Galectin-3 depletion tames pro-tumoural microglia and restrains cancer cells growth

Cance Le e s 591 (2024) 216879 A ailable online 16 Ap il 2024 0304-3835/© 2024 The Au ho s. Published by Else ie B.V. This is an open access a icle unde he CC BY-NC-ND license (h p://c ea i ecommons.o g/licenses/by- nc-nd/4.0/). O iginal A icles Galec in-3 deple ion ames p o- umou al mic oglia and es ains cance cells g ow h Albe o Ri e a-Ramos a , b , 1 , Luis C uz-He n´ andez a , b , 1 , Rocío Tala e ´ on a , b , 1 , Ma ía Te esa S´ anchez-Mon e o a , b , Juan Ga cía-Re illa a , b , Ma a Mule o-Ace edo a , b , Tomas Deie bo g c , Jos´ e Luis Vene o a , b , Manuel Sa mien o So o a , b , * a Ins i u o de Biomedicina de Se illa, IBiS/Hospi al Uni e si a io Vi gen Del Rocío/CSIC/Uni e sidad de Se illa, Se illa, Spain b Depa amen o de Bioquímica y Biología Molecula , Facul ad de Fa macia, Uni e sidad de Se illa, Spain c Expe imen al Neu oin lamma ion Labo a o y, Depa men o Expe imen al Medical Science, Lund Uni e si y, Lund, Sweden ARTICLE INFO Keywo ds: Galec in-3 Mic oglia B ain me as asis Glioblas oma Tumou -associa ed mic oglia and mac ophages (TAMs) ABSTRACT Galec in-3 (Gal-3) is a mul i unc ional p o ein ha plays a pi o al ole in he ini ia ion and p og ession o a ious cen al ne ous sys em diseases, including cance . Al hough he in ol emen o Gal-3 in umou p og ession, esis ance o ea men and immunosupp ession has long been s udied in di e en cance ypes, mainly ou side he cen al ne ous sys em, i s ele a ed exp ession in myeloid and glial cells unde sco es i s p o ound impac on he b ain’s immune esponse. In his con ex , mic oglia and in il a ing mac ophages, he p edominan non- cance ous cells wi hin he umou mic oen i onmen , play c i ical oles in es ablishing an immunosupp essi e milieu in di e se b ain umou s. Th ough he u ilisa ion o p ima y cell cul u es and immo alised mic oglial cell lines, we ha e elucida ed he cen al ole o Gal-3 in p omo ing cance cell mig a ion, in asion, and an immu- nosupp essi e mic oglial pheno ypic ac i a ion. Fu he mo e, employing wo dis inc in i o models encom- passing p ima y (glioblas oma) and seconda y b ain umou s (b eas cance b ain me as asis), ou his ological and ansc ip omic analysis show ha Gal-3 deple ion igge s a obus p o-in lamma o y esponse wi hin he umou mic oen i onmen , no ably based on in e e on- ela ed pa hways. In e es ingly, his esponse is p om- inen ly obse ed in umou -associa ed mic oglia and mac ophages (TAMs), esul ing in he supp ession o cance cells g ow h. 1. In oduc ion The umou mic oen i onmen (TME) o e s a ich spa io empo al dynamic o cellula componen s. Apa om malignan cance cells, he e a e a complex my iad o issue esiden and in il a ed pe iphe al immune sys em cells. This e ol ing mosaic o mul i unc ional cells makes he TME a a he complica ed opic o s udy in he ield, wi h g ea impo ance o de elop no el he apies ha may e ec i ely be used in he mos challenging cance s o ea . B eas cance has been ecen ly placed as he mos diagnosed cance wo ldwide, wi h g ea a ini y o colonise dis an o gans, specially he b ain [1]. Thus, depending on umou sub ype, b eas cance b ain me as ases (BCBM) a e e y common and ep esen a majo heal h issue wi h meag e su i al a es amongs pa ien s (ca. 12 mon hs) [1]. In u n, one o he mos common o ms o p ima y b ain umou s is glioblas oma (GB), he deadlies ypes o b ain cance , whe e pa ien s who unde go ea men ha e an a e age su i al measu ed jus in mon hs (ca. 15 mon hs) [2]. Despi e p ima y b ain umou s and he me as a ic coloni- sa ion o he b ain a e comple ely di e en diseases, hey g ow unde he in luence o he unique immune mic oen i onmen ound in he cen al ne ous sys em, which clea ly d i es he disease p og ession [1, 3]. BCBM and GB a e conside ed immunologically “cold”, owing o hei lack o umou an igens, ine icien T-cell in il a ion and an abundan p esence o myeloid-de i ed immune cells [4]. Amongs he b oad ple ho a o cells ha cons i u e he TME, he mos abundan non-cance ous cell ype a e mic oglia and in il a ed mac ophages. Tumou -associa ed mic oglia and mac ophages (TAMs) ha e been * Co esponding au ho . Dep . Biochemis y and Molecula Biology. P o eso Ga cía Gonz´ alez s , Uni e si y o Se ille, Spain. E-mail add ess: [email p o ec ed] (M. Sa mien o So o). 1 These au ho s con ibu ed equally. Con en s lis s a ailable a ScienceDi ec Cance Le e s jou nal homepage: www.else ie .com/loca e/canle h ps://doi.o g/10.1016/j.canle .2024.216879 Recei ed 26 Oc obe 2023; Recei ed in e ised o m 8 Ap il 2024; Accep ed 9 Ap il 2024 Cance Le e s 591 (2024) 216879 2 desc ibed o de elop an an i-in lamma o y pheno ype and ul ima ely suppo BCBM and GB g ow h [5,6]. In ha sense, one key componen o he immunosupp essi e beha iou o TAMs is Galec in-3 (Gal-3) [7,8]. This β-galac oside-bind- ing p o ein is long known o be a poo p ognos ic ma ke in b eas cance [9] and glioblas oma [10]. Gi en he immunomodula o y ole o Gal-3, a po en ial ole o his lec in in d i ing a p o- umou al TME in b ain umou s is wo h being conside ed. P e ious wo ks ha e shown how ac i a ed mic oglia a e he majo sou ce o Gal-3 wi hin he b ain [7,11], oge he wi h as ocy es [12]. The e o e, his s udy has ocused on he ole o TAMs-associa ed Gal-3, in d i ing immunosupp ession in BCBM and GB. Ou esul s desc ibe a s ong p esence o Gal-3 closely associa ed o TAMs in BCBM and GB human esec ions, and in i o models. Blocking Gal-3 in i o, we managed o shape mic oglia ac i a ion, supp essing hei an i-in lamma o y pheno ype and inhibi ing cance cells mig a ion and in asion. Ou in i o models p o ided compelling e idence sup- po ing a c i ical ole o Gal-3 in d i ing TAMs-associa ed immunosup- p ession wi hin BCBM and GB mic oen i onmen . T ansc ip omic analysis showed a s ong ac i a ion o key inna e and adap i e immune esponse pa hways, hus, empe ing p o- umou al TAMs pheno ype o- wa ds a mo e p o-in lamma o y s a e, con ibu ing o cy o oxic T-cell in il a ion and umou size educ ion. Las ly, GB-bea ing Gal-3 KO mice who unde wen he s anda d chemo he apeu ic agen , Temozolomide (TMZ), showed a syne gis ic e ec in a es ing umou g ow h [13]. 2. Ma e ial and me hods 2.1. Cell cul u e and ans ec ion GL261 (mouse glioblas oma), BV2 (mouse mic oglia) and EO771 (mouse mamma y ca cinoma) cell lines we e pu chased om he Ame ican Type Cul u e Collec ion (ATCC, USA). Fo de ailed cul u e condi ions see Supp. da a. BV-2 cells Gal-3 knockou we e kindly p o- ided by P o Tomas Deie bo g. Fo p ima y mic oglia cul u e, mig a- ion and in asion, please ead Suppplemen a y da a. 2.2. RNA ex ac ion and quan i a i e PCR RNA was ex ac ed om he cell lines using he RNeasy Ki (Qiagen, Ne he lands). Fo p o ocol and p ime s in o ma ion see Supp. da a. 2.3. Wes e n blo ing assay Memb anes we e incuba ed wi h an i-Gal 3 an ibody (1:1000). GAPDH an ibody (Sigma-Ald ich; 1:2000; USA) was used as a loading con ol. Fo p o ocol in o ma ion see Supp. da a. 2.4. Animals and su ge y Expe imen s we e pe o med in 12-week-old mice C57BL/6 (wild- ype, w , n =8) and Galec in-3 null mu an mice (Gal-3 KO, n =8) wi h he same backg ound, bo h lines ob ained om Cha les Ri e s. Female mice we e used o he BCBM model [14] and male mice o he GB model [15]. Animal expe imen a ion was ca ied ou in acco dance wi h he Eu opean Communi y Council Di ec i es (86/609/EU) and Spanish law (R.D. 53/2013 BOE 34/11370-420, 2013) o he use and ca e o labo a o y animals. Fo in i o p o ocol de ails, please see Supplemen- a y da a. 2.5. Immunohis ochemis y and mic oscopy S aining p o ocol was pe o med as p e iously desc ibed [15]. Fo ull in o ma ion, please see Supp. da a. 2.6. Quan i a ion o umou bu den All da a we e analysed blinded o expe imen al condi ions. Tumou g ow h was his ologically assessed as p e iously desc ibed [16,17], (Supp. Figu e 2). Fo ull in o ma ion see Supp. da a. 2.7. TAMs analysis Colocalisa ion s udies we e analysed using he ImageJ so wa e package (NIH) as p e iously desc ibed [18]. Full in o ma ion on he plugins and so wa e can be ound in he Supp. Da a. 2.8. Mic oa ay analysis To s udy he di e ences in genes exp ession be ween w (n =3) and Gal-3 KO mice (n =3), T ansc ip ome Analysis Console (TAC) So wa e was used. Gene Se En ichmen analysis (GSEA) and he Da abase o Anno a ion, Visualiza ion and In eg a ed Disco e y (DAVID) we e also used o u he analyse gene clus e s compa isons and key biological pa hways. Fo ull in o ma ion see Supp. da a. 2.9. Human da a and samples om umou Bank collec ion Samples and da a we e p o ided by Biobank HUB-ICO-IDIBELL, in- eg a ed in he Spanish Biobank Ne wo k and unded by Ins i u o de Salud Ca los III (PT20/00171) and by Xa xa de Bancs de Tumo s de Ca alunya (XBTC, Pla Di ec o d’Oncologia de Ca alunya). We down- loaded he Cance Genome A las (TCGA) da a o he b eas cance (TCGA-BRCA) and he lowe -g ade glioma and glioblas oma (GBMLGG) da abase om UCSC Xena (h ps://xena.ucsc.edu/). 2.10. S a is ical analysis All da a a e p esen ed as mean ±SD. Unless o he wise s a ed, da a we e analysed using G aphPad P ism ( 8.4). Fo 2 g oups, S uden ’s - es (2- ailed) was used. Fo mo e han 2 g oups, 1-way ANOVA wi h Tukey’s mul iple compa ison es o wo-way Ano a wi h Bon e oni es we e used and adjus ed P alues a e epo ed. S a is ical signi icance was de ined as p <0.05. 2.11. Da a a ailabili y s a emen T ansc ip omic da a we e gene a ed a he Genome and Sequencing acili ies in he Ins i u e o Biomedicine o Se ille (IBiS, Spain). Raw da a suppo ing he indings o his s udy a e comple ely a ailable om he co esponding au ho (MSS) on eques . 3. Resul s 3.1. TAMs a e a majo sou ce o Gal-3 in he TME o BCBM and GB Fi s ly, we in es iga ed whe he he TME o BCBM and GB showed signi ican le els o Gal-3 and, was closely ela ed o TAMs. The in i o EO771 (Fig. 1A–C, BCBM) and GL261 (Fig. 1D–F, GB) models showed a s ong p esence o Gal-3 in he TME. Using Iba-1, a pan-ma ke o TAMs, we obse ed ha mos Gal-3 exp ession colocalised wi h hese cells. Fu he mo e, o alida e he Gal-3 KO (Fig. 1C and F) model, no p es- ence o Gal-3 was ound, excep o a negligible amoun likely de i ed om he umou cells (Fig. 1A and D). We also aimed a deciphe ing whe he Gal-3 (LGALS3) exp ession had a p ognos ic alue in b eas cance pa ien s. Thus, a e using The Cance Genome A las (TCGA) da abase, we ound ha high exp ession o LGALS3 was signi ican ly linked wi h poo e su i al (Fig. 1G). Fu he mo e, human b ain esec ions om pa ien s wi h BCBM (Fig. 1H) had a s ong exp ession o Gal-3 in he umou oci, wi h clea inc ease in he co e compa ed wi h he ma gins o he umou . Tha Gal-3 A. Ri e a-Ramos e al. Cance Le e s 591 (2024) 216879 3 Fig. 1. A.-) Quan i a ion o TAMs exp essing Gal-3 in he TME in w and Gal-3 KO mice (n =8) injec ed wi h EO771 cells. B.-) Imuno uo escence (IF) image (me ge –le - inse – igh -) showing he me as a ic a ea in a w mouse. (DAPI, blue), Gal-3 (G een) and TAMs (Iba1, ed). C.-) Me as a ic a ea o a Gal-3 KO mouse. D.-) Quan i a ion o TAMs exp essing Gal-3 in he TME in w and Gal-3 KO mice (n =8) injec ed wi h GL261. E.-) IF image (me ge –le - inse – igh -) showing he GB a ea in a w mouse. F.-) Me as a ic a ea o a Gal-3 KO mouse. G.-) GDC TCGA b eas cance da abase analysing he dele e ious impac o high le els o LGALS3. H.-) Quan i a ion (n =6) o Gal-3-exp essing TAMs. BCBM esec ion om a 59-yea s-old woman, showing a s ong p esence o TAMs (Iba1, ed) colocalising wi h Gal-3 (g een). I.-) TCGA da abase showing he s iking nega i e in luence o high le els o LGALS3 in low and high-g ade glioma pa ien s. J.-) Quan i a ion (n =6) o Gal-3- exp essing TAMs. GB esec ion om a 67-yea s-old man, highligh ing (inse ) he a ea wi h s ong p esence o TAMs (Iba1, ed) and Gal-3 (g een). Scale ba s: 100 μ m. *, p <0,0001 (Welch’s - es . A. Ri e a-Ramos e al. Cance Le e s 591 (2024) 216879 4 exp ession was closely ela ed wi h TAMs and umou cells, amongs o he cell ypes. Rega ding TCGA s udies in GB samples, high LGALS3 was also closely linked wi h meag e su i al in low g ade and glio- blas oma pa ien s, sugges ing he dele e ious impac ha his p o ein in hose pa ien s (Fig. 1I). Again, Gal-3 p esence was ound in GB e- sec ions, wi h he same pa e n ound in BCBM samples (Fig. 1J). In gene al, he he e ogeneous p esence o Gal-3 in he TME, likely o ming a eas o pseudopalisades, unde sco es i s in ol emen in he adap i e s a egies o umou cells and TAMs wi hin he s essed TME (e.g. nu ien dep i a ion o hypoxia). Al oge he , wi h hese esul s, we demons a ed a s ong p esence o Gal-3-exp essing TAMs in wo di e en o ms o b ain umou s. 3.2. Gal-3 igge s an an i-in lamma o y pheno ype in mic oglia cell line P e ious wo ks om ou esea ch eam desc ibed he key ole o Gal- 3 in TREM2-associa ed mic oglial ac i a ion [7,11,19]. Tha pheno ypic change in mic oglial esponse upon Gal-3 p esence s ongly sugges ed he need o elucida e he po en ial ole o his lec in in he ac i a ing pheno ype in TAMs. 3.2.1. EO771-condi ioned medium We aimed a deciphe ing how b eas cance -de i ed Gal-3 may egula e mic oglia pola isa ion. To his end, we moni o ed iNOS, A g1, CD206 and PD-L1. iNOS, he p o o ypic ma ke o p o-in lamma o y mic oglia [20] (Fig. 2A), su e ed a signi ican d op when soluble Gal-3 was added o he medium. In e sely, A g1, CD206 and PD-L1 (Fig. 2B–D), p o o ypic ma ke s o an i-in lamma o y (p o- umou al) mic oglia [6,21,22], once TCM was added, hei exp ession was signi ican ly inc eased. Howe e , when Gal-3 was silenced in umou cells and hei condi ioned medium was hen added (Gal3KD), le els o e e y ma ke signi ican ly d opped. Fig. 2. Top ow is de o ed o BV-2 cells ea ed wi h EO771-condi ioned medium. Ba g aphs show he quan i a i e analysis o p o-in lamma o y ma ke iNOS (A), and immunosupp essi e A g1 (B), CD206 (C) and PD-L1 (D). Cells we e ea ed (24h) wi h umou -condi ioned medium om pa en al EO771 (EO) o EO771 cells wi h Gal-3 knockdown (siRNA, Gal3KD). Las expe imen al condi ion was used o s udy he e ec o he exogenous addi ion o Gal-3 o EO-Gal3KD medium. Con ocal Images showing he exp ession o iNOS (E, ed), A g1 (G een, F), CD206 (G een, G) and PD-L1 (H, ed). Le images e e o con ol (DMEM) and igh images o cells ea ed wi h exogenous Gal-3. As pe EO771 g oup, Bo om ow is de o ed o BV-2 cells ea ed wi h GL261-condi ioned-medium. Con ocal images show exp ession o iNOS ( ed, I), A g1 (G een, J), CD206 (K, G een) and PD-L1 (L, ed). Ba g aphs show quan i a i e analysis o he exp ession o iNOS (M), A g1 (N), CD206 (O) and PD-L1 (P). Scale ba s: 50 μ m. A. Ri e a-Ramos e al. Cance Le e s 591 (2024) 216879 5 In con as o iNOS, when exogenous Gal-3 was added back o he me- dium, he le els o e e y immunosupp essi e ma ke we e up egula ed. Rep esen a i e images o he abo e-men ioned ma ke s a e depic ed in Fig. 2E–H. A a gene le el, iNOS showed a simila beha iou as p e iously desc ibed, since i s gene exp ession was signi ican ly up egula ed (>3- old) a e EO-TCM and Gal-3 exposi ion. Simila ly, A g-1, CD206 and PD-L1 we e up egula ed once BV2 we e exposed o TCM and Gal-3 (Supp. Figu e 3). Rega ding he s udy using BV-2 Gal-3 KO cells, hese cells we e less esponsi e o all expe imen al challenges (Supp. Figu e 4). Howe e , iNOS was down egula ed when exogenous Gal-3 was added. In e sely, A g1, CD206 and PD-L1 we e up egula ed upon Gal-3 addi ion, mimicking he p e ious esul s ound wi h pa en al BV-2 cells. All hese da a suppo he ole o Gal-3 as a c i ical p omo e o an an i-in lamma o y pheno ype in mic oglia, by up egula ing classical immune-supp essi e ma ke s. 3.2.2. GL261-condi ioned medium When TCM om Gal-3 KD umou cells was added (Gal3KD), iNOS was signi ican ly inc eased in BV-2 (Fig. 2M). In e es ingly, when Gal-3 was added, iNOS d opped o con ol le els. Con e sely, he absence o Fig. 3. A.-) Sc a ch-wound assay pe o med wi h EO771 cells in DMEM ( ed do ed lines) condi ions. Time-cou se s udy wi h pic u es a 0, 6 and 24h. EO771 cells a e 24h exposi ion o GB1107 (black do ed line). B.-) G aph measu ing gap closu e abili y o cells a di e en expe imen al condi ions: con ol (DMEM, ed line), mic oglia-condi ioned medium (BCV2, G een line), selec i e Gal-3 inhibi o (GB1107, black line) and a e addi ion o exogenous Gal-3 (blue line). C.-) Pic u es o he sc a ch assay pe o med o GL261 cells in con ol (DMEM, ed do ed line) and GB1107 condi ion (black do ed line). D.-) Quan i a ion o gap closu e unde he same condi ions as desc ibed in B. E.-) In asion assay pe o med in Boyden Chambe s o EO771 cells, s ained wi h c ys al iole . Cells we e coa ed wi h pa en al BV2 cells (DMEM) a he bo om well, o BV2 wi h Gal3 knockdown (siRNA –Gal3KD-). Nex expe imen al condi ion was pe o med using exogenous Gal-3 wi h Gal3KD BV2 cells seeded a he bo om (+Gal3), and inally, using pa en al BV2 cells wi h GB1107 (0,1 μ M). F.-) Quan i a ion o numbe o in ading cells. G.- Same as E, wi h GL261 cells. H.-) Same as F wi h GL261 cells. *p <0,05; **p <0,01; ***p <0,005. One-way ano a wi h Tukey pos hoc analysis. Scale ba 750 μ m. A. Ri e a-Ramos e al. Cance Le e s 591 (2024) 216879 6 Gal-3 in he TCM (Gal3KD) d opped A g1 le els (Fig. 2N), whils Gal-3 addi ion had an opposing e ec , inc easing A g1 exp ession. Rega ding CD206, i s le els emained unal e ed excep when cells we e coa ed wi h Gal3KD-medium (ca. 2- old inc ease, Fig. 2O). This esul s may be expec ed since p e ious wo ks showed ha glioma cells up egula e CD206. Finally, PD-L1 was up egula ed a all expe imen al condi ion (Fig. 2P), wi h special ele ance upon Gal-3 addi ion (ca. 6- old inc ease). qPCR analysis showed how Gal-3 inc eased he gene exp ession o PD-L1 and A g-1 e en highe han he p esence o TCM. In he case o iNOS, bo h Gal-3 and TCM inhibi ed i s exp ession, hus, sugges ing an an i-in lamma o y beha iou in BV-2 cells upon Gal-3 p esence (Supp. Figu e 3) BV-2 Gal-3 KO cells showed almos no pheno ypic changes a e he expe imen al challenges, sugges ing he impo an ole o endogenous Gal-3 o igge he mic oglial esponse in he p esence o TCM. In e - es ingly, PD-L1 was signi ican ly up egula ed when Gal-3 was added (Supp. Figu e 4). These da a a e in ag eemen wi h ou p e ious indings (EO771- condi ioned medium), highligh ing he key ole o Gal-3 up egula ing classical immunosupp essi e ma ke s (e.g. A g1 and PD-L1) in BV-2 [23]. In e es ingly, we epea ed he same expe imen al challenges in p i- ma y mic oglia cells, and he same end owa d he ac i a ion p o- in lamma o y and he supp ession o immunosupp essi e ma ke s was ound [24,25] in cells om Gal-3 KO pups (Supp. Fig. 5 and Supp. Da a). 3.3. Pha macological blocking o Galec in-3 hinde s umou cells mig a ion We pos ula ed ha pha macological blocking o Gal-3 (GB1107) [26], could es ain umou cells mig a ion. We used EO771 and GL261 o s udy hei mig a ing p ope ies in esponse o he cell pe meable GB1107. In e es ingly, bo h cell lines (Fig. 3A–D) showed simila sensi i i y o Gal-3 modula ion. Bo h umou cells exposed o mic oglial-condi ioned medium (BV-2, Fig. 3B and D, g een line) had simila gap closu e speed compa ed wi h con ol (DMEM) condi ions (Fig. 3A–D, ed line). A sligh inc ease in he mig a ion pa e n was obse ed when 0.1 μ M soluble Gal-3 (Gal-3, Fig. 3B–D, blue line) was added. Impo an ly, when bo h umou cell lines we e exposed o GB1107, hei mig a ion was g ea ly educed (Fig. 3A–D, black line). O e all, ou esul s sus ain he ele an ole o Gal-3 in umou cell mig a ion. 3.4. Mic oglial Galec in-3 suppo s umou cell in asion We aimed a s udying he ole o mic oglial Gal-3 in he in ading p ope ies o EO771 and GL261 cells. Pe o ming an in asion assay wi h an a i icial basemen memb ane [27], umou cells we e co-cul u ed wi h pa en al BV2 cells, and BV2 cells wi h Gal-3 exp ession knocked-down (Gal3 KD) (Supp. Figu e 1B). EO771 cells ha eached he luminal side o he memb ane was signi ican ly lowe wi h BV2 Gal-3 KD cells (Fig. 3E–F and Supp. Figu e 1C, Gal3KD). Adding back exogenous Gal-3 had a signi ican inc ease in he numbe o in ading cells (+Gal3). The use o GB1107 (0.1 μ M), mimicked he esul s om Gal-3 KD condi ion, and success ully educed he numbe o in aded umou cells (Fig. 3E–F, GB1107). Simila ly, GL261 cells mi o ed esul s ound in EO771 cells (Fig. 3G). When coa ed wi h Gal-3KD BV-2 cells (Gal3KD), numbe o in ading cells d opped ((Fig. 3G–H). Again, when Gal-3 was added back o he medium (+Gal3), he numbe o cells eached con ol alues (DMEM). Impo an ly, when GB1107 was added, he numbe o in asi e cells signi ican ly d opped, hus, con i ming he impo an e ec o Gal- 3 ega ding in ading capaci y o bo h umou cell lines (Fig. 3H, GB1107). These esul s unde sco e po en ial ole o Gal-3 egula ing he TME, exe ing a obus in luence on umou cells mig a ion and in asion. 3.5. Glioblas oma and b eas cance b ain me as asis bu den is educed in Galec in-3 knockou mice EO771 and GL261 cells we e implan ed in wild- ype C57Bl/6 and Gal-3 KO mice (n =8, pe expe imen al condi ion), and he umou bu den was quan i ied a di e en ime poin s (Fig. 4, 10 and 21 days). 3.5.1. B eas cance b ain me as asis Since 99 % o b eas cance cases accoun o women [28], we ocused ou BCBM in i o s udy in emale mice. E alua ion o he me - as a ic bu den wi hin he b ain o w and Gal-3 KO mice emained simila a day 10 (Fig. 4A–B). Howe e , a day 21, Gal-3 KO mice showed a signi ican d op o he me as a ic bu den (Fig. 4C–E, ca. 8- old). 3.5.2. Glioblas oma Simila o he BCBM model, GB size showed no di e ence a day 10 (Fig. 4F–G). Again, a day 21, he umou size was signi ican ly educed in Gal-3 KO mice (Fig. 4H–J, ca. 5- old). 3.6. Galec in-3 a ec s ascula densi y and angiogenesis in he TME The ole o Gal-3 in p omo ing angiogenesis in cance s ou side he b ain has been ex ensi ely s udied [24]. Ou indings show ha Gal-3 KO mice exhibi ed educed angiogenic essels compa ed o w (Supp. Figu e 6). Fu he mo e, ascula densi y was signi ican ly dec eased in he BCBM model, while he numbe o blood essels emained simila in he GB model. These indings highligh he po en ial impac o Gal-3 on angiogenesis and ascula isa ion in di e en b ain cance models. 3.7. Galec in-3 deple ion shi ed TAMs ac i a ion owa ds a p o- in lamma o y pheno ype We quan i ied immuno luo escen colocalisa ion o key p o- and an i-in lamma o y ma ke s in TAMs. In bo h models, we clea ly obse ed a pheno ype shi owa ds a p o-in lamma o y pheno ype o e immunosupp ession (Fig. 5). No di e ences we e ound a day 10 a e umou implan a ion in ei he model (da a no shown). Howe e , a day 21, in he BCBM model, he exp ession o iNOS wi hin he TAMs popula ion was inc eased om 15 % in w animals, o 40 % in he Gal-3 KO model (Fig. 5A–C). Impo an ly, A g-1-exp essing TAMs we e educed 50 % in Gal-3 KO mice (Fig. 5D–F). Again, hese indings sugges he impo an ole o Gal- 3 in d i ing he p o- umou al TAMs pheno ype. Wi h ega d o he GB model, pe cen age o iNOS-exp essing TAMs inc eased om 23 % (w ) up o 58 % (Gal-3 KO) (Fig. 5G–I). On he con a y, he s ong immunosupp essi e pheno ype in w mice (high exp ession o A g-1), was signi ican ly d opped in Gal-3 KO mice (4- old dec ease, Fig. 5J-L). Unde s anding he complexi y o cha ac e ising he TME, owing in g ea pa o he b oad numbe o po en ial p o- s an i-in lamma o y ma ke s o be used, he iNOS/A g-1 balance (Supp. Figu e 7A) may sugges ha a e Gal-3 deple ion, TAMs showed a shi owa ds an an i- umou al ac i a ion by inc easing p o-in lamma o y iNOS and educing he exp ession o immunosupp essi e A g-1 [15]. In e es ingly, he iNOS/A g-1 balance in o al numbe o cells (beyond TAMs) was il ed owa ds he p o-in lamma o y side (Supp. Fig. 7). Fu he suppo ing he an i- umou al scena io gene a ed by TAMs in Gal-3 KO mice, CD68-posi i e cells we e inc eased in he BCBM Gal-3 KO model, sug- ges ing ha Gal-3 deple ion migh enhance phagocy ic ea u es in TAMs. (Supp Fig. 8). Las ly, in alignmen wi h ou p e ious indings, we employed GB1107 (28), o in es iga e i s po en ial impac on TAMs pheno ype. Gi en ha b ain me as ases g ow in close p oximi y o ce eb al blood A. Ri e a-Ramos e al. Cance Le e s 591 (2024) 216879 7 essels, we aimed o de e mine whe he GB1107 could shi TAMs beha iou owa ds a p o-in lamma o y s a e. Rema kably, despi e no signi ican educ ion in b ain me as asis size (da a no shown), he TAMs pheno ype in GB1107-injec ed mice mi o ed ou esul s om Gal-3 KO mice. We obse ed a signi ican inc ease in iNOS and a down egula ion o A g-1 ac oss he TAMs popula ion (Supp. Fig. 9). These indings sugges ha GB1107 has he po en ial o modula e TAMs pheno ype in he TME. 3.8. Gal-3 KO mice showed a obus an i- umou T cell in il a ion Tumou -in il a ing lymphocy es (TILs) a e also impo an componen s o he TME. High le el o CD4 + TILs combined wi h low CD8 + TILs is associa ed wi h poo p ognosis in b ain umou s [28]. Since A g-1 exp essing TAMs impai T cell esponses inducing immunosup- p ession [15], we pu sui he po en ial e ec o Gal-3-dependen TAMs pola isa ion in TILs popula ion. In BCBM, he in a umou al CD4 + T cells in il a ion was signi i- can ly inc eased in w mice (Fig. 6A–B). Impo an ly, CD8 + T cells in il a ion had an opposing pa e n and we e mo e abundan in Gal3 KO mice (Fig. 6C–D). This CD8 + /CD4 + T Cell a io has long been desc ibed as a good p ognosis ma ke in b ain umou s [28]. In he GB model, TILs mi o ed he esul s ound in he BCBM mice. CD4 + T cells we e less in il a ed in he Gal3 KO mice (Fig. 6F–G), whils Fig. 4. Con ocal images o he me as a ic a ea o w (n =8) (A) and ansgenic Gal3KO (n =8) (B) emale C57Bl/6 a day 10 and day 21 (C and D). E.-) Quan i a ion o he me as a ic a ea in w ( ed) and Gal-3 KO mice (ligh ed ba s). Con ocal images o glioblas oma a ea in w (F) and Gal3KO (G) mice injec ed wi h GL261 cells a day 10. H and I a e images aken a day 21 pos injec ion. J.-) quan i a ion o glioblas oma a ea compa ing w (da k G een) and Gal3KO mice (ligh G een) a 10 and 21 days. S a is ical analysis o umou olume p og ession was made independen ly ac oss bo h imes. One-way ano a and Tukey pos hoc es s was pe o med. ***, p <0,005. ****, p <0,001. A. Ri e a-Ramos e al. Cance Le e s 591 (2024) 216879 8 CD8 + T we e signi ican ly inc eased (Fig. 6H–I). We obse ed a ypical clus e ing pa e n in TILs, p esumably a si es o an igen p esen a ion [29]. TAMs, as an igen-p esen ing cells, had a deep impac in TILs, since he p o-in lamma o y pheno ype ound in Gal-3 KO mice, suppo ed a umou icidal CD8 + /CD4 + TILs a io (Fig. 6E and J). 3.9. Tumou size was u he dec eased a e emozolomide ea men Following he signi ican educ ion in umou size obse ed in Gal-3 KO mice, we sough o de e mine whe he TMZ adminis a ion could po en ially enhance he al eady desc ibed an i- umou al mic oen i- onmen . Ou indings e ealed ha animals ea ed wi h wo ounds o TMZ exhibi ed a signi ican educ ion in GB compa ed o hose ea ed wi h TMZ alone (Supp. Figu e 10). Mo eo e , ou da a indica e ha TMZ shi ed TAMs owa ds a p o-in lamma o y s a e. Speci ically, TMZ e oked an inc ease in iNOS exp ession in w and Gal-3 KO mice, while A g1 was signi ican ly educed in TMZ- ea ed Gal-3 KO mice (Supp. Figu e 10). These indings sugges ha Gal-3 de iciency may enhance he an i- umou al e ec s o TMZ, which could ha e impo an implica ions o GB ea men . 3.10. Gal-3 deple ion enhanced a mul i a ious p o-in lamma o y gene exp ession signa u e in he TME In an ini ial e o o explo e he p e-clinical ele ance o ou ind- ings, we pe o med nex gene a ion ansc ip ome-wide gene-le el exp ession p o iling s udies in he TME. Tumou s om BCBM and GB animals we e ha es ed a day 21 pos -implan a ion, and hen analysed. 3.10.1. B eas cance b ain me as asis Compa ing he TME o w and Gal3 KO mice, we obse ed 180 dys egula ed genes (100 genes up egula ed and 80 genes down- egula ed). The mos o up egula ed genes in he Gal-3 KO model we e immune sys em- ela ed, closely linked o p o-in lamma o y ma ke s (e. g. Tn s 10, Tn s 1b), commonly exp essed in mic oglia and mac o- phages (e.g. CD68, P2 y14), T cells (e.g. CD2, CD2, CD8b, G anzyme A- B) and Kille cells (Kl b1, KL 1a). On he con a y, w animals showed p o-me as a ic genes exp ession (e.g. Se pinb9 , Se pinb3b) [30] and a ginase-1 [31] (Fig. 7A). Gene se en ichmen analysis (GSEA) Fig. 5. A.-) Con ocal images o he me as a ic a ea o w mouse a day 21 a e being injec ed wi h EO771 cells. Pic u es shows he TME (le image) and he inse ( igh ) wi h nuceli s ained in blue (DAPI), TAMs in ed (Iba-1) and iNOS (g een). B.-) Quan i a ion o TAMs exp essing iNOS wi hin he TME (n =5). C.-) As pe A, images showing he TME o Gal3KO mice. D.-Same s udy as abo e-men ioned bu analysing TAMs exp essing A g1 (E) in w (D) and Gal3KO (F). G.-) Glioblas oma a ea o w mouse injec ed wi h GL261 cells a day 21. Pic u es shows he TME (le image) and he inse ( igh ) wi h nuceli s ained in blue (DAPI), mic oglia in g een (Iba-1) and iNOS in ed. H.-) Quan i a ion o TAMs exp essing iNOS (n =5). I.-) As pe G, images showing he TME o Gal3KO mice. J.-) Same s udy as abo e- men ioned bu analysing TAMs exp essing A g1 (K) in w (J) and Gal3KO mice (L). Scale ba s 250 μ m (50 μ m inse s). A. Ri e a-Ramos e al. Cance Le e s 591 (2024) 216879 9 compa ing o e 22000 genes ac oss bo h expe imen al models, showed se e al p o-in lamma o y gene clus e s in he Gal-3 KO model. Specially mic oglial esponse o in e e on gamma was one o he mos dys egu- la ed clus e s ac oss g oups (Fig. 7B–C), which is well-known o play a pi o al ole in d i ing he p o-in lamma o y pheno ype in TAMs [32]. Mo eo e , using wikipa hways analysis om TAC so wa e, Gal-3 KO e oked an inc ease in di e en in lamma o y gene signa u es, wi h ADAR1 edi ing de iciency immune esponse and Type II in e e on sig- nalling as he mos up egula ed pa hways (Fig. 7D). These pa hways a e mu ually ela ed since ADAR1 is a c i ical IFN-inducible o m, display- ing a pi o al ole in d i ing p o-in lamma o y esponses o he immune sys em [33,34]. 3.10.2. Glioblas oma We obse ed 194 dys egula ed genes, 111 we e down egula ed and 83 up egula ed. Again, mos dys egula ed genes a e ypically exp essed in ac i a ed immune sys em cells including mic oglia and mac ophages (Clec7a, CD68, CD180) (Fig. 8A). GSEA showed he in e e on gamma mic oglial pa hway as one o he mos up egula ed clus e s in he immunologic signa u e gene se s (Fig. 8B) Using he Da abase o Anno a ion, Visualiza ion and In eg a ed Disco e y (DAVID) we obse ed ha GO TERM biological p ocesses showed chemokine-media ed pa hways, oge he wi h in e e on gamma and in lamma o y esponses (Fig. 8C). Finally, gene clus e analysis ound up egula ion in Vi al p o ein in e ac ions, which is closely ela ed wi h ADAR-1 [34], and cellula esponse o in e e on gamma (Fig. 8D). Al oge he , ou ansc ip omic analysis o he TME clea ly show an immune- ela ed p o-in lamma o y mic oen i onmen , which signi i- can ly a ec ed BCBM and GB g ow h. 4. Discussion Galec in-3 is a membe o he e olu iona y conse ed animal lec ins amily. I is exp essed in se e al cells ypes playing key oles in di e en biological p ocesses such as apop osis, au ophagy, phagocy osis, in lamma ion, and cance [8,35–38]. Howe e , in he las yea s, ou g oup has unmasked he pi o al ole o Gal-3 in immune sys em cells ac i a ion, speci ically mic oglia [7,19]. Thus, his wo k desc ibes a no el ole o his pleio opic p o ein, ac i a ing an immunosupp essi e pheno ype in TAMs in a b ain umou con ex . Gal-3 induces an i-in lamma o y mic oglial ac i a ion in p ima y and immo alised mic oglia cell lines. We ound a complex beha iou , unde lining he ac i a ing e ec ha exogenous addi ion o Gal-3 exe ed on classical immunosupp essi e ma ke s (e.g. A g-1 and PD- L1), simila o he e ec s obse ed a e he exposi ion o umou - condi ioned medium. In e es ingly, such Gal-3-dependen an i-in lam- ma o y esponse was ab oga ed once Gal-3 was inhibi ed. These esul s a e in ag eemen wi h o he wo ks whe e blocking Gal-3 diminished le els o classical immunosupp essi e ma ke s [25,39]. Howe e , ou in i o esul s also sugges ha umou al Gal-3 plays an impo an ole d i ing such immunosupp essi e mic oglial esponse. In pa icula , he Fig. 6. A.-) Quan i a ion o CD4 + T cells wi hin he TME (g een) o BCBM. Nuclei in blue (DAPI) B.-) Con ocal images o a BCBM showing CD4 + T cells (g een) in he TME in w and Gal3KO mice. C.-) Con ocal images showing CD8 + T cells (g een) in he TME o EO771 mice. D.-) Quan i a ion o CD8 + T cells wi hin he TME o BCBM mice. E.-) Ra io o CD8 s CD4 T cells wi hin he TME o BCBM animals. F.-) Con ocal images o CD4 + T cells in he TME o GL261-injec ed mice. G.-) Quan i a ion o CD4 + T cells (g een) in he TME o GB model. H.-) con ocal images o CD8 + T cells (g een). I.-) Quan i a ion o in a umou al CD8 + T cells in ou GB model J.-) CD8/CD4 T cells a io wi hin he TME o ou GB model. Scale ba 250 μ m (100 μ m inse ). A. Ri e a-Ramos e al.