DOI: 10.1161/CIRCULATIONAHA.114.010962
Non-Invasive Imaging of Early Venous Thrombosis by 19F MRI Using Targeted Perfluorocarbon Nanoemulsions Running title: Temme et al.; 19F MRI of deep venous thrombosis Sebastian Temme, PhD1; Christoph Grapentin, PhG2; Christine Quast, MD1,3; Christoph Jacoby, PhD1,3; Maria Grandoch, MD4; Zhaoping Ding, MD1; Christoph Owenier, BSc1; Friederike Mayenfels, PhD2; Jens W. Fischer, PhD4,5; Rolf Schubert, PhD2; Jürgen Schrader, MD1,5*; Ulrich Flögel, PhD1,3,5* 1
Institut für Molekulare Kardiologie, Heinrich-Heine-Universität Düsseldorf Düsseldorf, rff, Dü Düsseldorf, ssel ss sel eldo dorf do rf,, rf 2 Germany; Institut für Pharmazeutische Technologie und Biopharmazie, Albert-LudwigsUniversität, Freiburg, Germany; 3Klinik für Kardiologie, Pneumologie und Angiologie, Univ Un Universitätsklinikum iver iv ersi er sitä t tsskl tä klin i ikum Düsseldorf, Düsseldo in Düsseldorf, orf rf, Ge G Germany; rmany; 4Institutt ffür ü Pharmakologie und ür klinische klin kl inisch ch he Pharmakologie, Phar Ph arma mako kolo ogi gie, H Heinrich-Heine-Universität e nrrichei h He H inee Unnivversit ität ät Düsseldorf, Düsse seeld l orrf, Dü Düssel Düsseldorf, e do dorf, Ge G Germany; rmany; rm y; 5 Cardiovascular Cardio iova vaascu scular Research Res esea eaarc rchh In Inst Institute stit st itut it u e Dü ut Düs Düsseldorf ssel elddorrf (CARID), (CA CARI CA RID) RI D , Heinrich-Heine-Universität D) Heein inri rich ri ch-H ch Hei eine ne-U ne -Uni -U n ve ni vers rssit ität ät Düsseldorf, Düüss ssel elldorf, D Düsseldorf, üsseeld eldorf rff, Ge Germ Germany man ny **Shared *S haare redd senior seeni nioor aauthorship utthoors rshhip hip
Address forr Co Correspondence: Corr rres rr esspo p nd n en encce: Ulrich Flögel, PhD Institut für Molekulare Kardiologie Heinrich-Heine-Universität Düsseldorf Postfach 101007 40001 Düsseldorf, Germany Tel: 49-211-8102933 Fax: 49-211-8102911 E-mail: [email protected] Journal Subject Codes: Diagnostic testing:[30] CT and MRI, Diagnostic testing:[124] Cardiovascular imaging agents/techniques, Thrombosis:[173] Deep vein thrombosis, Thrombosis:[175] Fibrinogen/fibrin
1
DOI: 10.1161/CIRCULATIONAHA.114.010962
Abstract Background—Non-invasive detection of deep venous thrombi and subsequent pulmonary thromboembolism are serious medical challenges and are difficult to detect by conventional ultrasound techniques. Methods and Results—Here we report a novel technique for the sensitive and specific identification of developing thrombi making use of background free 19F MRI together with Į2antiplasmin peptide (Į2AP) targeted perfluorocarbon nanoemulsions (PFCs) as contrast agent, which is cross-linked to fibrin by active factor XIII. Ligand functionality was ensured by mild coupling conditions using sterol-based post-insertion (SPIT). Developing thrombi with a diameter smaller than 0.8 mm could be unequivocally visualized in the murine vena cava f p in vivo by y simultaneous acquisition q g 1H and 19F MR inferior as hot spots of anatomical matching mages at 9.4 T with both excellent signal- and contrast-to-noise ratio (71r22 and ndd 117r5, 7r5, images espectively). Furthermore, Į2AP-PFC could be successfully applied for diagnosis of respectively). expe experimentally p ri rimental ally l iinduced nduc nd u ed pulmonary thromboembo thromboembolism. bolism s . In line with thee reported rep e orted half-life of FX XIIIIa, Ia ap ppl plic icattio ic ionn of o Į -PFCs FCss later laaterr than th han 60 60 min miin after afteer thrombus thro th r mb bus induction ind nduc ucttionn did did not not o any any llonger onge ger FXIIIa, application Į22AP-PFC result detectable esuult in detect ctab ab blee 199F MRI MRI signals. sign si gnal alss. al s. Corresponding Corrrespponding ng results res essults ults were wer eree obtained obta tain in ned in in ex ex vvivo ivoo ggenerated iv eneerat erated ed human huuma mann clots. c ot cl ots. s. T Thus, huus, Į2 Į2APP-P -PFCs PFC Cs ca cann vi vvisualize sualiz alizee freshly freshhlyy de fres developed evelo velo opeed tthrombi hromb mbi whic w which hic i h mi migh might ghtt st gh still tilll be susceptible usc scep eptible le to ppharmacologic h rm ha rmac a ollog ogicc iintervention. nterrventio nt on. n Conclusion Conclusions—Our ns—Ou Ourr re Ou rresults esu suult ltss de demo demonstrate ons n tr traatee th that att 1H/19F MRI M I together MR toge toge geth th her with wit ithh Į Į22AP-PFC -PFCs PFC F s iss a sensitive non-invasive technique for the diagnosis of acute deep venous thrombi and pulmonary thromboemboli. Furthermore, ligand coupling by SPIT represents a unique platform for specific targeting of PFCs for in vivo 19F MRI.
Key words: magnetic resonance imaging, pulmonary embolism, deep vein thrombosis, perfluorocarbon, molecular imaging
2
DOI: 10.1161/CIRCULATIONAHA.114.010962
Thrombosis plays a crucial role in a variety of cardiovascular diseases like myocardial infarction, deep venous thrombosis, and pulmonary embolism which are a major cause of morbidity and mortality. More recently, it has been shown that tumor patients have an up 7-fold and obese people a 2.5-fold increased risk to develop thromboembolic events1,2 caused by systemically increased pro-inflammatory conditions and the release of pro-thrombotic factors. Thus, visualization and specific identification of thrombi by imaging techniques addresses an important clinical problem. Thrombi can be non-invasively visualized by ultrasound3, computer tomography (CT)4 or magnetic resonance imaging (MRI)5. MRI is free of ionizing radiation and has high spatial resolution esolution also in deep tissues which are not accessible by ultrasound. However,, thrombus thrrom mbu buss detection by conventional MRI using 1H MR angiography or T1/T2-weighted 1H MRI is difficult incce small smal sm alll non-occlusive al non--oc no occl c usive thrombi have only minor miino n r impact on blo ood o fflow low lo w and may not give rise since blood too a clear clear signal signa nall in weighted wei e ghhte tedd images. imag im ag ges es. To To overcome oveerccomee this thiss limitation limiitaati tion onn Gadolinium-based Gad dol o iniu umm ba bassedd probes prrobbes -11 raised ais ised ed against aga g in nst fibrin fib briin within with th hin the the h thrombus thro rom mbus have mbus hav avee been b en developed be deevel evel elop oped op ed d (EP-2104R) (EP EP-2 -210 -2 1004R)66-11 . An Another nothe ot er sp sspecific ecciffic
deeve velo l pi lo ping ng thrombi thr h om mbi b is is factor f ct fa ctor orr XIIIa XII I Ia (FXIIIa) (FX FXII IIIIa Ia)) wh whic ichh cr ic cros o s--li os l nk nkss Į 2 an 2ntiipl plas a mi as m n with marker of developing which cross-links Į2-antiplasmin fibrin during the early phase of thrombus formation12,13. Probes based on Į2-antiplasmin have been used for ex vivo or in vivo labelling of thrombi by near-infrared fluorescence (NIRF), scintigraphy or gadolinium (Gd) enhanced 1H MRI14-17. Recently 19F MRI emerged as a novel promising technique for molecular imaging. For this, emulsified biologically inert perfluorocarbons (perfluorocarbon nanoemulsions = PFCs) are used as contrast agent to follow the fate of ex vivo or in vivo PFC-labelled cells18-20. Since 19F is physiologically found in biological tissue only in trace amounts, the resulting fluorine signal displays an excellent degree of specificity. Merging of 19F images with corresponding 1H data
3
DOI: 10.1161/CIRCULATIONAHA.114.010962
sets enables the exact anatomical localization of the 19F signal. Since 19F MRI generates a positive contrast as “hot spot”, it is especially useful for heterogeneous tissue where 1H-based contrast is challenged by susceptibility artefacts or sparse proton density complicating the interpretation of effects obtained by other contrast agents. Up to now, 19F MRI has been used predominantly for immune cell tracking in a variety of clinically relevant inflammation models2129
. This approach is based on the rapid uptake of i.v. injected PFCs by circulating monocytes that
subsequently migrate into the inflamed area resulting in a local accumulation of 19F-loaded immune cells. In the present study, we report a novel procedure for the generation of targeted PFCs, which makes use of a sterol-based post-insertion technique (SPIT) to generate Į 2-an 2-an ntiipl plas asmi as minn mi Į2-antiplasmin abelled PFCs (Į2AP-PFC). SPIT allows modification of preformed PFCs under mild conditions labelled hatt m aint ai ntai nt ainn thee fu ai fun nctionality of labile ligands. U sing Į2AP-PFC we we were werre we re able to detect the that maintain functionality Using formation form mation of developing dev vellop pin i g deep deep venous ven nou ouss thrombi thrrom th mbi andd ppulmonary mbi ulm mon onaaryy em embolism mboli bolism sm m by by 1H/ H/199F MRI MR iinn vivo v vo vi vo with wiith high hig ighh specificity speecif sp ecifiiciity ity aand ndd ssensitivity. ens nssittiv ivit ity. it y.
Methods An expanded methods section can be found in the online-only data supplement. Perfluorocarbon nanoemulsions Perfluorocarbon nanoemulsions (PFCs) were prepared as previously reported (see data supplement for more details)22,23. Sterol-based post-insertion technique (SPIT) Generation of the cholesterol-PEG2000-maleimide anchor An equimolar mixture of maleimide-PEG2000-NH2 (Jenkem Technology, Plano, TX), cholesteryl chloroformate (Sigma Aldrich, Seelze, Germany) and the activator triethylamine (Carl Roth, 4
DOI: 10.1161/CIRCULATIONAHA.114.010962
Karlsruhe, Germany) in water-free methylene chloride was prepared. The mixture was stirred for 24 h under exclusion of light in a nitrogen atmosphere. The resulting cholesterol-PEG2000maleimide was purified by chromatography using a sephadex LH20 column and validated by 1H NMR spectroscopy. Aliquots were stored at -80 °C under argon. Coupling of Į2-antiplasmin peptide to the cholesterol-PEG anchor To generate PFCs for site specific targeting of thrombi, we used a 14 amino acid peptide derived from Į2-antiplasmin (Į2AP) which is known to be cross-linked to fibrin at the glutamine Q3 by FXIIIa15-17. As control, Q3 was converted to alanine (Q3A) leading to a low affinity substrate for FXIIIa17. Both peptides were further functionalized with a cysteine residue at amino acid position 13 (supplemental Fig. S1) for coupling to the cholesterol-PEG anchor.. Fo Forr im immunommu muno no-no fluorescence studies, carboxyfluorescein (CF) was linked via an additional lysine at the cterminal erm min inal al ttryptophan rypt ry ptoppha pt hann (W14). peptides Germany) All pept ptid idees ((Genaxxon, id G na Ge naxx xxon xx on,, Ulm, on Ulm, G Ulm erm manny) we were dissolved disso issoolv ved inn sterile ster st errilee phosphate phhosph osphhat a e buffer buff bu fffer er (10 (100 with glycerol, mg/ml. peptides were added mM M phosphate, pho h sp spha hate ha te,, iisotonized sooton otonnizzed w i h gl it glyc ycer yc errol ol,, pH 77.4) . ) to .4 o5m g/m g/ ml. Th ml. Thee pe pept ptid pt des w eree ad er adde dedd to to tthe he cholesterol-PEG cholesterol-PE PE EG200 -maleimide mal a eiimi mide d ((Chol-PEG de Chol Ch o -P PEG200 -Mal) Mal a ) anchor anch an chor ch or aand nd tthe hee m mixture ix xtu ture re w was as sshaken hake ha k n at 17 °C C 2000 000 0-m 2000 0-M for 20 h at 700 rpm. The cholesterol-anchor was used in 10-fold excess compared to the ligand, thus allowing a quantitative coupling of peptides. During the incubation period the thiol-group of the cysteine residue and the maleimide group form a stable thioether bond30. Free maleimide groups were subsequently deactivated by addition of mercaptoethanolamine. By this way, a mixture of cholesterol-PEG2000-peptide and deactivated cholesterol-PEG2000 (approximate ratio of 1:10) was obtained and used in the following insertion step. Post-insertion Preformed PFCs were incubated with the obtained mixture on a rotary shaker at 17 °C for one
5
DOI: 10.1161/CIRCULATIONAHA.114.010962
hour. As illustrated in Fig. 1A, this leads to the spontaneous insertion of the cholesterol moiety into the phospholipid layer of the PFC31 (molar ratio phospholipid to cholesterol derivative of 20:1). PEGylated but non-targeted PFCs were formed by incubation with non-modified cholesterol-PEG2000 only. Characterization of PFCs The resulting emulsions were characterized by photon correlation spectroscopy on a Malvern Zetasizer Nano ZS (Malvern Instruments, Herrenberg, Germany) to determine the hydrodynamic diameter, the polydispersity index (PDI) and the ȗ-potential. In comparison to non-modified PFCs, we observed a slight increase in size (diameter: non-targeted PFCs = 149r15 nm; Į2APPFCs = 165r13 nm), a similar size distribution (PDI: non-targeted PFCs = 0.14r 0.14r0.01; r0. 0 01 0 ; Į2 Į2AP-P -PFCs -PFC FC = 0.16r0.05), but a less negative ȗ-potential (non-targeted PFCs = -37.2r4 mV; Į2AP-PFCs = 11.7r7 11 1.7 .7r 7r7 mV) mV V) for for th thee targ targeted geted or PEGylated PFCs PFCs. Cs. This This indicates tthe h suc he successful ucce cessful incorp incorporation poration of chol-PEG-Į2 ch cho hol-PEG-Į2AP iinto ntto the th he PFC PF FC nanoparticles. na opa nano part rtic iclles. 19F MRI MR me meas measurements asuureemeentts con as confirmed nfirm med tha that hatt all all nanoemulsions nano na nooem emul ulsi ul sion si onss ex exhi exhibited hibi hi bite bi tedd th te the sa same me ffluorine luor lu orrin inee co con content nten nten entt ((supplemental supp su pple pp leme le ment me nttal Fig. Fig ig. S2 S2B S2B). B). ). Animal experiments exp xper erim er imen im e ts en ts Animal experiments were in accordance with institutional guidelines on animal care. Male mice (C57BL/6; 25-30 g body weight; 8-10 weeks of age) used in this study were bred at the central animal facility of the Heinrich Heine University (Düsseldorf, Germany) and fed with a standard chow diet and received tap water ad libidum. Induction of venous thrombi and pulmonary thromboembolism During surgery, mice were kept under anaesthesia with 1.5% isoflurane. Buprenorphin was injected for analgesia. A median laparotomy was performed, and the inferior vena cava was exposed at the anatomical level of both kidneys. Subsequently, a filter paper (1u4 mm2) soaked
6
DOI: 10.1161/CIRCULATIONAHA.114.010962
with 10% FeCl3 was placed on the top of the vessel and incubated for 8 min. To assure the location of the FeCl3-soaked filter paper on top of the vessel surface, two stretches of parafilm were placed on both sides of the vessel. After removal of the filter paper the vessel was washed with 0.9% NaCl to remove residual FeCl3. PFCs (3 mmol/kg body weight) were injected into the tail vein approximately 5 min prior to thrombus induction or 5, 15, 30, 60 or 90 min post thrombus induction. Subsequently, MRI scans were performed at 2, 8 or 24 h after surgery. To induce pulmonary thromboembolism a mixture of human thrombin (Sigma-Aldrich, Seelze, Germany, 10U/25 g body weight) and Į2AP-PFCs (or unmodified PFCs/Q3A-PFCs as control) was injected which resulted in an 80% survival rate. 1H/19F MRI measurements were performed 24 h later. MRI studies Experiments Ex xpe peri rime ri ment me ntss we nt were re pperformed erformed at a vertical 9.4 T B Bruker ru uker AVANCE EIII W Wide idee Bore NMR id spectrometer pecctr t ometer (Bruker, (Br Brukker er,, Rheinstetten, Rhei Rh eins ei nste tett te tten en,, Germany) en German ny) operating opperrating ng at at frequencies frreq equuenc uencie iess of of 400.21 400 00.2 .211 MHz MHz for for 1H an and nd 37 376. 376.54 6.54 544 MHz MHzz for for o 1199F me m measurements asur as urrem emeents entss uusing sinng si ng m microimaging iccro oim mag agin ingg un in unit units itts as ddescribed esccrib crib bed ppreviously reevi viouuslyy2225,27,32 5,27 27,32 27 32
. Mice w were eree an er anae anaesthetized aeest s heeti t ze z dw with itth 1. 1.5% .5% iisoflurane soofl flur u an ur anee an and nd we were re ke kkept pt aatt 37 °C C du duri during ring ri ng tthe h he
measurements. Data were acquired using a 25-mm birdcage resonator tuneable to 1H and 19F. After acquisition of the morphological 1H images, the resonator was tuned to 19F and anatomically matching 19F images were recorded (see data supplement for a more detailed description of MRI setup, acquisition parameters, and quantification procedures). An overview over all imaging parameters used for 1H/19F MRI is given in supplemental table 1. In vitro thrombus studies Human blood was obtained by venous puncture and collected on ice. 100 μl of blood were transferred to a round-bottom 96-well plate and incubated at 37 °C for 15 min. Next, 25 μl of
7
DOI: 10.1161/CIRCULATIONAHA.114.010962
PFCs (non-targeted PFCs or Į2AP-PFCs) were added to each well and the plate was further incubated for 90 min at 37 °C under constant motion. The blood clots were extensively washed with cold PBS and subjected to 1H/19F MRI. Details about PFC uptake studies by murine blood are given in the expanded method section of the supplemental material. Cytotoxicity assay Murine splenocytes were obtained from C56BL/6 mice, incubated with the different PFCs and analysed as described in the expanded method section of the supplemental material. Flow cytometry Mice were killed by cervical dislocation, thrombi were excised and digested in streptokinase/plasmin treptokinase/plasmin (150/2 U/ml) for 30 min att 37 °C under constant shaking. To To generate geene nera rate ra te a single ingle cell suspension, thrombi were passed through a 70 μm cell strainer (BD Biosciences, Heidelberg, Heid He idel id elbe el berg be rg,, Ge rg G Germany), rm many), an washed with FACS buff buffer fer e and and stained forr 30 30 min min with fluorochromecoupled co oupled up antibodies antibo odi diess as as described deesccribe ribe bedd inn the the h ddata ata sup ssupplement. uppleement ntt. Histology Hiist stol olog ol ogyy and og and fluorescence flluooreesc scen e cee microscopy en mic icro rosc ro scop opyy op Excised thro thrombi omb mbii an andd lu lung lungs n s we ng weree ffixed ixed ix e iin ed n 4% P PFA FA oorr em embedded mbe bedd dded dd ed iin n Ti T Tissue-Tek ssue ss ue-T ue -T Tek k ((Weckert Weck We cker ck et Labortechnik, Kitzingen, Germany) and frozen at -20 °C. Sections of 8 μm or 14 μm were cut and processed for immunohistochemical staining as described previously23,24,32 (see data supplement for more details). Statistics All data were evaluated for normal distribution using the Shapiro-Wilk test and are given as mean values r standard deviations (SD). Statistical difference was assessed by Welch´s test (for unequal variances) and a level of p
Non-Invasive Imaging of Early Venous Thrombosis by 19F MRI Using Targeted Perfluorocarbon Nanoemulsions Running title: Temme et al.; 19F MRI of deep venous thrombosis Sebastian Temme, PhD1; Christoph Grapentin, PhG2; Christine Quast, MD1,3; Christoph Jacoby, PhD1,3; Maria Grandoch, MD4; Zhaoping Ding, MD1; Christoph Owenier, BSc1; Friederike Mayenfels, PhD2; Jens W. Fischer, PhD4,5; Rolf Schubert, PhD2; Jürgen Schrader, MD1,5*; Ulrich Flögel, PhD1,3,5* 1
Institut für Molekulare Kardiologie, Heinrich-Heine-Universität Düsseldorf Düsseldorf, rff, Dü Düsseldorf, ssel ss sel eldo dorf do rf,, rf 2 Germany; Institut für Pharmazeutische Technologie und Biopharmazie, Albert-LudwigsUniversität, Freiburg, Germany; 3Klinik für Kardiologie, Pneumologie und Angiologie, Univ Un Universitätsklinikum iver iv ersi er sitä t tsskl tä klin i ikum Düsseldorf, Düsseldo in Düsseldorf, orf rf, Ge G Germany; rmany; 4Institutt ffür ü Pharmakologie und ür klinische klin kl inisch ch he Pharmakologie, Phar Ph arma mako kolo ogi gie, H Heinrich-Heine-Universität e nrrichei h He H inee Unnivversit ität ät Düsseldorf, Düsse seeld l orrf, Dü Düssel Düsseldorf, e do dorf, Ge G Germany; rmany; rm y; 5 Cardiovascular Cardio iova vaascu scular Research Res esea eaarc rchh In Inst Institute stit st itut it u e Dü ut Düs Düsseldorf ssel elddorrf (CARID), (CA CARI CA RID) RI D , Heinrich-Heine-Universität D) Heein inri rich ri ch-H ch Hei eine ne-U ne -Uni -U n ve ni vers rssit ität ät Düsseldorf, Düüss ssel elldorf, D Düsseldorf, üsseeld eldorf rff, Ge Germ Germany man ny **Shared *S haare redd senior seeni nioor aauthorship utthoors rshhip hip
Address forr Co Correspondence: Corr rres rr esspo p nd n en encce: Ulrich Flögel, PhD Institut für Molekulare Kardiologie Heinrich-Heine-Universität Düsseldorf Postfach 101007 40001 Düsseldorf, Germany Tel: 49-211-8102933 Fax: 49-211-8102911 E-mail: [email protected] Journal Subject Codes: Diagnostic testing:[30] CT and MRI, Diagnostic testing:[124] Cardiovascular imaging agents/techniques, Thrombosis:[173] Deep vein thrombosis, Thrombosis:[175] Fibrinogen/fibrin
1
DOI: 10.1161/CIRCULATIONAHA.114.010962
Abstract Background—Non-invasive detection of deep venous thrombi and subsequent pulmonary thromboembolism are serious medical challenges and are difficult to detect by conventional ultrasound techniques. Methods and Results—Here we report a novel technique for the sensitive and specific identification of developing thrombi making use of background free 19F MRI together with Į2antiplasmin peptide (Į2AP) targeted perfluorocarbon nanoemulsions (PFCs) as contrast agent, which is cross-linked to fibrin by active factor XIII. Ligand functionality was ensured by mild coupling conditions using sterol-based post-insertion (SPIT). Developing thrombi with a diameter smaller than 0.8 mm could be unequivocally visualized in the murine vena cava f p in vivo by y simultaneous acquisition q g 1H and 19F MR inferior as hot spots of anatomical matching mages at 9.4 T with both excellent signal- and contrast-to-noise ratio (71r22 and ndd 117r5, 7r5, images espectively). Furthermore, Į2AP-PFC could be successfully applied for diagnosis of respectively). expe experimentally p ri rimental ally l iinduced nduc nd u ed pulmonary thromboembo thromboembolism. bolism s . In line with thee reported rep e orted half-life of FX XIIIIa, Ia ap ppl plic icattio ic ionn of o Į -PFCs FCss later laaterr than th han 60 60 min miin after afteer thrombus thro th r mb bus induction ind nduc ucttionn did did not not o any any llonger onge ger FXIIIa, application Į22AP-PFC result detectable esuult in detect ctab ab blee 199F MRI MRI signals. sign si gnal alss. al s. Corresponding Corrrespponding ng results res essults ults were wer eree obtained obta tain in ned in in ex ex vvivo ivoo ggenerated iv eneerat erated ed human huuma mann clots. c ot cl ots. s. T Thus, huus, Į2 Į2APP-P -PFCs PFC Cs ca cann vi vvisualize sualiz alizee freshly freshhlyy de fres developed evelo velo opeed tthrombi hromb mbi whic w which hic i h mi migh might ghtt st gh still tilll be susceptible usc scep eptible le to ppharmacologic h rm ha rmac a ollog ogicc iintervention. nterrventio nt on. n Conclusion Conclusions—Our ns—Ou Ourr re Ou rresults esu suult ltss de demo demonstrate ons n tr traatee th that att 1H/19F MRI M I together MR toge toge geth th her with wit ithh Į Į22AP-PFC -PFCs PFC F s iss a sensitive non-invasive technique for the diagnosis of acute deep venous thrombi and pulmonary thromboemboli. Furthermore, ligand coupling by SPIT represents a unique platform for specific targeting of PFCs for in vivo 19F MRI.
Key words: magnetic resonance imaging, pulmonary embolism, deep vein thrombosis, perfluorocarbon, molecular imaging
2
DOI: 10.1161/CIRCULATIONAHA.114.010962
Thrombosis plays a crucial role in a variety of cardiovascular diseases like myocardial infarction, deep venous thrombosis, and pulmonary embolism which are a major cause of morbidity and mortality. More recently, it has been shown that tumor patients have an up 7-fold and obese people a 2.5-fold increased risk to develop thromboembolic events1,2 caused by systemically increased pro-inflammatory conditions and the release of pro-thrombotic factors. Thus, visualization and specific identification of thrombi by imaging techniques addresses an important clinical problem. Thrombi can be non-invasively visualized by ultrasound3, computer tomography (CT)4 or magnetic resonance imaging (MRI)5. MRI is free of ionizing radiation and has high spatial resolution esolution also in deep tissues which are not accessible by ultrasound. However,, thrombus thrrom mbu buss detection by conventional MRI using 1H MR angiography or T1/T2-weighted 1H MRI is difficult incce small smal sm alll non-occlusive al non--oc no occl c usive thrombi have only minor miino n r impact on blo ood o fflow low lo w and may not give rise since blood too a clear clear signal signa nall in weighted wei e ghhte tedd images. imag im ag ges es. To To overcome oveerccomee this thiss limitation limiitaati tion onn Gadolinium-based Gad dol o iniu umm ba bassedd probes prrobbes -11 raised ais ised ed against aga g in nst fibrin fib briin within with th hin the the h thrombus thro rom mbus have mbus hav avee been b en developed be deevel evel elop oped op ed d (EP-2104R) (EP EP-2 -210 -2 1004R)66-11 . An Another nothe ot er sp sspecific ecciffic
deeve velo l pi lo ping ng thrombi thr h om mbi b is is factor f ct fa ctor orr XIIIa XII I Ia (FXIIIa) (FX FXII IIIIa Ia)) wh whic ichh cr ic cros o s--li os l nk nkss Į 2 an 2ntiipl plas a mi as m n with marker of developing which cross-links Į2-antiplasmin fibrin during the early phase of thrombus formation12,13. Probes based on Į2-antiplasmin have been used for ex vivo or in vivo labelling of thrombi by near-infrared fluorescence (NIRF), scintigraphy or gadolinium (Gd) enhanced 1H MRI14-17. Recently 19F MRI emerged as a novel promising technique for molecular imaging. For this, emulsified biologically inert perfluorocarbons (perfluorocarbon nanoemulsions = PFCs) are used as contrast agent to follow the fate of ex vivo or in vivo PFC-labelled cells18-20. Since 19F is physiologically found in biological tissue only in trace amounts, the resulting fluorine signal displays an excellent degree of specificity. Merging of 19F images with corresponding 1H data
3
DOI: 10.1161/CIRCULATIONAHA.114.010962
sets enables the exact anatomical localization of the 19F signal. Since 19F MRI generates a positive contrast as “hot spot”, it is especially useful for heterogeneous tissue where 1H-based contrast is challenged by susceptibility artefacts or sparse proton density complicating the interpretation of effects obtained by other contrast agents. Up to now, 19F MRI has been used predominantly for immune cell tracking in a variety of clinically relevant inflammation models2129
. This approach is based on the rapid uptake of i.v. injected PFCs by circulating monocytes that
subsequently migrate into the inflamed area resulting in a local accumulation of 19F-loaded immune cells. In the present study, we report a novel procedure for the generation of targeted PFCs, which makes use of a sterol-based post-insertion technique (SPIT) to generate Į 2-an 2-an ntiipl plas asmi as minn mi Į2-antiplasmin abelled PFCs (Į2AP-PFC). SPIT allows modification of preformed PFCs under mild conditions labelled hatt m aint ai ntai nt ainn thee fu ai fun nctionality of labile ligands. U sing Į2AP-PFC we we were werre we re able to detect the that maintain functionality Using formation form mation of developing dev vellop pin i g deep deep venous ven nou ouss thrombi thrrom th mbi andd ppulmonary mbi ulm mon onaaryy em embolism mboli bolism sm m by by 1H/ H/199F MRI MR iinn vivo v vo vi vo with wiith high hig ighh specificity speecif sp ecifiiciity ity aand ndd ssensitivity. ens nssittiv ivit ity. it y.
Methods An expanded methods section can be found in the online-only data supplement. Perfluorocarbon nanoemulsions Perfluorocarbon nanoemulsions (PFCs) were prepared as previously reported (see data supplement for more details)22,23. Sterol-based post-insertion technique (SPIT) Generation of the cholesterol-PEG2000-maleimide anchor An equimolar mixture of maleimide-PEG2000-NH2 (Jenkem Technology, Plano, TX), cholesteryl chloroformate (Sigma Aldrich, Seelze, Germany) and the activator triethylamine (Carl Roth, 4
DOI: 10.1161/CIRCULATIONAHA.114.010962
Karlsruhe, Germany) in water-free methylene chloride was prepared. The mixture was stirred for 24 h under exclusion of light in a nitrogen atmosphere. The resulting cholesterol-PEG2000maleimide was purified by chromatography using a sephadex LH20 column and validated by 1H NMR spectroscopy. Aliquots were stored at -80 °C under argon. Coupling of Į2-antiplasmin peptide to the cholesterol-PEG anchor To generate PFCs for site specific targeting of thrombi, we used a 14 amino acid peptide derived from Į2-antiplasmin (Į2AP) which is known to be cross-linked to fibrin at the glutamine Q3 by FXIIIa15-17. As control, Q3 was converted to alanine (Q3A) leading to a low affinity substrate for FXIIIa17. Both peptides were further functionalized with a cysteine residue at amino acid position 13 (supplemental Fig. S1) for coupling to the cholesterol-PEG anchor.. Fo Forr im immunommu muno no-no fluorescence studies, carboxyfluorescein (CF) was linked via an additional lysine at the cterminal erm min inal al ttryptophan rypt ry ptoppha pt hann (W14). peptides Germany) All pept ptid idees ((Genaxxon, id G na Ge naxx xxon xx on,, Ulm, on Ulm, G Ulm erm manny) we were dissolved disso issoolv ved inn sterile ster st errilee phosphate phhosph osphhat a e buffer buff bu fffer er (10 (100 with glycerol, mg/ml. peptides were added mM M phosphate, pho h sp spha hate ha te,, iisotonized sooton otonnizzed w i h gl it glyc ycer yc errol ol,, pH 77.4) . ) to .4 o5m g/m g/ ml. Th ml. Thee pe pept ptid pt des w eree ad er adde dedd to to tthe he cholesterol-PEG cholesterol-PE PE EG200 -maleimide mal a eiimi mide d ((Chol-PEG de Chol Ch o -P PEG200 -Mal) Mal a ) anchor anch an chor ch or aand nd tthe hee m mixture ix xtu ture re w was as sshaken hake ha k n at 17 °C C 2000 000 0-m 2000 0-M for 20 h at 700 rpm. The cholesterol-anchor was used in 10-fold excess compared to the ligand, thus allowing a quantitative coupling of peptides. During the incubation period the thiol-group of the cysteine residue and the maleimide group form a stable thioether bond30. Free maleimide groups were subsequently deactivated by addition of mercaptoethanolamine. By this way, a mixture of cholesterol-PEG2000-peptide and deactivated cholesterol-PEG2000 (approximate ratio of 1:10) was obtained and used in the following insertion step. Post-insertion Preformed PFCs were incubated with the obtained mixture on a rotary shaker at 17 °C for one
5
DOI: 10.1161/CIRCULATIONAHA.114.010962
hour. As illustrated in Fig. 1A, this leads to the spontaneous insertion of the cholesterol moiety into the phospholipid layer of the PFC31 (molar ratio phospholipid to cholesterol derivative of 20:1). PEGylated but non-targeted PFCs were formed by incubation with non-modified cholesterol-PEG2000 only. Characterization of PFCs The resulting emulsions were characterized by photon correlation spectroscopy on a Malvern Zetasizer Nano ZS (Malvern Instruments, Herrenberg, Germany) to determine the hydrodynamic diameter, the polydispersity index (PDI) and the ȗ-potential. In comparison to non-modified PFCs, we observed a slight increase in size (diameter: non-targeted PFCs = 149r15 nm; Į2APPFCs = 165r13 nm), a similar size distribution (PDI: non-targeted PFCs = 0.14r 0.14r0.01; r0. 0 01 0 ; Į2 Į2AP-P -PFCs -PFC FC = 0.16r0.05), but a less negative ȗ-potential (non-targeted PFCs = -37.2r4 mV; Į2AP-PFCs = 11.7r7 11 1.7 .7r 7r7 mV) mV V) for for th thee targ targeted geted or PEGylated PFCs PFCs. Cs. This This indicates tthe h suc he successful ucce cessful incorp incorporation poration of chol-PEG-Į2 ch cho hol-PEG-Į2AP iinto ntto the th he PFC PF FC nanoparticles. na opa nano part rtic iclles. 19F MRI MR me meas measurements asuureemeentts con as confirmed nfirm med tha that hatt all all nanoemulsions nano na nooem emul ulsi ul sion si onss ex exhi exhibited hibi hi bite bi tedd th te the sa same me ffluorine luor lu orrin inee co con content nten nten entt ((supplemental supp su pple pp leme le ment me nttal Fig. Fig ig. S2 S2B S2B). B). ). Animal experiments exp xper erim er imen im e ts en ts Animal experiments were in accordance with institutional guidelines on animal care. Male mice (C57BL/6; 25-30 g body weight; 8-10 weeks of age) used in this study were bred at the central animal facility of the Heinrich Heine University (Düsseldorf, Germany) and fed with a standard chow diet and received tap water ad libidum. Induction of venous thrombi and pulmonary thromboembolism During surgery, mice were kept under anaesthesia with 1.5% isoflurane. Buprenorphin was injected for analgesia. A median laparotomy was performed, and the inferior vena cava was exposed at the anatomical level of both kidneys. Subsequently, a filter paper (1u4 mm2) soaked
6
DOI: 10.1161/CIRCULATIONAHA.114.010962
with 10% FeCl3 was placed on the top of the vessel and incubated for 8 min. To assure the location of the FeCl3-soaked filter paper on top of the vessel surface, two stretches of parafilm were placed on both sides of the vessel. After removal of the filter paper the vessel was washed with 0.9% NaCl to remove residual FeCl3. PFCs (3 mmol/kg body weight) were injected into the tail vein approximately 5 min prior to thrombus induction or 5, 15, 30, 60 or 90 min post thrombus induction. Subsequently, MRI scans were performed at 2, 8 or 24 h after surgery. To induce pulmonary thromboembolism a mixture of human thrombin (Sigma-Aldrich, Seelze, Germany, 10U/25 g body weight) and Į2AP-PFCs (or unmodified PFCs/Q3A-PFCs as control) was injected which resulted in an 80% survival rate. 1H/19F MRI measurements were performed 24 h later. MRI studies Experiments Ex xpe peri rime ri ment me ntss we nt were re pperformed erformed at a vertical 9.4 T B Bruker ru uker AVANCE EIII W Wide idee Bore NMR id spectrometer pecctr t ometer (Bruker, (Br Brukker er,, Rheinstetten, Rhei Rh eins ei nste tett te tten en,, Germany) en German ny) operating opperrating ng at at frequencies frreq equuenc uencie iess of of 400.21 400 00.2 .211 MHz MHz for for 1H an and nd 37 376. 376.54 6.54 544 MHz MHzz for for o 1199F me m measurements asur as urrem emeents entss uusing sinng si ng m microimaging iccro oim mag agin ingg un in unit units itts as ddescribed esccrib crib bed ppreviously reevi viouuslyy2225,27,32 5,27 27,32 27 32
. Mice w were eree an er anae anaesthetized aeest s heeti t ze z dw with itth 1. 1.5% .5% iisoflurane soofl flur u an ur anee an and nd we were re ke kkept pt aatt 37 °C C du duri during ring ri ng tthe h he
measurements. Data were acquired using a 25-mm birdcage resonator tuneable to 1H and 19F. After acquisition of the morphological 1H images, the resonator was tuned to 19F and anatomically matching 19F images were recorded (see data supplement for a more detailed description of MRI setup, acquisition parameters, and quantification procedures). An overview over all imaging parameters used for 1H/19F MRI is given in supplemental table 1. In vitro thrombus studies Human blood was obtained by venous puncture and collected on ice. 100 μl of blood were transferred to a round-bottom 96-well plate and incubated at 37 °C for 15 min. Next, 25 μl of
7
DOI: 10.1161/CIRCULATIONAHA.114.010962
PFCs (non-targeted PFCs or Į2AP-PFCs) were added to each well and the plate was further incubated for 90 min at 37 °C under constant motion. The blood clots were extensively washed with cold PBS and subjected to 1H/19F MRI. Details about PFC uptake studies by murine blood are given in the expanded method section of the supplemental material. Cytotoxicity assay Murine splenocytes were obtained from C56BL/6 mice, incubated with the different PFCs and analysed as described in the expanded method section of the supplemental material. Flow cytometry Mice were killed by cervical dislocation, thrombi were excised and digested in streptokinase/plasmin treptokinase/plasmin (150/2 U/ml) for 30 min att 37 °C under constant shaking. To To generate geene nera rate ra te a single ingle cell suspension, thrombi were passed through a 70 μm cell strainer (BD Biosciences, Heidelberg, Heid He idel id elbe el berg be rg,, Ge rg G Germany), rm many), an washed with FACS buff buffer fer e and and stained forr 30 30 min min with fluorochromecoupled co oupled up antibodies antibo odi diess as as described deesccribe ribe bedd inn the the h ddata ata sup ssupplement. uppleement ntt. Histology Hiist stol olog ol ogyy and og and fluorescence flluooreesc scen e cee microscopy en mic icro rosc ro scop opyy op Excised thro thrombi omb mbii an andd lu lung lungs n s we ng weree ffixed ixed ix e iin ed n 4% P PFA FA oorr em embedded mbe bedd dded dd ed iin n Ti T Tissue-Tek ssue ss ue-T ue -T Tek k ((Weckert Weck We cker ck et Labortechnik, Kitzingen, Germany) and frozen at -20 °C. Sections of 8 μm or 14 μm were cut and processed for immunohistochemical staining as described previously23,24,32 (see data supplement for more details). Statistics All data were evaluated for normal distribution using the Shapiro-Wilk test and are given as mean values r standard deviations (SD). Statistical difference was assessed by Welch´s test (for unequal variances) and a level of p
