New, fully implantable device for selective clearance of csf-target molecules: proof of concept in a murine model of alzheimer’s disease

Ci a ion: Co o-Vilcapoma, M.A.; Cas illa-Silgado, J.; Fe nández-Ga cía, B.; Pin o-He nández, P.; Cip iani, R.; Cape illo-Za a e, E.; Menéndez-González, M.; Ál a ez-Vega, M.; Tomás-Zapico, C. New, Fully Implan able De ice o Selec i e Clea ance o CSF-Ta ge Molecules: P oo o Concep in a Mu ine Model o Alzheime ’s Disease. In . J. Mol. Sci. 2022,23, 9256. h ps://doi.o g/10.3390/ijms23169256 Academic Edi o : S e anie Kü en Recei ed: 29 July 2022 Accep ed: 13 Augus 2022 Published: 17 Augus 2022 Publishe ’s No e: MDPI s ays neu al wi h ega d o ju isdic ional claims in published maps and ins i u ional a il- ia ions. Copy igh : © 2022 by he au ho s. Licensee MDPI, Basel, Swi ze land. This a icle is an open access a icle dis ibu ed unde he e ms and condi ions o he C ea i e Commons A ibu ion (CC BY) license (h ps:// c ea i ecommons.o g/licenses/by/ 4.0/). In e na ional Jou nal o Molecula Sciences A icle New, Fully Implan able De ice o Selec i e Clea ance o CSF-Ta ge Molecules: P oo o Concep in a Mu ine Model o Alzheime ’s Disease Ma ía Almudena Co o-Vilcapoma 1,2,† , Juan Cas illa-Silgado 1,2,†, Benjamín Fe nández-Ga cía2,3 , Paola Pin o-He nández 1, Ra aela Cip iani 4, Es ibaliz Cape illo-Za a e 4,5,6 , Manuel Menéndez-González 2,7,8,* , Ma co Ál a ez-Vega 9,10,‡ and C is ina Tomás-Zapico 1,2,‡ 1Depa amen o de Biología Funcional, Á ea de Fisiología, Uni e sidad de O iedo, 33006 O iedo, Spain 2Ins i u o de In es igación Sani a ia del P incipado de As u ias, 33011 O iedo, Spain 3Depa amen o de Mo ología y Biología Celula , Á ea de Ana omía, Uni e sidad de O iedo, 33006 O iedo, Spain 4Achuca o Basque Cen e o Neu oscience, Depa amen o de Neu ociencias, Uni e sidad del País Vasco (UPV/EHU), 48940 Leioa, Spain 5Cen o de In es igación en Red de En e medades, Neu odegene a i as (CIBERNED), 28029 Mad id, Spain 6IKERBASQUE, Basque Founda ion o Science, 48009 Bilbao, Spain 7Se icio de Neu ología, Hospi al Uni e si a io Cen al de As u ias, 33011 O iedo, Spain 8Depa amen o de Medicina Á ea de Medicina, Uni e sidad de O iedo, 33006 O iedo, Spain 9Se icio de Neu oci ugía, Hospi al Uni e si a io Cen al de As u ias, 33011 O iedo, Spain 10 Depa amen o de Ci ugía, Á ea de Ci ugía, Uni e sidad de O iedo, 33006 O iedo, Spain *Co espondence: [email p o ec ed] † These au ho s con ibu ed equally as i s au ho s. ‡ These au ho s con ibu ed equally as senio au ho s. Abs ac : We ha e p e iously p oposed a adical change in he cu en s a egy o clea pa hogenic p o eins om he cen al ne ous sys em (CNS) based on he ce eb ospinal luid (CSF)-sink he apeu ic s a egy, whe eby pa hogenic p o eins can be emo ed di ec ly om he CNS ia CSF. To his aim, we designed and manu ac u ed an implan able de ice o selec i e and con inuous aphe esis o CSF enabling, in combina ion wi h an i-amyloid-be a (A β ) monoclonal an ibodies (mAb), he clea ance o A β om he CSF. He e, we p o ide he i s p oo o concep in he APP/PS1 mouse model o Alzheime ’s disease (AD). De ices we e implan ed in wen y- ou mice (se en een APP/PS1 and se en W ) wi h low a es o complica ions. We con i med ha he aphe esis module is pe meable o he A β pep ide and impe meable o mAb. Mo eo e , ou esul s showed ha con inuous clea ance o soluble A β om he CSF o a ew weeks dec eases co ical A β plaques. Thus, we conclude ha his in e en ion is easible and may p o ide impo an ad an ages in e ms o sa e y and e icacy. Keywo ds: immuno he apy; Alzheime ; be a-amyloid; CSF-sink; ce eb ospinal luid; implan able de ice; blood–b ain ba ie ; nanopo ous memb anes 1. In oduc ion The social impac o neu odegene a i e diseases (NDD) is undeniable. They comp ise a wide ange o di e en neu opa hologies, which can be ei he spo adic o inhe i ed. Howe e , mos o hem sha e he common hallma k o deposi s o disease-speci ic p o eins; hus, NDD can be unde s ood as p o einopa hies. In e es ingly, in he symp oma ic Hun ing on’s disease (HD) mouse model, blocking mu an hun ing in exp ession a o s he disappea ance o agg ega es and amelio a es he beha io al pheno ype [ 1 ]. This opened an insigh in o he pa hophysiology o NDD whe e p o ein agg ega ion o e whelms he p o eos asis capaci y o neu ons (e.g., ubiqui in- p o easome and au ophagy-lysosome sys ems), in e e ing wi h he abili y o neu ons In . J. Mol. Sci. 2022,23, 9256. h ps://doi.o g/10.3390/ijms23169256 h ps://www.mdpi.com/jou nal/ijms In . J. Mol. Sci. 2022,23, 9256 2 o 13 o cope wi h pa hogenic p o eins [ 2 , 3 ]. The o ma ion o agg ega es o hese p o eins can ha e di e en o igins, including, among o he s, inc eased syn hesis o syn hesis o s uc u ally abno mal o ms and/o dec eased deg ada ion, ei he by enzyma ic o cellula sys ems [ 4 , 5 ]. A dec ease in hei clea ance o compa men s ou side he b ain pa enchyma has also been desc ibed, which has been linked o he impai men o he blood–b ain ba ie (BBB) [6], he ce eb ospinal luid (CSF) low [7], and he glympha ic sys em [8,9]. Accumula ion o amyloid-be a (A β ) is he main pa hological hallma k o Alzheime ’s disease (AD). In bo h ea ly-onse and la e-onse o ms o AD, A β clea ance seems al eady impai ed a he p od omal s age o AD, and i is emo ed om he b ain by a ious o e lapping and in e ac ing clea ance sys ems: deg ada ion, BBB anspo , in e s i ial luid (ISF) bulk low, he glympha ic pa hway, and CSF abso p ion in o he ci cula o y and lympha ic sys ems [10,11]. Di e en app oaches ha e been in es iga ed o emo e A β bo h biologically and mechanically, om dec easing p oduc ion (i.e., BACE inhibi o s) o inc easing clea ance in he pe iphe y (immuno he apy, plasmaphe esis, enzyma ic deg ada ion) [ 12 ]. Among hem, immuno he apy wi h an i-A β monoclonal an ibodies (mAb) is he mos ex ensi ely explo ed in humans, showing he capaci y o clea b ain plaques and es o e le els o soluble A β in he CNS [ 13 ]. Howe e , none o hese he apies ha e shown clinically ele an bene i s o AD pa ien s, and se ious side e ec s ha e been epo ed, including amyloid- ela ed imaging abno mali ies (ARIA) a e an i-A β mAb he apies [ 14 ]. These ailu es led o seek al e na i e me hods o elimina e pa hogenic p o eins om he b ain using o he chemical o physical p inciples, such as hemodialysis o plasmaphe esis. Among he esul s ob ained om he esea ch on hese in e en ions, i has been ound ha blood dialysis and plasmaphe esis educe A β le els in plasma and CSF in AD pa ien s and a enua e AD symp oms and pa hology in AD mouse models [ 15 – 21 ]. This sugges s ha emo ing A β om he plasma migh be an e ec i e o m o gene a ing an e lux o b ain Aβ h ough he BBB [22]. The BBB p e en s he ee mo emen o molecules be ween he in e s i ial luid (ISF)/CSF and plasma [ 11 ], while ISF soluble molecules mo e in cons an equilib ium be ween he CSF and he ISF, bo h being compa men s in di ec communica ion [ 7 , 22 ]. Gi en his, we ha e p e iously p oposed a adical change in he cu en pa adigm based on clea ing a ge molecules om he CSF using implan able de ices [ 23 ]. Ou hypo hesis elies on he CSF-sink he apeu ic s a egy, whe eby pa hogenic p o eins ha a e in equi- lib ium be ween ISF and CSF can be emo ed di ec ly om CSF [ 24 – 26 ]. This in ol es an al e a ion o his equilib ium, a o ing he clea ance o hese p o eins in hei soluble s a e and hus dec easing hei a ailabili y o o m agg ega es in he b ain pa enchyma. In e es ingly, he equilib ium be ween ISF and CSF emains s able in symp oma ic AD models, in con as o he balance loss be ween he ISF and plasma [ 22 ]. Thus, he e is a much mo e di ec way o emo ing a ge p o eins om he ISF han clea ing hem om he plasma: clea ing hem om he CSF. To his aim, we p opose o use a p ocedu e we named CSF aphe esis based on an implan able de ice wi h a nanopo ous sys em allowing selec i e and con inuous aphe esis o A β om he CSF. To e alua e he biological e icacy o he sys em, we ha e de eloped a minia u ized p o o ype o conduc s udies in mu ine models o NDD. He e, we p esen he i s p oo o concep in he APP/PS1 mouse model o AD. 2. Resul s 2.1. Implan a ion o he Minia u ized P o o ype Was Feasible in a Mu ine Model o Alzheime ’s Disease Fo his p oo o concep , we ha e used a well-cha ac e ized mu ine model o AD o which he ime a which A β plaques can be obse ed in he b ain is well known (six mon hs) [ 27 ]. To be su e ha plaques would be p esen a he ime o he in e en ion, we used se en-mon h-old mice. On he o he hand, a his age he plaques o med a e no as dense as hey can be in olde mice, hus making i easie o iden i y he e ec o con inuous il a ion ea men in a sho - e m s udy. In . J. Mol. Sci. 2022,23, 9256 3 o 13 The minia u ized p o o ype o he implan able de ice consis ed o an aphe esis module wi h a ese oi (Pa A; Figu e 1A) and a ca he e connec ing Pa A o a e minal cannula o in ace eb o en icula implan a ion (Pa B; Figu e 1A). Pa A o he de ice was placed subcu aneously on he back o he mouse, as was he cannula, and pa B was implan ed in he la e al en icle (Figu e 1B,C). Due o he weigh (4.29 ± 0.41 g), bu especially he size o he de ice (3.5 × 1 × 1 cm), only male mice we e used in he s udy. In he case o he se en-mon h-old APP/PS1 mice, he weigh o he de ice ep esen s 12.86 ± 0.44% o he mouse’s weigh , while in he case o W mice o he same age his pe cen age is 14.92 ±0.38% . On he o he hand, he mos limi ing ac o was he leng h o he de ice, which, in ela ion o he leng h o he back o he mice, ep esen s 66.51 ± 1.77% o he leng h in he case o he APP/PS1 mice and 69.56 ±1.86% in he case o he W mice. In . J. Mol. Sci. 2022, 23, x FOR PEER REVIEW 3 o 14 2. Resul s 2.1. Implan a ion o he Minia u ized P o o ype Was Feasible in a Mu ine Model o Alzheime ’s Disease Fo his p oo o concep , we ha e used a well-cha ac e ized mu ine model o AD o which he ime a which Aβ plaques can be obse ed in he b ain is well known (six mon hs) [27]. To be su e ha plaques would be p esen a he ime o he in e en ion, we used se en-mon h-old mice. On he o he hand, a his age he plaques o med a e no as dense as hey can be in olde mice, hus making i easie o iden i y he e ec o con inuous il a ion ea men in a sho - e m s udy. The minia u ized p o o ype o he implan able de ice consis ed o an aphe esis mod- ule wi h a ese oi (Pa A; Figu e 1A) and a ca he e connec ing Pa A o a e minal cannula o in ace eb o en icula implan a ion (Pa B; Figu e 1A). Pa A o he de ice was placed subcu aneously on he back o he mouse, as was he cannula, and pa B was implan ed in he la e al en icle (Figu e 1B,C). Due o he weigh (4.29 ± 0.41 g), bu es- pecially he size o he de ice (3.5 × 1 × 1 cm), only male mice we e used in he s udy. In he case o he se en-mon h-old APP/PS1 mice, he weigh o he de ice ep esen s 12.86 ± 0.44% o he mouse’s weigh , while in he case o W mice o he same age his pe cen age is 14.92 ± 0.38%. On he o he hand, he mos limi ing ac o was he leng h o he de ice, which, in ela ion o he leng h o he back o he mice, ep esen s 66.51 ± 1.77% o he leng h in he case o he APP/PS1 mice and 69.56 ± 1.86% in he case o he W mice. Figu e 1. Componen s o he de ice o selec i e and con inuous aphe esis o CSF in mice. ( A ) Pa A consis s o he aphe esis module whe e he nanopo ous memb ane o CSF il a ion is loca ed; coupled o his module is he ese oi whe e he he apeu ic agen is con ained. Bo h he aphe esis module and he ese oi a e p o ec ed by a capsule. Pa B is composed o a cannula connec ed o he aphe esis module and, a he o he end, is coupled o a ca he e ha allows i s implan a ion in he b ain. ( B ) Scheme o he loca ion o pa A on he back o he mouse (made in BioRende .com), as well as i s connec ion h ough pa B o one o he ce eb al hemisphe es. ( C ) The whole sys em is implan ed subcu aneously so ha i emains p o ec ed om he no mal ac i i y o he mouse. In . J. Mol. Sci. 2022,23, 9256 4 o 13 Conside ing his, a o al o wen y- ou male mice we e success ully ope a ed on, which se en een we e APP/PS1 and se en WT (Table 1). Rese oi s we e illed wi h mAb in en mice ( wo W and eigh APP/PS1) and wi h ehicle (a i icial CSF, aCSF) in en mice ( h ee W and se en APP/PS1). Among hem, ou mice we e eu hanized due o se e e le ha gy ha appea ed 48–72 h a e su ge y in h ee cases, while ano he mouse was sac i iced one week la e . In addi ion, one mouse showed signi ican weigh loss and was eu hanized wo weeks a e su ge y. The emaining six een mice showed no appa en complica ions and exhibi ed no mal cage beha io , such as nes o ma ion and g ooming, eaching he end o he h ee-week in usion s udy. Table 1. Summa y o he di e en p ocedu es o echniques ca ied ou on each o he mice used. Mice ID Geno ype Con en In used Time Implan ed Complica ions AβPe meabili y mAb Impe meabili y Aβ Rese oi Aβ Plasma Aβ Co ex IHC IF Fluo escence 7 APP/PS1 mAb 48 h Le ha gy + + 20 W mAb 48 h Le ha gy 21 W aCSF 72 h Le ha gy + 5 APP/PS1 mAb 1 week Le ha gy + 11 APP/PS1 mAb 2 weeks Weigh loss + + + 26 APP/PS1 mAb 2 weeks Le ha gy + 27 APP/PS1 aCSF 2 weeks Le ha gy + 28 APP/PS1 aCSF 2 weeks Le ha gy + 2 APP/PS1 aCSF 3 weeks None + + + 8 APP/PS1 aCSF 3 weeks None + + + + 9 APP/PS1 aCSF 3 weeks None + + + 12 APP/PS1 aCSF 3 weeks None + + + + 4 APP/PS1 aCSF 3 weeks None + + 17 W aCSF 3 weeks None + + 19 W aCSF 3 weeks None + + 3 APP/PS1 mAb 3 weeks None + + + 6 APP/PS1 mAb 3 weeks None + + + 10 APP/PS1 mAb 3 weeks None + + + + 1 APP/PS1 mAb 3 weeks None + + 18 W mAb 3 weeks None + 13 APP/PS1 mAb 488 3 weeks None + + 14 APP/PS1 mAb 488 3 weeks None + + 15 W mAb 488 3 weeks None + + 16 W mAb 488 3 weeks None + + + 29 APP/PS1 naï e - None + 30 APP/PS1 naï e - None + + 31 APP/PS1 naï e - None + + 32 APP/PS1 naï e - None + + 33 APP/PS1 naï e - None + + 34 APP/PS1 naï e - None + 35 APP/PS1 naï e - None + + 36 W naï e - None + 37 APP/PS1 naï e - None + 2.2. The Aphe esis Module Is Pe meable o Aβand Impe meable o An i-AβAn ibody In Vi o The nanopo ous memb anes (NPMBs) used in his s udy we e i s es ed in i o o de e mine hei e icacy in e ms o A β pe meabili y (Supplemen a y Figu e S1A,B). Thei impe meabili y o molecules la ge han he po e size o he NPMBs used he e (9 ± 2 nm) was also es ed wi h albumin (~68 kDa), which is smalle han mAb (~150 kDa), as a e e ence molecule (Supplemen a y Figu e S1C,D). To de e mine whe he he soluble A β pep ide p esen in CSF was able o pass h ough he NPMBs in he aphe esis module in i o , A β le els wi hin he ese oi we e quan i ied wi h SIMOA a e being explan ed om sac i iced mice. One- hi d o he ese oi s analyzed showed he p esence o A β in hei con en (Figu e 2A; Table 1). On he o he hand, he impe meabili y o he sys em o he he apeu ic agen used in his s udy, an i-A β mAb (clone 6E10), was s udied. Fo his pu pose, ou addi ional mice ( wo APP/PS1 and wo WT) we e ea ed wi h he Alexa 488-conjuga ed e sion o he same mAb (Table 1). The le els o luo escence emi ed by he mAb we e hen de e mined bo h in he ese oi and in di e en issues o he mice. The able in Figu e 2B shows luo escence (emission: 600 nm) de e mined in he ese oi , in pe iphe al issue (li e homogena e), and sys emic blood (whole blood and plasma). As a con ol o he luo escence emi ed by he mAb, a 1:200 dilu ion was measu ed. The same dilu ion was In . J. Mol. Sci. 2022,23, 9256 5 o 13 also pe o med on he samples ob ained om he ese oi s. In his way, i was de e mined ha hei luo escence was highe han he emission le els a 600 nm de ec ed in undilu ed pe iphe al samples om he same mice. These samples also had simila emissions o he alues de ec ed in naï e mice, which we e used as a con ol o basal au o luo escence. In . J. Mol. Sci. 2022, 23, x FOR PEER REVIEW 5 o 14 2.2. The Aphe esis Module Is Pe meable o Aβ and Impe meable o An i-Aβ An ibody In Vi o The nanopo ous memb anes (NPMBs) used in his s udy we e i s es ed in i o o de e mine hei e icacy in e ms o Aβ pe meabili y (Supplemen a y Figu e S1A,B). Thei impe meabili y o molecules la ge han he po e size o he NPMBs used he e (9 ± 2 nm) was also es ed wi h albumin (~68 kDa), which is smalle han mAb (~150 kDa), as a e - e ence molecule (Supplemen a y Figu e S1C,D). To de e mine whe he he soluble Aβ pep ide p esen in CSF was able o pass h ough he NPMBs in he aphe esis module in i o, Aβ le els wi hin he ese oi we e quan i ied wi h SIMOA a e being explan ed om sac i iced mice. One- hi d o he ese oi s analyzed showed he p esence o Aβ in hei con en (Figu e 2A; Table 1). Figu e 2. Pe meabili y o Aβ and impe meabili y o mAb analysis a e in i o s udy. (A) G aph shows he le els o Aβ de ec ed (pg/mL) in he ese oi o W and APP/PS1 mice, which could be loaded wi h ehicle (aCSF) o wi h an i-Aβ mAb. Da a a e p esen ed as mean ± SEM. Each do ep- esen s one mouse. (B) Emi ed luo escence da a a 600 nm. Table shows he emission alues a 600 nm de ec ed in liquid samples om ou APP/PS1 mice. These samples we e ob ained om he ese oi , whole blood, plasma, and li e homogena e om wo mice wi h he de ice and ea ed wi h Alexa 488-conjuga ed mAb (clone 6E10) and wo mice ha had no been ope a ed (naï e). Da a o hese mice can be ound in Table 1. The 600 nm emission alues o a 1:200 dilu ion o he mAb, as well as o aCSF used as ehicle and li e homogeniza ion bu e a e also shown. (C) Rep esen a- i e images o ce eb al co ex and li e sec ions om APP/PS1 mice ea ed wi h Alexa 488-conju- ga ed mAb (clone 6E10). F om le o igh , immuno luo escence in consecu i e sec ions was de el- oped using a p ima y an ibody o Aβ plaque de ec ion (clone 6C3), a p ima y an ibody mo e Figu e 2. Pe meabili y o A β and impe meabili y o mAb analysis a e in i o s udy. ( A ) G aph shows he le els o A β de ec ed (pg/mL) in he ese oi o W and APP/PS1 mice, which could be loaded wi h ehicle (aCSF) o wi h an i-A β mAb. Da a a e p esen ed as mean ± SEM. Each do ep esen s one mouse. ( B ) Emi ed luo escence da a a 600 nm. Table shows he emission alues a 600 nm de ec ed in liquid samples om ou APP/PS1 mice. These samples we e ob ained om he ese oi , whole blood, plasma, and li e homogena e om wo mice wi h he de ice and ea ed wi h Alexa 488-conjuga ed mAb (clone 6E10) and wo mice ha had no been ope a ed (naï e). Da a o hese mice can be ound in Table 1. The 600 nm emission alues o a 1:200 dilu ion o he mAb, as well as o aCSF used as ehicle and li e homogeniza ion bu e a e also shown. ( C ) Rep esen a i e images o ce eb al co ex and li e sec ions om APP/PS1 mice ea ed wi h Alexa 488-conjuga ed mAb (clone 6E10). F om le o igh , immuno luo escence in consecu i e sec ions was de eloped using a p ima y an ibody o A β plaque de ec ion (clone 6C3), a p ima y an ibody mo e speci ic o soluble A β (clone 6E10), only Alexa 488-conjuga ed seconda y an ibody, o no an ibody a all. As shown in he images, no signal is de ec ed in he absence o any an ibody. In addi ion, an immuno luo escence s udy was pe o med on c yop ese ed sec ions o bo h he b ain and li e (Figu e 2C). Fi s , we used a p ima y an ibody ha de ec s oligome ic o ms o A β42 (clone 6C3) and a seconda y an ibody conjuga ed wi h Alexa 488. Resul s showed ha , in he ce eb al co ex, high abundance o A β plaques we e de ec ed, while no signal was ound in he li e o hese mice. In consecu i e sec ions, ano he In . J. Mol. Sci. 2022,23, 9256 6 o 13 immuno luo escence was pe o med using he same p ima y an ibody as o ea men (clone 6E10), wi h highe a ini y o soluble and in acellula o ms o A β [ 28 ]. In his case, sligh labelling was obse ed in he ce eb al co ex bu no in he li e . Howe e , incuba ion o he sec ions wi h only labelled seconda y an ibody o no an ibody showed no signal in ei he he ce eb al co ex o in he li e . All in all, hese esul s indica e ha he mAb con ained in he ese oi s is no able o pass h ough he il a ion sys em and he e o e does no appea in b ain issue o pe iphe al issues. 2.3. Con inuous CSF Aphe esis o Rapidly Dec eased A ea Co e ed by A β Plaques in he Ce eb al Co ex Once we de e mined ha A β eaches he ese oi while he mAb emained e ained in he il a ion sys em in i o , we assessed he e ec o he h ee-week in e en ion. To his aim, we quan i ied he a ea occupied by A β plaques in ela ion o he o al assessed a ea o he ce eb al co ex in APP/PS1 mice (Figu e 3). Resul s showed ha con inuous aphe esis o CSF wi h an i-A β mAb signi ican ly educed he a ea occupied by plaques ela i e o aCSF APP/PS1 mice (100 ± 18.80% s. 58.84 ± 7.01%; Figu e 3A,B,D). In some o hese mice, we also de e mined soluble A β42 le els in ce eb al co ex homogena es and in plasma (Figu e 3E,F). Howe e , we obse ed no di e ences be ween mice ea ed wi h aCSF o mAb in ela ion o soluble A β42 le els in APP/PS1 mice o he same age and ha had no unde gone su ge y a his le el. Thus, hese esul s indica ed ha con inuous CSF aphe esis educes he su ace a ea occupied by A β plaques in he ce eb al co ex wi h no changes in b ain o plasma soluble Aβ42 le els a e his in e en ion ime. In . J. Mol. Sci. 2022, 23, x FOR PEER REVIEW 7 o 14 Figu e 3. E alua ion o he e ec o selec i e and con inuous CSF aphe esis in i o. (A–C) Rep e- sen a i e images o immunohis ochemis y de eloped o he quan i ica ion o Aβ plaques in APP/PS1 mice wi h he de ice and ea ed wi h aCSF (A) o mAb (B) and hei compa ison wi h a W mouse (C). Scale ba : 100 µm. (D) Quan i ica ion o he a ea occupied by plaques wi h espec o he o al su ace a ea o he co ex is ep esen ed as he pe cen age o change wi h espec o aCSF- ea ed mice. Da a a e p esen ed as mean ± SEM. Each do ep esen s one mouse. S a is ical com- pa ison be ween he wo g oups was pe o med using unpai ed S uden ’s - es . *, p < 0.05. (E) G aphical ep esen a ion o he quan i ica ion o Aβ le els (pg/mL) in ce eb al co ex homogena es om APP/PS1 mice wi h he de ice, con aining aCSF o mAb, and in age-ma ched APP/PS1 mice ha ha e no unde gone su ge y (naï e). Da a a e p esen ed as mean ± SEM. Each do ep esen s one mouse. (F) G aphical ep esen a ion o he quan i ica ion o Aβ le els (pg/mL) in plasma om APP/PS1 mice implan ed wi h he de ice con aining aCSF o mAb and in in age-ma ched APP/PS1 mice ha ha e no unde gone su ge y (naï e). Da a a e p esen ed as mean ± SEM. Each do ep e- sen s one mouse. 3. Discussion In his wo k, we p esen a new implan able de ice o clea ing selec i ely a ge ed molecules di ec ly om he CSF. The inno a i e mechanism o ac ion o his de ice is based on he p ope y o selec i e molecula pe meabili y o NPMBs. Thus, we bioengi- nee ed a modula de ice whe e he key componen is he aphe esis module endowed wi h ailo ed NPMBs wi h speci ic physicochemical p ope ies. Addi ionally, we in eg a ed he aphe esis module wi h o he componen s o access o he CSF and o in usion o he a- peu ic agen s/aCSF. As he mechanism o ac ion is based on he p ope y o selec i e mo- lecula pe meabili y o NPMBs, we i s es ed he sys em bo h in i o and in i o, demons a ing he e lux o small molecules, such as Aβ, and he impe meabili y o mol- ecules la ge han he memb ane po e size, such as albumin o immunoglobulins. Albei we showed ha memb anes a e pe meable o Aβ, he e we e some inconsis encies among he di e en samples aken om he in i o s udy ha may ela e o di e en a iables, such as he use o he apeu ic agen o aCSF, he espec i e doses, ime o de ice im- plan ed, and sample p ocessing. Impo an ly, we ha e no assessed he le el o bio ouling ha may ha e occu ed in memb anes and po en ially a ec he pe meabili y p ope ies o e ime [29]. Fu he s udies a e needed o be e unde s and bio ouling o memb anes and he ime-dependen molecula dynamics o Aβ and mAb h ough NPMBs in i o. Figu e 3. E alua ion o he e ec o selec i e and con inuous CSF aphe esis in i o . ( A – C ) Rep esen a i e images o immunohis ochemis y de eloped o he quan i ica ion o A β plaques in APP/PS1 mice wi h he de ice and ea ed wi h aCSF ( A ) o mAb ( B ) and hei compa ison wi h a W mouse ( C ). Scale ba : 100 µ m. ( D ) Quan i ica ion o he a ea occupied by plaques wi h espec o he o al su ace a ea o he co ex is ep esen ed as he pe cen age o change wi h espec o aCSF- ea ed mice. Da a a e p esen ed as mean ± SEM. Each do ep esen s one mouse. S a is ical compa ison be ween he wo g oups was pe o med using unpai ed S uden ’s - es . *, p< 0.05. ( E ) G aphical ep esen a ion o he quan i ica ion o A β le els (pg/mL) in ce eb al co ex homogena es om APP/PS1 mice wi h he de ice, con aining aCSF o mAb, and in age-ma ched APP/PS1 mice ha ha e no unde gone su ge y (naï e). Da a a e p esen ed as mean ± SEM. Each do ep esen s one mouse. ( F ) G aphical ep esen a ion o he quan i ica ion o A β le els (pg/mL) in plasma om APP/PS1 mice implan ed wi h he de ice con aining aCSF o mAb and in in age-ma ched APP/PS1 mice ha ha e no unde gone su ge y (naï e). Da a a e p esen ed as mean ±SEM. Each do ep esen s one mouse. In . J. Mol. Sci. 2022,23, 9256 7 o 13 3. Discussion In his wo k, we p esen a new implan able de ice o clea ing selec i ely a ge ed molecules di ec ly om he CSF. The inno a i e mechanism o ac ion o his de ice is based on he p ope y o selec i e molecula pe meabili y o NPMBs. Thus, we bioenginee ed a modula de ice whe e he key componen is he aphe esis module endowed wi h ailo ed NPMBs wi h speci ic physicochemical p ope ies. Addi ionally, we in eg a ed he aphe esis module wi h o he componen s o access o he CSF and o in usion o he apeu ic agen s/aCSF. As he mechanism o ac ion is based on he p ope y o selec i e molecula pe meabili y o NPMBs, we i s es ed he sys em bo h in i o and in i o , demons a ing he e lux o small molecules, such as A β , and he impe meabili y o molecules la ge han he memb ane po e size, such as albumin o immunoglobulins. Albei we showed ha memb anes a e pe meable o A β , he e we e some inconsis encies among he di e en samples aken om he in i o s udy ha may ela e o di e en a iables, such as he use o he apeu ic agen o aCSF, he espec i e doses, ime o de ice implan ed, and sample p ocessing. Impo an ly, we ha e no assessed he le el o bio ouling ha may ha e occu ed in memb anes and po en ially a ec he pe meabili y p ope ies o e ime [ 29 ]. Fu he s udies a e needed o be e unde s and bio ouling o memb anes and he ime- dependen molecula dynamics o Aβand mAb h ough NPMBs in i o. E en when implan a ion o he minia u ized p o o ype in mice is easible and globally sa e, complica ions may p esen in he immedia e and de e ed pos -ope a o y. Some complica ions, such as weigh loss and le ha gy, a e mo e likely o be de i ed om he su gical p ocess i sel o om he side e ec s o pos -su gical medica ion han om he de ice o he he apeu ic agen . Howe e , he size o pa A in he de ice is wi hin he uppe limi o wha is easible o an implan able subcu aneous de ice in mice. The e o e, we will wo k o ine- une he de ice by mainly educing he size o pa A o u u e s udies. In addi ion, we a e cu en ly conduc ing a ho ough e iew o he p o ocols o inc ease sa e y and imp o e he wel a e o he mice. The main inding o his s udy is ha con inuous selec i e aphe esis o soluble A β om he CSF apidly clea s amyloid plaques. This educ ion could be due o he en apmen o Aβin o he aphe esis module, hus gene a ing an e lux om he ISF o he CSF. P e ious s udies ha e al eady epo ed ha CNS-deli e ed immuno he apies wi h an ibodies agains A β , including he one used in his s udy, was e ec i e in educing bo h A β plaques and in acellula A β [ 28 , 30 ]. Those esul s we e ob ained ei he a e an in ahippocampal injec ion in which case he educ ion occu ed on he ipsila e al side [ 30 ] o a e an injec ion in o he hi d en icle wi h educ ion obse ed bila e ally in he hippocampus bu no in he amygdala [ 28 ]. Ne e heless, in bo h cases he e ec s we e obse ed a se en days pos -injec ion and we e los 30 days a e injec ion. In his p oo o concep , he educ ion obse ed in up o h ee weeks o selec i e CSF aphe esis con i ms ha his sys em elici s in he sho ime he same e ec on A β plaques as A β -immuno he apies and o he A β - clea ing me hods. Howe e , heo e ically, he mo e sus ained he A β clea ance e ec , he mo e p o ound and ex ensi e he e e se in he neu opa hological and subsequen clinical changes will be in he long e m. Consequen ly, sys ems o con inuous and sus ained clea ance o a ge molecules in he CNS a e needed, and he de ice he e p esen ed may con ibu e o his aim. One o he main limi a ions o his s udy is ha we ha e no pe o med unc ional s ud- ies o de e mine whe he con inuous and selec i e aphe esis no only educes A β plaques bu also has cogni i e bene i s. Howe e , a educ ion o bo h plaques and in acellula A β has p e iously been epo ed o be associa ed wi h some cogni i e impai men s in mu ine models o AD [ 28 ]. Ne e heless, his is no he case in clinical ials wi h sys emically adminis e ed an i-A β mAb which, while showing e ec i e emo al o plaques, no signi i- can clinical bene i was ound [ 13 , 14 ]. Indeed, plaque o ma ion i sel could be a de ense mechanism [ 4 ], and densely deposi ed A β may no esul as oxic as p e iously hough . Fu he mo e, he mos oxic species o A β a e oligome ic soluble o ms [ 31 , 32 ], which is now pe cei ed as an ad an age o mAb speci ically add essed agains hem [ 13 ]. Al hough In . J. Mol. Sci. 2022,23, 9256 8 o 13 in his s udy we ha e no de ec ed changes in soluble Aβin b ain pa enchyma o plasma, i is possible ha longe aphe esis imes may ha e he capaci y o modi y he le els o hese mo e oxic A β species. The e o e, in u u e long- e m s udies, a mo e comp ehensi e analysis o he di e en compa men s in o which soluble o ms o A β may mo e will be needed o ully unde s and he dynamics o A β be ween plaques, ISF, CSF, and he aphe esis module. Al hough he implan able de ice o CSF con inuous and selec i e aphe esis is an in asi e p ocedu e in small animal models, in i s cu en o m, he esul s ob ained in his p oo -o -concep s udy o e a sho pe iod o ime a e p omising. Cos /bene i implica ions o his he apeu ic s a egy o NND pa ien s may be e y high as his ou e p o ides se e al ad an ages o e “s anda d” pe iphe ally adminis e ed d ugs, including: 1, he sys em allows ac ing on he CSF di ec ly and con inuously, p o iding a mo e powe ul clea ance e ec han he apies based on pe iphe al clea ance; 2, being a selec i e he apy p e en s po en ial sa e y issues ha may occu i he le els o o he molecules in he CSF we e modi ied; 3, immunoisola ion o mAb impedes immune esponses, ully a oiding one o he mos se ious side e ec s epo ed wi h mAb sys emically adminis e ed. All in all, he inno a ion he e p esen ed opens up he doo o a subs an ial imp o emen in he ea men o de as a ing NDD o which he e is cu en ly no cu e. 4. Ma e ials and Me hods 4.1. Mice Se en een se en-mon h-old male APP/PS1 mice, ansgenic o mouse/human amy- loid p ecu so p o ein (Mo/HuAPP695swe) and a mu an human p esenilin 1 (PS1-dE9) in 129S backg ound and se en non- ansgenic li e ma es (W ) we e used o de ice implan- a ion. These mice we e andomly di ided in o wo g oups: con ol ( ehicle, aCSF) and ea men (mAb). Nine addi ional male mice o he same age, bu which had no unde gone su ge y, we e also used (eigh APP/PS1 and one W ; naï e). The dis ibu ion o hese mice in he di e en g oups and hei use o he espec i e p ocedu es can be seen in Table 1. Mice we e main ained wi h ood and wa e ad libi um on a 12 h ligh /da k cycle (onse a 8:00 a.m.) and unde con olled empe a u e (22 ± 2 ◦ C). All p ocedu es we e conduc ed du ing he ligh po ion o he cycle be ween 8:30 and 14:00 a.m. a he Animal Facili y o Uni e sidad de O iedo and we e pe o med in acco dance wi h ins i u ional guidelines app o ed by The Resea ch E hics Commi ee o he Uni e si y o O iedo (PROAE 32/2020). 4.2. Implan able De ice o Mice We ha e designed a p o o ype o an implan able de ice o con inuous and selec i e aphe esis o CSF (Figu e 1A). The de ice was minia u ized o adap i o he size o mice and o manu ac u e in biocompa ible ma e ials. I consis s o wo main componen s, a subcu aneous ese oi and an aphe esis module endowed wi h ailo ed NPMBs [33,34] , which we e encapsula ed oge he in a capsule made by Neu oscience Inno a i e Technolo- gies (Figu e 1A, pa A). Fo his s udy, NPMBs ha e a po e size o 9 ± 2 nm. This allows he passi e low o smalle subs ances, such as soluble A β (~4 kDa in i s monome ic o m), on bo h sides while p e en ing he passage o molecules la ge han 10 nm, e.g., an ibodies (immunoglobulin G, ~150 kDa) o albumin (~68 kDa) ha may ac as he apeu ic agen s (Supplemen a y Ma e ial). These agen s a e e ained in he de ice wi hin he subcu aneous ese oi which has a olume o 100 µ L. A he op o he ese oi is a sel -sealing po , which allows illing using a 50 µ L sy inge (800 se ies mic oli e sy inges; Hamil on, Reno, NV, USA). This design allows d ugs con ined in he ese oi o in e ac wi h he a ge molecule ha is il e ed h ough he NPMBs. The aphe esis module is connec ed h ough a b ain in usion cannula (B ain In usion ki 2 ; Alze , Cupe ino, CA, USA) o one o he la e al en icles o he b ain (Figu e 1A, pa B). In . J. Mol. Sci. 2022,23, 9256 9 o 13 4.3. Filling he Rese oi wi h An i-AβmAb o Vehicle The ese oi was illed unde s e ile condi ions wi h an an i-A β1–42 mAb, which ecognizes human bu no mu ine A β1–42 . To es ima e he mos app op ia e dose o an ibody, we conside ed p e ious s udies in which a single dose o an ibody was injec ed in ace eb o en icula ly, and A β clea ance was e alua ed [ 28 ]. Since hese s udies showed ha he dose applied, 10 µ g, had an e ec up o se en days bu no up o hi y days, we conside ed a dose o 50 µ g o an ibody and a s udy ime o h ee weeks in ou expe imen o assess he changes associa ed wi h con inuous CSF aphe esis in he p esence o his he apeu ic agen . Thus, en de ices (Table 1) we e loaded wi h a solu ion o non-conjuga ed an i- β -Amyloid an ibody (clone 6E10; Biolegend, San Diego, CA, USA) dilu ed in a i icial CSF (aCSF; ACSF, Toc is Bioscience, B is ol, UK). Addi ionally, ano he ou we e illed wi h luo opho e-conjuga ed an i-A β an ibody (Alexa Fluo ® 488 an i- β -Amyloid, clone 6E10; Biolegend). Final mAb concen a ion wi hin he de ice was 500 µ g/mL. As con ol, en de ices we e illed wi h he same olume o aCSF. 4.4. Su ge y and Mice Wel a e The su gical p ocedu e was de eloped acco ding o he p o ocol desc ibed in [ 35 ]. B ie ly, anes he ized mice (2% iso luo ane and 0.4 L/min O 2 ) we e placed in he s e eo axic ame (Digi al Compac Mouse S e eo axic Ins umen ; Ha a d Appa a us, Hollis on, MA, USA), and he su ge y a ea was hen sha ed. To expose he skull, an incision was made om he middle o he skull o he base o he neck om whe e a subcu aneous pocke was opened, and he de ice was inse ed. Finally, he cannula was implan ed in o he le la e al en icle. Su ge y coo dina es, wi h espec o B egma, we e AP—1.26 mm, ML—0.7 mm, and—2.5 mm o he skull. The cannula was glued wi h cyanoac yla e (Cicas ik Su u e ; Chemical Ibe ica, Salamanca, Spain), and he incision was su u ed using 5–0 su u e h ead (S e ile Su gical Su u e; Lo caMa ín, Mu cia, Spain). A e su ge y, all he mice we e placed in an in ensi e ca e cage wi h exhaus i e empe a u e con ol (28–30 ◦ C; Ve a io S50, Kanal e , Bizkaia, Spain) du ing he i s 24 h o ensu e p ope eco e y. All mice ecei ed p e- and pos -su gical analgesic and an ibio ic medica ion; addi ionally, hey we e housed indi idually a e su ge y o p e en damage o he su u es. Te mina ion was scheduled a 21 days. Th oughou he s udy, he mice we e moni o ed daily, obse ing hei weigh , cage beha io (nes o ma ion), ood, and wa e consump ion. 4.5. Tissue and Fluids Collec ion and P ocessing Th ee weeks a e su ge y—o ea lie i se e e complica ions p esen ed—mice we e deeply anaes he ized wi h ke amine (100 mg/kg) and xylazine (10 mg/kg) in saline solu ion. The de ice was ca e ully emo ed and ho oughly cleaned ex e nally wi h 70% e hanol. Subsequen ly, he luid was emo ed om he ese oi and s o ed a − 80 ◦ C o u he s udy. Pe iphe al blood samples we e collec ed om he in e io ena ca a (~0.5–1 mL). Fo his, EDTA (as an icoagulan ; 1.5 mg/mL blood) p e ea ed 1 mL sy inges (BD, F anklin Lakes, NJ, USA) and in a enous ca he e s (BD Insy e-W, 21 GA, BD) we e used. Blood was immedia ely cen i uged a 2000 g, and plasma was s o ed a − 80 ◦ C un il use. A e blood ex ac ion, he abdominal ao a was cu a he same le el o he ena ca a, and s e ile ice-cold phospha e bu e saline (PBS; ~30 mL) was pe used o clean issues. 4.5.1. Tissue P ocessing o Biochemical Analysis In mice used o biochemical analysis, he b ains we e sagi ally dissec ed, and he ce eb al co exes we e ex ac ed om he igh hemisphe es and subsequen ly ozen a − 80 ◦ C. In addi ion, li e po ions (equal o 200 mg) we e emo ed and spli in o wo hal es, one o which was also ozen un il use a −80 ◦C. The ce eb al co exes and li e s (100 mg) we e homogenized in lysis bu e (20 mM HEPES pH 7.4, NaCl 100 mM, NaF 50 mM, EDTA 5 mM, T i on X-100 1%) wi h p o ease inhibi o s (cOmple e ™ , Mini P o ease Inhibi o Cock ail; Roche, Basel, Swi ze land) in