 ~5 V/ _& n( K. J. |
19 u( @ b1 B1 V! `
2$ R2 F0 Y7 f/ N8 o' X
3 9 c6 C6 k6 Q" w, p2 `43 e3 u2 s0 Z- L4 a* M6 S' N& J* m
5 e% y' ~& h! }/ N D( b$ d* |
+ + – – 9 K% f! y, L9 b6 M. [Anti-CCL50 ^ r2 D. F6 w) O; C6 Y
0- J: j8 ~4 W; [7 W- u
1* q7 a1 g3 P! y& O' E. T
2 ' H( }8 l: t0 R/ p* ?3 0 u' D, u; t D( C) K+ _( a4 0 a% U- W9 Y7 n5 + z/ l" W0 t, V" Q6* [% M, o0 C. t7 k( p7 @
– – + + . s9 L0 i L S% P* p9 OIgG 4 d3 o* {! V/ ]/ ~$ ]( H7 PMDA / ~2 Z7 C! V4 o5 ~) Y% B: K- iMDA+MSC) l B, O1 u$ A; B8 q
<9 k3 r- e$ M9 I4 A" V/ l0 [
Figure 5 | CCL5–CCR5 interaction is essential& Z( ~, C0 e8 I2 w
for the MSC-induced metastasis.7 W6 h1 N" g& O) E
a, Immunofluorescence analysis of CCR5: x( f# o! _& ]" v }5 A8 c4 @
distribution in MDA-MB-231 cells cultured with/ v, \3 _- Y3 L( U
MSCs. DAPI (for nuclei staining) is in blue; ! d/ q- B$ D& DCCR5 detected in green. Arrowheads denote8 D. B4 M1 D6 i3 c2 M- U9 V2 \
MSCs. b, Western blot analysis showing CCR5 - g. }( i- T' v* \. u- a4 Iexpression in MDA-MB-231/silacZ, MDA-MB- 2 H" A, H9 Z5 K# f1 R4 _8 U231/siCCR5(809) and MDA-MB-231/* l; F2 @0 q( M4 ?8 j
siCCR5(186) lysates. b-Actin was used as a. B' S1 `: m7 a
loading control. c, A total of 500,000 cells of the % { _& ^& X7 G, }1 I7 j' v1 A7 jMDA-MB-231 variants in b were co-mixed with& B3 I% F2 r& ?3 z9 _8 C! O: L3 q! ]0 [
1.53106 MSCs and injected subcutaneously into( M& `7 H; I/ Y5 M; `- s+ M
nude mice. Mice were killed when tumours , f- b4 X1 w* W( n5 R7 |) mreached 1 cm in diameter and the metastasis" h& p7 H9 c, F7 T3 Z
index was calculated for each cohort (n55 per1 H0 |2 q8 z/ ^0 h; S! V2 r: Z
group). Results represent means6s.e.m.;8 T* A6 r- \2 [
asterisk, P,0.05 using one-tailed Student’s+ ^, c& |2 S* W$ N3 [
t-test. d, Anti-CCL5 neutralizing antibody or * n1 a) k% O' b& O- b7 Lcontrol IgG was administered intraperitoneally % L! k; E' A& \0 K" {5 u' _% }0 Vtwice weekly in SCID mice bearing MDA-MB-2313 t. G* N) Q3 O M
(n59) or MDA-MB-2311MSC tumours2 W9 f: j8 h/ Q8 A
(n511). Representative lung pictures of the 4 E0 }( Q# ?. a% xindicated cohorts are shown. e, Lung metastasis0 \6 ?9 B% w" q# s
indices of mice in d. Data shown are + c, G5 [! x9 L* @" {8 ], brepresentative of means6s.e.m. Asterisk, , m \. X v5 b; `8 ^4 bP,0.05 in one-tailed Student’s t-test. 2 g6 N5 t6 d A8 B3 B6 {kinase inhibitor LY294002. Drug concentrations that did not inhibit 3 u! W. {2 G9 @) W7 j+ a2 Lthe basal motility levels of MDA-MB-231 cells blocked the elevation+ M1 D- `: f! h
of motility induced by ectopic CCL5 expression (Fig. 4g). These$ D: `/ t# e& C/ o
results, when taken together, suggest that the observed CCL5- V5 P7 h4 c; e9 ^) o/ Q
enhanced lung colonization could be ascribed, in significant part, 9 w9 j% l, y; Y6 u* I- g& ]6 Mto its ability to promote extravasation and/or motility of cancer cells5 N1 S% u/ @+ o
at sites of dissemination rather than promoting the survival and/or ) D9 U# g, [0 u) [' \proliferation of these cells. 8 V9 l+ F& x0 I( ~Essential role for the CCL5–CCR5 loop 1 n' O/ \9 p- u: SCCL5 acts through three G-protein-coupled receptors, termed # M# |+ G: w# {7 ECCR1, CCR3 and CCR5 (ref. 23). CCR5 has been determined to be+ x5 ^4 o! \3 R3 n8 |
the main receptor for CCL5 in MDA-MB-231 cells, as inhibition of its ) J2 r% W" j/ _/ b2 X4 m2 Osurface expression through dominant-negative mutants abrogated Y! t, D6 l# v5 @- z! o. M9 rthe ability of these cells to respond to CCL5 chemotaxis24. We therefore) G- i8 _* o' H' ^* U+ |# C7 [ Z8 H
focused our efforts on evaluating the importance of the CCL5–- P" }; \2 D% f" p4 N: v" ]
CCR5 interactions in MSC-induced metastasis. : ^5 s- U- M0 p" nWe confirmed that CCR5 is expressed by MDA-MB-231 cells and5 M3 Q. _+ E( C! e
not by MSCs (Fig. 5a), supporting the notion that MSC-derived # `( F9 R, v% a3 kCCL5 acts primarily in a paracrine fashion on MDA-MB-231 cells' U* e7 @4 o' `; k$ \
in the BCC and MSC mixed cell populations described above. To: A& c$ E% |" c: C
probe whether the observed MSC-induced metastasis required) }$ \8 C* v& _: m0 s0 U
CCL5–CCR5 interactions, we inhibited CCR5 expression in MDAMB-+ ]: W4 i5 s' s2 ]% s% d
231 cells by more than 85% through shRNA knockdown (ref. 25 - @$ d7 F7 s9 R3 Xand Fig. 5b), and mixed these cells with MSCs before implantation $ |1 z1 y' M& Q' e# ` }0 Hinto host mice. Indeed, inhibition of CCR5 expression in the BCCs,! t: B% i1 p$ h* k3 ]! e" I! Y% I5 i, Q
achieved using either of two different shRNA constructs, abrogated ( {; F' A# w9 S# c4 ethe ability of MSCs to enhance the metastasis of MDA-MB-231 cells g* D( H. p! ` {# k6 v" a(Fig. 5c). Furthermore, neutralization of CCL5 protein using intraperitoneal' M8 W' I- J: N* p6 \
injections of an anti-human CCL5 monoclonal antibody 1 i' R! o3 l. s/ `3 k7 halso abrogated the MSC-induced metastasis by MDA-MB-231 cells9 `1 ~5 a# B ]" R7 V
(Fig. 5d, e). In addition, MSCs in which CCL5 expression was inhibited & N" R! }" y. S* d1 N1 gby shRNA knockdown failed to promote metastasis of the$ s q( |& Q0 `0 w8 t
admixed MDA-MB-231 cells (data not shown). Taken together, these" [. {' X9 p# X' T4 ~8 g6 k
results underscore the critical importance of the CCL5–CCR5 paracrine# d; y% k5 C- L9 @8 H
interactions in enabling MSCs to induce metastasis of the1 h- s. v% l( Y
MDA-MB-231 cells. + u/ x P& E% D8 g0 \Discussion/ U) Y: O( t" {# J
Certain models of metastatic progression propose that cancer cell , ]+ m4 A) j. c6 J, y' f& winvasion and metastasis from the primary tumour site are strongly ! _. [8 o4 r2 x% c1 o# l& {$ binfluenced by contextual signals emanating from the stroma of the; A* [0 q# m, F( o- i2 Q+ n
primary tumour. It follows that if carcinoma cells are subsequently6 R+ X' M: M( O7 Y0 N
deprived of such signals, they may revert to an earlier phenotypic' J4 ]; k# J1 v) J& B
state in which they no longer display the traits of high-grade malignancy.! V4 V( U0 ^5 S6 c4 ~* |# I
Indeed, such a model has been proposed previously by others & V& {: @" N: |4 a3 z/ g: won the basis of indirect evidence21. Here, we demonstrate that at least1 U1 W) v- Z9 t6 U# Y
one mesenchymal cell type, the MSC, can expedite tumour metastasis,0 j$ e* j, V2 w1 g
and suggest that after primary human carcinomas recruit MSC( Z1 A; l: W9 n: N: u! H
populations into their midst, subsequent interactions between the+ w# B2 o3 o: N2 `9 V# n
MSCs (or their derivatives) and the BCCs endow the latter with ' t6 B& ?# s2 _invasive and metastatic properties. : G6 C; w4 x# y# z( n, P9 [Although the recruitment of labelled MSCs to tumour xenografts) r Z/ T1 w7 Q3 i" [, E+ V9 Y9 E# ~
has been established in a variety of experimental models of tumorigenesis,3 [1 p& F& e; c* P
there is currently no available way to quantify with any accuracy" Z& v9 p% {: v6 b7 f
the number ofMSCs in actual human tumours, in part because no set . b, M* A% C) p$ zof markers has been identified that can uniquely stain these cells without 4 z( Z' z" n( o5 D* r( [concomitantly staining other mesenchymal types in the tumourassociated L$ C3 y, }: h4 G/ M [. ]* o
stroma6. Our demonstration that the stroma derived from : x3 a. q/ N/ ~1 |. b" Rtumour xenografts contained appreciable numbers of murine MSCs 4 l0 d- x1 [7 Q" N# z. @6 v$ j$ `indicates that significant steady-state levels of these cells aremaintained 1 D% o2 {2 M9 Q8 ^in developing tumours. Interestingly, the use of CD10—one of the! Z0 `4 F4 [5 I6 T! O r$ C! ]+ p) b
markers associated withhumanMSCs—to purify cells fromthe stroma * q$ _ A8 u* L& ]8 aof human primary invasive breast carcinomas yielded a population of + i1 V3 A& d% B+ w- R( wcells that expresses a number of other markers collectively used to2 H6 r5 Z1 @+ o" ` j! p
characterize human MSCs (for example, CD44, CD105 and CD106;3 ^/ c4 [! |. i, U
Fig. 6a). This suggested that, similar to tumour xenografts, human3 Q; p0 L, S% O% i, M6 o9 p* W5 t) G
carcinomas also acquire significant numbers of MSCs. Furthermore,: F" a4 G1 T6 I* b: ?1 g
we note that CCL5, which is prominent in the stromal gene expression : Y E5 X: w$ {3 m( Lsignature associated with poor prognosis of breast cancers26 (SFT;+ f! r& i- M7 ?
Fig. 6b, c), is also enriched in the leukocyte- and endothelial cell-free 8 q4 H' _0 j$ G# gstroma of primary invasive ductal carcinomas (Fig. 6d), specifically in 1 `, c+ G6 U" |* z$ lthe CD10-positive compartment27 (Fig. 6e). Collectively, these observations " S" N' C! Y( c# O, ? Yargue strongly for a significant association between stromal- p* k+ N" e) L# n3 l; x' J
CCL5 levels, MSCs and human invasive breast cancers.5 l9 k' M/ ]8 v3 Z/ U& c
c . e" \& D# N/ K3 M2 [STT1969B ^; J' P' S+ u+ F) k) t* S/ o
STT3126 . I0 R% r! Y9 i% B+ cSTT3124 ! h% G3 d# K% @. P7 ~! C& L0 W) QSTT656B ) T5 w; G) X1 D0 `STT1968B' O9 m# N6 ]& D; V: G% o" i( X
STT3122 / [% ~1 }8 f& g: H: ^! m- kSTT3053/ y+ P& U" I! I: P: J& b @
STT1986B G$ ~) K7 K# k& M: Y. m
STT854' O, s+ K+ M4 d
STT3125 . S0 _. N3 A2 I# }! k- @* ASTT1975 ' P/ l K% l" [+ `9 H, QSTT1987B ) }+ }3 h6 W+ V! t; oSTT1079- d3 e$ ~' w* w- d" t# T Q+ j3 D
STT638 9 F! T. Q0 T+ p# r/ k5 CSTT1774- |. f+ ~3 {# c
STT19840 w- v' B6 e* X. l- v
STT1737C& K% ]& x3 e9 K7 ^6 } ]2 z# E- L
STT3068: T$ g9 ]7 ^3 G4 ^. _1 j: E
STT3120/ a6 _! ~& Q" b
STT850% L$ S/ ^0 X7 Q7 S
STT417B 0 d1 o! a% A. U9 jSTT31199 l, W0 E! E& h& H0 s
STT1776 ( i! o3 e3 Z3 A( u' k7 ? p2 DSTT1777B # {6 r3 c; {8 @1 o0 t* ]STT689B : i* V8 g8 N" \* eSTT1971 7 x) u Z X' @% N {. c5 X9 mSTT597 ( \" V1 U5 B+ t2 x8 P2 jSTT6267 A- q5 I, e5 S) `2 @
STT154 # B9 S+ i* m) p. n" tSTT2774# ]6 C% ]0 S' [1 G$ q. ]
STT1966 9 z/ \* ^9 O3 Z$ y6 I- k- rSTT27763 H! f9 w$ W7 g8 B& n
STT2775 , P: a: O/ W( xSTT2772+ ^& g. |2 N8 {; T, C# ]
STT1637 5 \+ g0 {5 s6 {! eSTT1220C2 p Z' c0 L) r n; j4 k
STT-094B-1; D$ B5 O8 w+ f
STT675 " f' Z0 Q" Q1 Y; _STT2770: H6 }1 y; I4 R O: r2 g
STT695B) x; ]4 p P) M9 \
STT17719 F) z [! [8 F
STT1778 ) \% K/ c/ m, {0 I1 c n( h1 H y3 Y/ ySTT491 ' [; k \4 T: ~; F+ V M8 eSTT18230 b5 p0 O7 l9 t' Q
STT200C7 q0 o8 e! [1 S I0 p: Q
STT741B ! p2 C# J' F NSTT335C& D2 g S; ~. g: \3 |1 k8 c
STT709B7 F( h, a- f1 ^. r f( D3 h; p% o
STT516D) S! w; z* ]4 |7 O0 _7 S- m: e" ]& `# x3 Q/ M
STT607B; U! }) l) C. b& A: l
STT680B 8 n I2 F7 _7 ySTT1148B $ o# d0 O; S) Q* G( SSTT523B q# }5 v- b0 l% M! a8 e" k
STT526E 4 _0 F. m3 Y/ J, M% N; e5 ?1 PSTT742F & T `, H- D. o8 Q! U$ B4 Qe ; ]3 ^5 u5 w9 ~) K4 jCD13/ m- \8 n: K- P9 c8 _
CD290 s3 P0 z! a0 r
CD444 z5 @# d: h6 A' D: u6 |
CD49e( v% h' C3 \% V+ [7 T8 ?
CD54 3 ^8 r% i7 l( M' L8 e( X8 ECD59; [& i$ _6 ^$ U7 ?6 {; O' u
CD633 }: N. ?4 ~: l ]& E/ h
CD1058 D* L6 d5 \; s2 T: M \
CD106 ~; w o4 `. U. s1 S# E
Nestin" R: M G0 z' i# f+ L0 z
HAS2 1 v" I* O! H UIGF24 @; w, B! W" L2 |) a8 U
PLAU# m6 U1 O3 E: N$ m4 T2 Y7 W
TIMP1 0 j4 U. Q& d, @$ ]/ M0 m2 gCAV1 ) ]% }$ M; [+ j$ X4 y# C7 H) e7 rIDC-71 i4 `# ?, v0 D8 v4 J- _
T1126030 i ]3 ~3 y3 r q
T392303 " @0 U+ d* C$ e& ~: m/ QNormal Invasive# ~! ~2 D" ] u9 P
CCL5 % g+ N& E& M3 y. R8 B; ?log2 ratios' K$ I' J* G- c. C) _7 S/ S
–2.0 –1.4 –0.9 –0.3 0.3 0.9 1.4 2.0 , u4 B& p/ X: `3 U; tlog2 ratios 1 o$ g* N: D& K$ w' ~d * {. }$ w6 x5 s–2, G5 H% W5 ^5 l( l' D/ b* n' z
–1 4 Z5 O8 F! m; Y* r0 1 2' B% k+ i9 ]) A$ s% |
–2.0 4 N4 L0 L- x/ l- H% l& k–1.5 2 B2 [' h& @9 F' m) o3 e8 |–1.09 E5 F7 W7 _! z9 J8 s
–0.53 X; [" m3 X0 A$ H# ]
0.0 ) O% R9 Y4 D* r0 UDTF SFT/ X; C E b, ^! J/ U
DTF * f$ y S( U R& |$ FSFT) z9 Y& l. ^- }$ I3 \0 ]
a b $ ~ h( O8 h s0 rF * `9 _6 d& c3 k* g- h( JFigure 6 | Stromal fibroblastic cells of human invasive ductal carcinomas are ) m! `+ S2 I8 ]; o# z7 krich in MSC markers and overexpress CCL5. a, SAGE TreeView display of6 c8 E" d, s8 Z R2 F! x
MSC markers expressed in stromal CD10-positive cells from invasive2 F4 h; z5 N- t9 Q8 x8 A
tumours27. b, Soft-tissue tumourswere ranked byCCL5 expression26, fromlow 6 f3 x; ?6 E2 p m(green) to high (red). Wide blocks indicate expression ratios of tumours $ ~/ ]: {5 e) p- S4 `classified as desmoid-type fibromatosis (DTF; yellow outline, n510) or4 Q- F W [# g
solitary fibrous tumours (SFT; blue outline, n513); narrow blocks are other4 [7 L5 Z: Z# e% D
soft-tissue tumours (n532). c, Box plot showing that CCL5 expression is3 I/ F2 S: }& I* _4 L5 b ?' T7 a3 Q
higher (P50.004) in SFT than inDTF. The difference in log2 expression ratios ! p' o% p8 |+ u& Q# j/ N. obetween SFT and DTF was tested with the Welch’s test. d, CCL5 Affymetrix 8 q# Q5 B- @% n) ~" J, e4 X xgene expression in the stroma of human invasive ductal cancers compared to- g5 v& |' H0 C3 R( _" F
that in normal cancer-free breast tissue (indicated as ‘Normal’; see Methods). 9 @ s% u: S' r5 u/ ie, CCL5 expression is mostly restricted to the CD10-positive fibroblastic cells ( k6 e3 }, o0 J# Qderived from invasive ductal cancers. The heatmap shown is a cluster of - |: F4 _2 b( r1 r9 B7 f1 W( t9 kCCL5.genelist obtained as in a.Details of thepurificationmethodologies of the+ O. b4 ^3 i' d6 ?
various groups indicated in a, d and e are found in ref. 27.Although we have focused here on CCL5 in the MSC–MDA-MB-2 T0 y' w+ {* v7 i1 W+ M
231 cell interactions, CCL5 seems to have an equally critical involvement' `+ K0 ^8 ^0 i
in the functional interaction of MSCs with MDA-MB-435 ( S! Z& Y3 l1 T3 M4 X: @ Xhuman BCCs. CCL5 levels accumulate synergistically when the two . n; o( l9 F; G0 ?7 ~+ ~& xcell types are co-cultured together (Supplementary Fig. 10a), and# j+ ]7 I4 @8 ^+ m+ q
MSCs in which CCL5 expression was compromised by shRNA knockdown + o! v' L/ e3 hfailed to promote metastasis by MDA-MB-435 cells to which 7 a8 F. N& a% }. N( ltheywere admixed (Supplementary Fig. 10b).With these facts in mind,/ H. y9 ?% z- U: j2 q- _1 [/ n/ t
we point out that CCL5 does not seemto be involved in regulating the8 Q& F2 Z% L z7 X
MSC-induced metastasis of MCF7/Ras or HMLER cells, which may6 y5 C# c) x: {& g8 T
depend on other paracrine factors such as VEGF and interleukin-8. % _1 n. S; d3 X8 K& U! H" M5 x: ^Nevertheless, our observations highlight the recently discovered critical . F% ]9 n& N" c& j- o; Kroles of chemokine networks in malignant progression28,29 and suggest' M9 u# b8 f3 t$ S# Y' t6 }
the possible utility of a variety of CCL5 analogues and CCR5 antagonists * W" K7 a3 H0 z& S5 i/ v. Q! ucurrentlyused in anti-HIVtherapy30 in treatingmetastatic disease. 1 F+ w7 _$ \7 |/ \Notably, we have observed that MSCs induce the metastasis of cells - B9 [( ~7 J9 w! q/ @to the lung that are, on isolation and re-injection into recipient mice,- K( L0 l6 |6 B5 h; j! D, t
no more metastatic than their predecessors in the primary tumour, B- |6 u% l9 p5 b
(Fig. 2e). This indicated that acquisition of increased metastatic* N1 v1 h r: c/ ?% l7 `
powers by these tumour cells was reversible, and suggested that the7 c4 V4 y( s2 e
maintenance of this phenotype depends on continuing contact with + L* s. T9 Y3 L! t6 M) i( Wstromal cells. If extended to other tumour types, the present results 9 @; Y6 k: O3 i1 |$ h; qhold important implications for the molecular analysis of malignant & q. ?9 q# L8 l8 E1 V' Tprogression. They suggest that many of the cellular functions associated1 a. n0 W: ]2 N& }/ ]: f
with invasion and metastasis are often not expressed constitutively, g0 q1 q% H" H! X+ @, r2 n$ q
by carcinoma cells, but rather only transiently in response to / x, B$ }0 ?- G. V: {3 Y4 Lcontextual signals that tumour cells receive from their stromal microenvironment.$ A b4 D" u% S" h: v. y( t: b/ ~
If so, analysis of the gene expression patterns of bulk) a) @- n: ^/ `0 Q d
primary tumour populations may fail to detect the expression of key ( a4 W; M: M; B, l1 igenes mediating invasiveness and metastasis, if only because they are # {* ^! l! F- |% Abeing transiently expressed in minor subpopulations of cells within ( v( m, K0 T+ I3 Lsuch tumours. Additionally, attempts at determining the metastatic 1 k9 n& C% i: g& L1 N9 @0 h* Vpropensities of tumours may need to be focused on the genes and 2 Q: l& _1 U" A; d; |2 Z: ]proteins that confer responsiveness of primary tumour cells to stromal 8 H; U- W6 t8 m5 m( u0 X$ i' C7 hsignals, rather than on the genes and proteins that directly mediate ( C4 @( q# y, f6 W, }the cellular phenotypes of invasion and metastasis.; d9 [, A, X( c8 [1 {
METHODS SUMMARY & d& ~8 V" X$ dCells labelled with GFP or ds-red, or harbouring various overexpression or J7 R0 { z6 r8 M+ rshRNA constructs, were generated by viral transduction followed by FACS3 W2 ^7 {5 |, h8 q8 G
enrichment or antibiotic selection. Xenograft experiments were conducted in : E5 T f+ n" ]) X7 G! c' bnude or NOD/SCID mice and metastasis was estimated using fluorescence ; h% o( w+ B7 S7 k7 F7 `microscopy. The levels of cytokines, growth factors and chemokines were - ^. L5 ~; \1 O/ Yassessed by immunoassays. Migration and invasion assays were conducted using6 _6 @/ u. D+ {4 |1 M* J2 _
transwell chambers. Antibody treatment of tumour-bearing mice was conducted( k4 n+ B7 D% k1 Y# U$ |7 n
by intraperitoneal injections. See Methods for detailed information regarding: \" G( T9 H7 Z5 `! t
cell culture, viral infections, in vivo colonization and extravasation assays, RT–1 Z+ W" s( q* {* B( f
PCR, TUNEL and anoikis assays, immunohistochemical and immunofluorescence( i+ G# F5 L% ]
determinations, western blotting, and antibodies used. ! k; M) d. d% x f: @* e7 ?$ vFull Methods and any associated references are available in the online version of- q5 I0 v0 b+ b* o0 `1 P0 z
the paper at www.nature.com/nature.作者: 科研人 时间: 2015-5-24 16:27