Tetrazine Carbon Nanotubes for Pretargeted In Vivo “Click‐to‐Release” Bioorthogonal Tumour Imaging

Abstract The bioorthogonal inverse‐electron‐demand Diels–Alder (IEDDA) cleavage reaction between tetrazine and trans‐cyclooctene (TCO) is a powerful way to control the release of bioactive agents and imaging probes. In this study, a pretargeted activation strategy using single‐walled carbon nanotubes (SWCNTs) that bear tetrazines (TZ@SWCNTs) and a TCO‐caged molecule was used to deliver active effector molecules. To optimize a turn‐on signal by using in vivo fluorescence imaging, we developed a new fluorogenic near‐infrared probe that can be activated by bioorthogonal chemistry and image tumours in mice by caging hemicyanine with TCO (tHCA). With our pretargeting strategy, we have shown selective doxorubicin prodrug activation and instantaneous fluorescence imaging in living cells. By combining a tHCA probe and a pretargeted bioorthogonal approach, real‐time, non‐invasive tumour visualization with a high target‐to‐background ratio was achieved in a xenograft mice tumour model. The combined advantages of enhanced stability, kinetics and biocompatibility, and the superior pharmacokinetics of tetrazine‐functionalised SWCNTs could allow application of targeted bioorthogonal decaging approaches with minimal off‐site activation of fluorophore/drug.


Introduction
Effective molecular imaging and therapy relies on selective accumulation of ad iagnostic probe or therapeutic agent at the site of interest. One such approach is on-demand activation [1] in which the potency/activity of the drug/probe is attenuated but which, upon chemical reaction or enzymatic interaction, is converted into its active form at the desired site of action. This method helps spare healthy tissues by preventing adverse and off-target side effects.M oreover,i t allows ap recise control of the drug/probe activity and has been employed in the course of pretargeted strategies for imaging and therapy. [1] Bioorthogonal chemistry is apowerful tool for on-demand activation. Notably,the inverse-electrondemand Diels-Alder (IEDDA) reaction between tetrazine and trans-cyclooctene (TCO) has enormous potential for in vivo bioconjugation by capitalizing on its fast reactivity, even at low concentrations in acomplex biological context, [2] and it is inert to biological functionalities.F urthermore, a" click-to-release" bioorthogonal cleavage reaction that enables instantaneous release of asubstance from TCO after tetrazine ligation was reported and has potential for prodrug activation. [3] Prodrug activation with this method was first validated by Robillard and co-workers who demonstrated efficient and selective activation of aTCO-caged doxorubicin prodrug [4] and am onomethyl auristatin Ep rodrug [5] from antibody-drug conjugates in tumours.M ore recently,t he same authors have reversed the roles of the TCO and tetrazine. [6] AT CO reacts with at etrazine linker that is substituted with methylene-linked carbamate,w hich leads to a1,4-elimination of the carbamate and release of asecondary amine.A lthough these examples highlight the potential of bioorthogonal decaging reactions between tetrazine and TCO under complex biological conditions,o ptimization of the reaction kinetics and pharmacokinetics of the molecules carrying the two complimentary functionalities remains challenging.
Herein, we describe ap retargeting strategy that takes advantage of the fast bioorthogonal cleavage between tetrazine and TCO and the tumour-targeting ability of nanomaterials equipped with bioorthogonal reagents to enable localized enrichment of active drug/probe with tumour specificity and spatiotemporal precision. Single-walled carbon nanotubes (SWCNTs) have attractive properties; au nique quasi one-dimensional structure,h igh loading efficacy, and multivalent effect. SWCNTs also accumulate in tumours [7] through enhanced permeability and retention (EPR) effects [8] when systemically injected, which makes them suitable as delivery vehicles for cancer therapy. SWCNTs coated with phospholipid polyethylene glycol (PEG) are stable in vivo with ab lood circulation time of approximately 1.2 hours.I na ddition, SWCNTs also show relatively low uptake by the reticuloendothelial system and near-complete clearance from the main organs after about 2months, [9] which makes these functionalized SWCNTs safe for in vivo applications.S WCNTs are used routinely to diagnose tumours [10] and therapeutically as carriers to deliver chemotherapeutic drugs, [8a, 11] proteins, [12] plasmid DNA, [13] and small interfering RNA. [14] Herein, we present an ew strategy in which the bioorthogonal activation of responsive drugs or probes in solid tumours is under strict spatiotemporal control ( Figure 1). Thestrategy comprises two bioorthogonal reagents:t etrazine-modified SWCNTs (TZ@SWCNTs) and the TCO-carbamate containing molecules,achemotherapeutic prodrug or diagnostic probe,w hich are sequentially administered in two steps.F irstly,T Z@SWCNTs administered systemically will accumulate in tumour tissues as aresult of EPR effect. In the second step,t he effector molecules whose functions are largely attenuated via TCO protection are systemically applied. TheT Z@SWCNTs enriched specifically in tumour sites will serve as abioorthogonal trigger and activate the effector molecules in situ while sparing normal tissues.T his tumour-pretargeted activation strategy,w hich explores carbon nanotubes,e nables localized enrichment of active drug/probe with spatiotemporal control and selectivity for tumour sites over normal tissues,o ffering improved effectiveness and safety for tumour imaging and therapy.
Furthermore,w ea im to expand the strategy of real-time targeted fluorescence imaging in vivo,i nanon-destructive way with ahigh signal-to-noise ratio.Fluorescence imaging is ap owerful tool in scientific research and medical diagnosis because of its sensitivity and non-invasiveness,but the limited penetration depth has restrained its further applicability in vivo.Asolution to address the problem is to develop and implement fluorophores that emit in the near infrared (NIR) range 700-1000 nm, allowing in-depth imaging of organisms with minimum background autofluorescence and minimum damage to tissues-a significant advantage over fluorophores that emit in the ultraviolet and visible range.N IR fluorescence imaging is particularly beneficial for image-guided tumour surgery since it can accurately visualize tumour margins in real time. [15] Furthermore,a ne xcellent signal-tonoise ratio is critical for an effective bioimaging probe. Accordingly,f luorogenic probes that enable turn-on fluores-cence of the reporter in response to as pecific trigger are powerful tools for biological imaging and diagnosis in an in vivo context. ATCO-caged coumarin probe was developed by Chen; [3c,16] however, its emission in the blue region (440-460 nm) limits its use in in vivo applications.A lthough some tetrazine-based probes with aturn-on emission upon IEDDA ligation have been developed, most emit in the UV/Vis range. [17] There are still al imited number of fluorogenic probes with emission in the NIR range available for chemically controlled illumination in living animals.W ith this in mind, we seek to develop afluorogenic bioorthogonal probe that allows for ar eal-time targeted fluorescence tumour imaging in vivo.
Foro ur approach, hemicyanine (HCA),acationic dye, was chosen because it is stable,e mits in the NIR, and its optical properties are tuneable through modification of the amine group. [18] Therefore,H CA was ligated to TCO to give TCO-carbamate HCA (tHCA). Thef luorescence of tHCA can be effectively quenched and restored upon tetrazinetriggered ligation and liberation. This small molecular probe can diffuse deeply into atumour and facilitate tumour-specific imaging when in combination with TZ@SWCNTs.
Overall, our new pretargeted strategy combines SWCNTs and tetrazine/TCO cleavage reaction to 1) enable spatiotemporal control over prodrug activation by integrating the specific targeting property of nanomaterials and the favourable pharmacokinetics and fast clearance of the small effector molecule,a nd 2) enable real-time,n on-destructive tumour fluorescence imaging with ahigh target-to-background ratio. We anticipate that this approach, and the reported tetrazine carbon nanotubes and NIR fluorogenic probe,w ill add significant value to existing targeting strategy for both diagnostic and therapeutic applications. Figure 1. Illustration of the pretargeteds trategy for selective activation in cancer cells. The approach takes advantage of the EPR effect-enabled accumulationo fTZ@SWCNTsa nd the bioorthogonal IEDDA cleavage reaction between tetrazine and TCO to selectivelyd eliver active therapeutic drugs or imaging probes to the tumour.

Results and Discussion
Pretargeted imaging with TCO-AMC carbamate conjugate To produce TZ@SWCNTs,t etrazine was loaded onto SWCNTs by conjugating methyl tetrazine amine (mTZ) to PEGylated phospholipid (DSPE-PEG2000) with an N-hydroxysuccinimide terminus to form DSPE-PEG-TZ (for synthetic details see the Supporting Information) and then attaching this to SWCNTs by sonication. Ther esulting TZ@SWCNTs were approximately 200 nm long.This method efficiently loads tetrazine onto SWCNTs,w hich can reach al oading efficiencyo fu pt o0 .24 mmol tetrazine per milligram of SWCNTs.I nt his study,w eu sed approximately 33 mmol tetrazine per milligram of SWCNTs.
To test the feasibility of our pretargeted activation strategy,wefirstly demonstrated the design with the reported TCO-AMC carbamate conjugate (tAMC). [16] We investigated the kinetics of tAMC liberation in the presence of mTZ, DSPE-PEG-TZ, or TZ@SWCNTs (2.5 equiv) in phosphate buffered saline (PBS) with DMSO (10 %) by monitoring the fluorescence recovery ( Figure 2a). AMC release is dramatically faster in response to DSPE-PEG-TZ and TZ@SWCNTs relative to the small molecule counterpart mTZ. In the presence of DSPE-PEG-TZ and TZ@SWCNTs,t AMC effected over 30 %fluorescence recovery relative to the same concentration of AMC in < 50 min, whereas with mTZ the percentage of recovery was 4.2 %a nd 29 %a fter 50 minutes and 50 hours,respectively ( Figure 2c). This difference can be explained by the improved solubility from fusing mTZ to the PEG chain. To assess whether the rate acceleration was due to the enhanced cycloaddition reaction or the sequential elimination, we measured the kinetics by following the decrease of tetrazine absorbance at 530 nm. Thes econd-order rate constants of mTZ and DSPE-PEG-TZ toward tAMC are comparable (Supporting Information, Figure S2), which indicates the accelerated tAMC fluorescence recovery in the presence of DSPE-PEG-TZ is caused by afaster elimination step.T oe nsure the reaction could proceed in physiological media, we tested the reaction in pure human plasma. Interestingly,t he reaction of DSPE-PEG-TZ and TZ@SWCNTs toward tAMC was slower compared to activity in 10 %DMSO/PBS and reached amaximum of 50 %release with ah alf-time of 88 and 103 minutes,r espectively (Figures 2b,d). This difference was presumably ar esult of the increased viscosity of pure human plasma. In contrast, the reaction with mTZ was faster with ahalf-time of 135 minutes; the increase is likely due to the improved solubility of mTZ in human serum albumin. Overall, we showed that tetrazine could be loaded onto SWCNTs with high loading efficiency and, importantly,t he resulting TZ@SWCNTs demonstrated enhanced decaging reactivity toward tAMC when compared to its small molecule counterpart mTZ.
Subsequently,w es tudied the pretargeted fluorogenic imaging of tAMC with TZ@SWCNTs on human breast adenocarcinoma (MCF-7) cells.T he cells were pretreated with 50 mm TZ@SWCNTs for 6hours before being exposed to 10 mm tAMC for another 3hours.Asshown in Figure 2e,the cells pretreated with TZ@SWCNTs revealed as trong fluorescence signal in cytosol, whereas the cells treated with tAMC only showed minimal background signal. This result indicates that TZ@SWCNTs could effectively target MCF-7 cells and trigger selective fluorescence restoration of tAMC in living cells.

Pretargeted prodrug activation
After validation of the pretargeted strategy with tAMC, we explored the application of the strategy in activating prodrugs in living cells.W ea ssessed the biocompatibility of mTZ, DSPE-PEG-TZ, and TZ@SWCNTs on three different breast cancer cell lines:MDA-MB-231 (triple-negative breast cancer, negative for HER2 expression), MCF-7 (breast adenocarcinomas,p ositive for HER2 expression), and SK-BR-3 (breast adenocarcinomas,HER2 over-expression). Cell viability was evaluated after cell exposure to various concentrations of tetrazines for 24 hours.T he three tetrazines exhibited diverse biocompatibilities toward the different cell lines tested. MCF-7 and SK-BR-3 cells exhibited an 80 %cell viability after being exposed to 100 mm TZ@SWCNTs for 24 hours,w hereas mTZ and DSPE-PEG-TZ showed am uch higher toxicity ( Figure S7) under the same conditions.T his reduced cytotoxicity of TZ@SWCNTs could be aresult of the reduced availability of free DSPE-PEG-TZ once attached to SWCNTs,s uggesting one more advantage of the nanoconstruction in addition to its intracellular and intratumoural transport capability.T hese results demonstrate that TZ@SWCNTs have improved biocompatibility relative to free mTZ and DSPE-PEG-TZ.
Guided by these results,w ee xplored anticancer doxorubicin (DOX) prodrug activation in living cells with TZ@SWCNTs.T he cytotoxicity of TCO-DOX carbamate (tDOX) prodrug was expected to be attenuated by the chemical modification. [5] tDOX prodrug was synthesized according to the procedure previously reported. [5] As illustrated in Figure 3a,t he cytotoxicity of tDOXp rodrug, free DOX, and the pretargeted strategy was investigated over 72 hours.F or the pretargeted strategy,c ells were pretreated with 25 mm TZ@SWCNTs for 6hours before being exposed to various concentrations of tDOX for another 72 hours.tDOX prodrug displayed significantly reduced cytotoxicity than free DOX with an EC 50 value of 5.59 mm against MDA-MB-231 and 6.05 mm against SK-BR-3, which are 138-and 100-fold higher than parent DOX, respectively (Figures 3b,c). For cells pretargeted with TZ@SWCNTs,t he anticancer capability of tDOX was notably restored, presenting a2 5-fold increase in cytotoxicity relative to tDOXa lone.T he significantly enhanced therapeutic efficiencyrepresents an effective tDOX prodrug activation in cells that contain TZ@SWCNTs. The > 25-fold enhancement of tDOXt oxicity in pretargeted cells relative to non-targeted cells demonstrates improved selectivity and safety of our pretargeted activation strategy over free DOX. Nevertheless,aprodrug approach is not ideal for MCF-7 cells,which are much less sensitive to DOX. [19] The EC 50 of DOX (0.19 mm)was only 43.9 times lower than that of tDOX prodrug (8.3 mm)( Figure S8). Fort he pretargeted approach, the EC 50 (1.8 mm)i ncreased 4.6-fold relative to tDOX alone.These results lead us to suggest that the prodrug activation strategy is more effective for cells with high sensitivity to the parent therapeutic drug. After evaluating the tDOX-concentration-dependent cytotoxicity,weassessed the cytotoxicityo f1mm tDOX in combination with various concentrations of TZ@SWCNTs (Figure 3d). With subsequent tDOX treatment, TZ@SWCNTs provoked intracellular DOXr elease and restored therapeutic potencya tac oncentration as low as 1 mm and reached the highest activation at 25 mm.T he trigger TZ@SWCNTs exhibited no obvious inhibition of cell viability up to 100 mm,s uggesting the wide selectivity window of the strategy on pretargeted cells over normal cells.T hese results clearly indicate the efficiency of TZ@SWCNTs to accumulate in cancer cells and liberate active doxorubicin from tDOX prodrug inside live cells, selectively exhibiting anticancer efficacyi np retargeted cells over normal cells.

Design of an NIR fluorogenic probe tHCA
After establishing the pretargeted strategy with the blue fluorescent dye tAMC and the prodrug tDOX, we expanded the strategy to include af luorogenic NIR tetrazine-uncaging probe for in vivo applications.T he fluorogenic probes only become fluorescent upon activation, enabling the minimization of background signal and omittance of cumbersome washing steps.The NIR emission in the 700-1000 nm window facilitates deep tissue penetration and non-destructive imaging. To date there are no examples of afluorogenic NIR probe that can be activated with bioorthogonal control. Therefore, we developed af luorogenic NIR probe with emission at 720 nm by protecting the amine group on HCAw ith TCO (Figure 4a). Starting from ac ommercially available IR780 iodide,tHCAwas obtained in three steps as adark-blue solid. Thea bsorption and fluorescence spectra of HCA and tHCA in ap H7.4 PBS solution containing 10 %D MSO were recorded ( Figure S3). HCA shows am aximum absorption at 677 nm;h owever,u pon masking using TCO,t he absorbance peak of tHCAexhibited ahypsochromic shift to 656 nm and excitation at 665 nm, resulting in significantly reduced emission at 720 nm (38-fold less relative to HCA). Thes hift in absorption and the significant decrease in fluorescent emission indicates that tHCAc an be effectively quenched largely by an internal charge-transfer (ICT) process.U pon addition of TZ@SWCNTs (50 mm), the cleavage reaction was immediate,asdetected by adistinct colour change from light blue to dark blue and ar emarkable eightfold fluorescence enhancement at 720 nm (Figure 4b), which demonstrated the fluorogenic nature of the tHCAp robe in response to the tetrazine-mediated cleavage reaction. Induced conversion of tHCAinto HCA by DSPE-PEG-TZ was further validated by high-performance liquid chromatography (HPLC;F igure S4). These results confirm that the fluorescence emission of tHCAiseffectively quenched by TCO masking and could be restored by tetrazine-triggered cleavage reaction.
Thes tability of tHCA in biological context was investigated by HPLC-PDA( Figure S6). ThetHCAwas incubated in 20 v/v %h uman plasma/PBS at 37 8 8Cf or 4hours and 24 hours before the proportion of intact tHCA was analysed Angewandte Chemie Research Articles by HPLC-PDA. Theresults demonstrated that 82 %oftHCA remained intact after 4hours and 52 %a fter 24 hours, indicating that tHCAc ould remain stable during circulation in the blood stream after intravenous injection in mice and, therefore,eligible for systemic in vivo administration. We also evaluated the stability of the three tetrazines in human plasma by measuring their reactivities toward equimolar tHCA. Thed ecaging efficiency of TZ@SWCNTs,D SPE-PEG-TZ, and mTZ toward tHCAremained 93, 94, and 77 % respectively after 4hours of incubation in pure human plasma at 37 8 8C( Figure S5). These results together confirm that the two components of the strategy,T Z@SWCNTs and tHCA, are stable in physiological media.
Once we verified the fluorogenic nature and the stability of our tetrazine reactive tHCAp robe,w ei nvestigated the release profile of HCA from tHCAi nr esponse to various tetrazines by following the fluorescence emission at 720 nm using amicroplate reader.Wefirst tested the reactions in PBS containing 10 %D MSO (Figure 4c). In the absence of tetrazines,the emission of tHCA(20 mm)remained negligible over the recording period of 40 hours.O nce TZ@SWCNTs (50 mm)was added, an immediate fluorescence enhancement of 7.8 times was detected, corresponding to over 25 % liberation of HCA. Thefluorescence intensity then gradually increased with time and reached an equilibrium of 41 % recovery at 40 hours.T he tHCAr elease profile triggered by DSPE-PEG-TZ mirrored the release in the presence of TZ@SWCNTs.I nc omparison, the reaction of mTZ with tHCAw as much slower with < 5% fluorescence recovery after 60 minutes,25%after 20 hours,and amaximum of 35 % recovery after 28 hours.
To further assess the stability and reactivity of the tHCA probe in ab iological context, we performed the tetrazinemediated tHCA decaging reactions in pure human plasma at 37 8 8C (Figure 4d). In the absence of tetrazines,tHCA (20 mm) showed no enhancement of fluorescence intensity in pure human plasma over 24 hours.W hen incubated with TZ@SWCNTs,D SPE-PEG-TZ, and mTZ (50 mm), tHCA reached am aximum 40 %r ecovery with half-times of 104, 105, and 211 minutes,respectively.The results were consistent with the fluorescence recovery profiles of tAMC.T hese results confirm the stability of tHCA and its inducible fluorescence by tetrazine-mediated decaging reaction in real biological scenarios.T hese results also highlight the advantage of the fast cleavage reaction between TZ@SWCNTs and tHCAt oinduce instantaneous fluorescence enhancement.

Pretargeted fluorogenic NIR imaging in live cells
After verifying the fluorogenic properties,s tability,a nd reactivity of tHCA, we explored the application of our tHCA probe for pretargeted live-cell imaging.F irst, we evaluated the biocompatibility of tHCAb ym easuring the cell viability of MCF-7 cells incubated with tHCA. Theviability of MCF-7 cells remained over 70 %after incubation with 2.5 mm tHCA for 2hours and then in complete DMEM for another 72 hours ( Figure S9). This result indicates that tHCA is biocompatible and eligible for fluorogenic imaging in vivo given the short blood circulation half-time of small molecules and the instantaneous reaction between TZ@SWCNTs and tHCA. Guided by the aforementioned results,w ei nvestigated the fluorogenic performance of tHCA in live cells.M CF-7 cells were preincubated with 50 mm TZ@SWCNTs for 6hours prior to tHCAt reatment. Accumulation of TZ@SWCNTs inside cells was visualized by labelling TZ@SWCNTs with Cy3 fluorophore for confocal fluorescence imaging (Figure S10). Given the advantage of low background from tHCA and the instantaneous cleavage reaction between tHCAa nd TZ@SWCNTs,wecarried out real-time fluorescence imaging immediately after exposure of cells to PBS containing 1 mm tHCA. MCF-7 cells treated sequentially with TZ@SWCNTs and tHCAshowed anotable fluorescence signal in 5minutes ( Figure S11) post tHCAa ddition, and the fluorescence intensity increased over 30 minutes (Figure 5a), whereas cells treated with tHCAa lone showed negligible fluorescence (Figure 5b). After 30 minutes of tHCAtreatment, pretargeted cells induced as ignificant fluorescence enhancement by af actor of 6.5 relative to cells treated with tHCA alone (Figure 5d). Furthermore,c onsidering the liberated HCAi s positively charged and prone to accumulate in mitochondria, we investigated the intracellular localization of liberated HCA using mitochondrial probe MitoSpy Green (Figure 5c). MitoSpy Green is ag reen-fluorescing stain that localizes to mitochondria regardless of mitochondrial membrane potential. As shown in Figure 5e,the NIR fluorescence signal from liberated HCA colocalized well with the Mito-Spy Green, with aPearsonscorrelation coefficient of 0.92. These results indicate the efficient and fast cell permeability of tHCAp robe and the almost instantaneous fluorescence signal in cells containing TZ@SWCNTs with no need for washing. Them itochondria-targeted property of the turn-on tHCAp robe brings further potential in mitochondria-based applications.

Pretargeted fluorogenic in vivo NIR tumour imaging
Based on the successful NIR fluorogenic imaging in live cells with our two-component strategy,weperformed tumour- pretargeted turn-on fluorescence imaging in living mice with tumour xenografts.C T26 tumour (undifferentiated colon carcinoma cell line)-bearing BALB/c mice were randomly allocated into 5g roups (n = 5) when the tumour volume reached approximately 300 mm 3 .T he pretargeted imaging group received ad ose of 25 mmol kg À1 TZ@SWCNTs and ad ose of 2.4 mmol kg À1 tHCA after a2hour interval, both through intravenous injection. Thet ime lag of 2hours was applied to allow TZ@SWCNTs to be cleared from the blood circulation and accumulate in the tumour regions via the EPR effect. TZ@SWCNTs alone,tHCAalone,and saline were also injected in parallel as control groups.I nvivo fluorescence images of live mice were obtained at various time points after tHCAi njection using an IVIS Lumina fluorescence/bioluminescence imaging system. Form ice treated with the pretargeted imaging strategy (TZ@SWCNTs tumour-pretargeting and subsequent tHCA injection), an indistinct signal was observed in the abdomen at 1.5 hours and became negligible at 3hours post tHCAi njection (Figure 6a). In tumour regions,p ronounced NIR fluorescence signals emerged at 3hours post tHCAi njection, and continued to increase over the course of 24 hours,indicating the release of HCA in the tetrazine-tagged tumours.T he fluorescence quantification of the tumours gave at umour-to-background ratio of 23:1 at 3hours post tHCA injection and the ratio increased to 42:1 after 24 hours (Figure 6b). In contrast, for mice treated with tHCAalone,only very low emission levels in mice abdomen were seen 1.5 hours post tHCAi njection and the signal was negligible after 3hours.T his result indicated rapid clearance of tHCAp robe from the body and its specific activation in tetrazine-tagged tumours.N otably, am uch lower dose of tetrazine is required by this strategy compared to the utilization of small molecule tetrazines and even antibody-tagged tetrazines [3c, 5] because of the high tumour delivery efficiency of TZ@SWCNTs.F urthermore, the result showed that an interval of 2hours between the administration of the two components is sufficient due to fast clearance of SWCNTs from the blood circulation, [9] which is much shorter than the 48 hours required by antibodyfacilitated tetrazine/TCO cleavage-reaction-triggered activation. [5] After 24 hours post tHCAi njection, the mice were euthanized, and the tumours and major organs were harvested for ex vivo fluorescence measurements and assessment. Thef luorescence images of the tumours and major organs were obtained to evaluate the in vivo release profile of tHCA as shown in Figures 6c and 6e.T he pretargeted strategy induced ar emarkable fluorescence signal in the tetrazinetagged tumours.T he other organs,s uch as liver and kidneys, showed much weaker fluorescence signals.Notably,the signal in the tumours was 3.6 times higher than that in liver.T hese non-specific signals were presumed to be from the reaction between tHCAa nd TZ@SWCNTs residues retained in the liver and kidneys during renal and hepatic clearance from the body.Ont he other hand, for mice treated with tHCA alone, we only observed negligible signals in the tumours and weak signals in the liver and kidneys.T his result further confirmed that the activation of the fluorescent signal is tetrazinedependent and thus tumour specific.
To compare the pretargeted activation strategy with direct targeted delivery,ag roup of mice were treated with as ingle dose of TZ@SWCNTs-Cy5 at aTZdose of 25 mmol kg À1 and acorresponding Cy5 dose of 0.7 mmol kg À1 for parallel study. Theexvivo fluorescence image of tumours and major organs at 26 hours post TZ@SWCNTs-Cy5 injection are shown in Figure 6d,a nd the corresponding fluorescence intensities are shown in Figure 6f.T he fluorescence signals of TZ@SWCNTs-Cy5 in the liver and kidneys were higher than in the tumours.T he resulting tumour-to-liver ratio is 0.18, which is about 20 times lower than that of the pretargeted activation strategy.T his result demonstrated that, compared to the direct tumour-targeted fluorophore delivery,t he twostep activation strategy largely avoided the signals of the active probe in normal tissues and enabled greater tumour specificity.

Conclusion
Our work presents an ew pretargeted activation strategy that integrates SWCNTs and an IEDDAc lick-to-release reaction to achieve tumour-specific delivery of active imaging agents and anticancer drugs with spatiotemporal control. Furthermore,w ed escribed the use of ab ioorthogonally applicable fluorogenic NIR probe for real-time in vivo imaging in which fluorescence was turned-on in response to tetrazines.I nt his pretargeted activation strategy,t he TZ@SWCNTs accumulate specifically in tumour cells as aresult of the EPR effect and tagged the tumour region with the bioorthogonal tetrazine handle.T he tetrazine handle serves as the trigger to react in situ with the TCO conjugate to release the active effector molecules.T Z@SWCNTs displayed an exceptionally high loading efficiency,a ccelerated activation rate,and improved biocompatibility relative to the corresponding small molecule mTZ. As shown in this study, our pretargeted activation strategy enabled selective tDOX activation in cells chemically tagged with tetrazines.T aking advantage of the pretargeted activation strategy and our tHCAf luorogenic probe,w eh ave developed aN IR fluorescence imaging technique in living cells with an instantaneous turn-on signal in mitochondria with no need for washing.With sequential systemic administration of TZ@SWCNTs and Figure 6. In vivo fluorescence imaging of tHCA via pretargeted strategy in mice. Mice bearing subcutaneous CT26 xenografts on the right flank (n = 5). a) Non-invasive fluorescence images of tumour-bearing mice over time and b) time-dependent tumour-specific NIR fluorescence intensity in live mice treated with saline, TZ@SWCNTsalone, pretargeted strategy,o rtHCA alone at different time points post-tHCA injection (n = 5). c) Ex vivo fluorescence images of dissected tumoursand major organs (liver,lungs, spleen, kidneys, heart) and e) release profiling of tHCA 24 h after tHCA injection. d) Ex vivo fluorescence images of dissected tumours and major organs and f) the biodistribution of Cy5f rom mice treated with TZ@SWCNTs-Cy5. Fluorescence is expressed in radiance:photons/second/cm 2 /steradian. tHCAi nl iving mice with xenograft tumours,t he tetrazinetriggered fluorogenic imaging resulted in higher tumour selectivity over liver and kidneys compared to direct fluorophore delivery.T he distinct fluorescent signals in tumour regions lasted up to 24 hours after intravenous administration of the probe.O ur results demonstrate that activatable NIR imaging guarantees tumour selectivity over normal tissues and facilitates in-depth imaging with ah igh signal-to-noise ratio in ar eal-time and non-destructive manner. We anticipate that the tumour-targeting efficiencywould be potentially improved further by incorporating tumour-targeting moieties. Moreover,w ith its fluorogenic property in instantaneous response to tetrazines and its mitochondria-orientedn ature, tHCAh as the potential to be used for other diagnostic and imaging applications,s uch as fluorescence-guided tumour surgery,s uper-resolution bioimaging and high-throughput screening,a sw ell as in combination with other tetrazinecontaining platforms such as antibodies.F inally,w ee xpect that this strategy could be applied to activate other TCOcaged effector molecules for both diagnostic and therapeutic applications.