李 奧，李凌軒，左翠翠，陳傳凱，樊一凡，步逸凡，林泓域，高錦豪

（廈門大學(xué)化學(xué)化工學(xué)院化學(xué)生物學(xué)系,譜學(xué)分析與儀器教育部重點(diǎn)實(shí)驗(yàn)室,福建省化學(xué)生物學(xué)重點(diǎn)實(shí)驗(yàn)室,廈門 361005）

1 Introduction

Reactive oxygen species（ROS）have been widely accepted to have deleterious consequences when they are excessive in the cells but also serve as signaling molecules at low concentrations［1—3］.Mounting evidence in a number of previous reports reveals that cells can manipulate ROS levels during various biological processes to maintain the homeostasis and fitness of living organisms［4—6］.Thus，the level of ROS is kept within a proper range during normal cellular metabolism［7，8］.In contrast，during the abnormal metabolism of many diseases，such as inflammation，tumor，and organ injuries，an aberrant level of ROS is often observed and regarded as one of the significant features［9］.Nowadays，the annual incidence of cancer has exceeded 20 million per year all around the world，which generates an urgent demand on research methods and diagnostic means for tumor［10］.Because of the high level of ROS in tumor，detection techniques for monitoring the level of ROS has become one of the most promising tools for tumor research and diagnosis.Plenty of interesting sensors based on the reaction-based indicated assay（RIA）have been reported for the detection and imaging of ROS［11—15］.Most of the chemical structures of these probes contain a boronate group both as a quencher for fluorescent dyes and a responsive moiety to ROS［16，17］.Nevertheless，some of these boronate-based probes are yet to overcome the challenges ranging from water-solubility to photo-stability.Moreover，fluorescence imaging methods are of several limitations when appliedin vivo，including shallow tissue penetration and autofluorescence interference［18］.

Over last decades，as a non-ionizing and multi-parameter imaging technique with deep penetration and high resolution，magnetic resonance imaging（MRI）has achieved eminent success in tumor imaging and diagnosis［19—21］.Nowadays，1H MRI，which takes advantage of the difference in relaxation times for abundant protons（in water or lipid molecules）in various tissues，demonstrates its outstanding capacity for obtaining anatomical details from soft tissues［22—24］.However，strong background signals from the complicated internal environment often impede further applications of1H MRI for detecting small bioactive molecules like ROS in deep tissues［25］.Fortunately，19F stands out as a promising nucleus complementary to1H for MRI due to its favorable properties，such as good sensitivity（83%of1H），100%natural abundance，and low biological distribution（＜10-6mol/L）［26—29］.These advantages render19F MRI promising as a feasible means for detecting low-concentration biomoleculesin vivoto provide“hot-spot”images with nearly negligible background［30—32］.Several strategies have been proposed for the design of19F MRI probes for bioactive molecules［33，34］.Among them，RIA has been extensively used in the construction of activatable19F MRI probes，which can switch19F MRI signals“on”or“off”in response to certain stimuli.Currently，several RIA-based19F MRI probes for imaging biomoleculesin vivohave been reported，including HClO and other specific ROS［35—38］，metal ions［39，40］，enzymes［41，42］，neurotransmitters［43］and acidosis［44—49］.Nevertheless，there are few reports regarding19F probes forin vivoimaging of generic ROS.

Herein，we report a generic ROS-responsive19F MRI probe（Gd-DPBF），which consists of a19F signal modulator Gd-chelate（Gd-DP）with a highly-fluorinated small molecule 3，5-bis（trifluoromethyl）-phenylboronic acid（BTFB）viaa widely used ROS-responsive boronate linker.The19F signal modulator（Gd chelate）imposes strong paramagnetic relaxation enhancement（PRE）effect to19F nuclei in Gd-DPBF，which abates their relaxation times（T1andT2）significantly，leading to considerably suppressed19F NMR/MRI signals.In the presence of ROS（H2O2，HClO，etc.），which cleave the boronate linkage in Gd-DPBF and release the product 3，5-bis（trifluoromethyl）phenol（BTFP），the PRE effect is substantially attenuated due to the increased distance between the Gd chelate and19F nuclei，resulting in the extension of the19F relaxation times.Consequently，the intensity of19F NMR/MRI signals is remarkably enhanced，which allows for the detection and imaging of generic ROS（Fig.1）.The relaxation times of19F before and after incubation with a typical ROS，hydrogen peroxide，were measured to confirm the mechanism of our probe.Meanwhile，we investigated the response of Gd-DPBF to ROS underin vitroconditionsvia19F NMR/MRI.Furthermore，ex vivoimaging experiments of hydrogen peroxide in a piece of pork by19F MRI with Gd-DPBF manifested the potential of the probe for imaging ROS in deep tissues（such as tumor）.Finally，real-time19F MRI of ROS in the tumors of living mice with Gd-DPBF revealed its feasibility for detection and imaging of generic ROS in deep tissues of living subjects.

Fig.1 Schematic illustration showing the chemical structure and functioning mechanism of the ROS-triggered 19F NMR/MRI probe(Gd-DPBF)

2 Experimental

2.1 Materials and Measurements

Bromoacetyl bromide（99%）was purchased from Aladdin（China）；propargylamine（98%）was purchased from Inno-Chem（China）；trifluoroacetic acid（99%），cyclen（99%），tert-butyl bromoacetate（99%），gadolinium（III）chloride hexahydrate and 3-hydroxytyramine hydrochloride were purchased from J&K Scientific（China）.All chemicals were used as received without further purification.Unless noted otherwise，all reactions were performed under inert（N2or Ar）environments.Milli-Q ultrapure water（Resistivity：18.2 MΩ·m）was used in all experiments.

The molecular weights of the synthesized compounds were measured on a Bruker Esquire 3000 Plus electrospray ionization instrument using an ICR analyzer（ESI-MS）and Bruker microflex MALDI-TOF-MS.All1H NMR and13C NMR experiments were carried out on a Bruker AVANCE III HD Ascend（600 MHz for1H，151 MHz for13C）with tetramethylsilane as an internal referencing standard.All19F magnetic resonance imaging（19F MRI）and relevant1H MRI were performed on a Bruker BioSpec 94/20 system（400 MHz for1H and 376 MHz for19F）equipped with a 40 mm（inner diameter）volume coil.Image acquisition，SNR analysis and pseudocolor rending were carried out with ParaVision 5.1（Bruker BioSpin）.

2.2 Experimental Methods

2.2.119F NMR Characterization19F NMR experiments were carried out on a Bruker AVANCE III HD Ascend spectrometer（564 MHz for19F）using a 5 mm BBFO cryoprobe.19F NMR spectra were acquired with 18μs delay and 200 scans.Samples were prepared in 10%D2O/H2O solution and CF3COONa（δ-75.4）was used as a reference for chemical shift.

2.2.2 Relaxation Time Measurements Relaxation measurements were performed on the same Bruker AVANCE III HD Ascend spectrometer（564 MHz for19F）.Samples were prepared in 10%D2O/H2O solution for shimming.Longitudinal relaxation times（T1）were measured using an inversion recovery（IR）sequence.Transverse relaxation times（T2）were measured using a Carr-Purcell-Meiboom-Gill（CPMG）sequence.The vdlist（T1fitting）and VClist（T2fitting）were set according to exponential equations.The following equations were used forT1andT2curve fitting，respectively：

whereT1was directly obtained from the fitting report，T2=fitting cycle number×（2D20+P2），D20=0.005 s or 0.004 s，P2=28μs.

2.2.3 Detection of ROS/RNS and Other Analytes with Gd-DPBF by19F NMR Gd-DPBF were incubated with indicated analytes in PBS buffer（50 mmol/L，pH=7.4）at 25℃for 2 h.The resulting solutions were subjected to19F NMR with the aforementioned parameters.ONOO-was generated by mixing NaNO2with H2O2；HO·was made by mixing（NH4）2Fe（SO4）2with H2O2；ROO·was made by dissolving 2，2′-azobis（2-amidinopropane）dihydrochloride（AAPH）in water；other analytes were purchased from commercial sources and used as received.

2.2.4In vitro19F MRI19F MRI were acquired on a 9.4 T Bruker BioSpec MRI scanner with commercially available19F/1H MRI coils.Gd-DPBF（final concentrations as indicated）were incubated with indicated analytes in PBS buffer（50 mmol/L，pH=7.4）for 1 h.The resulting solutions were subjected to19F MRI.A RARE sequence was used to acquire19F MR images with the following parameters：TR/TE=400 ms/8.1 ms，flip angle=90°，F(xiàn)OV=4 cm×4 cm，the slice thickness=20 mm，Matrix=32×32 and 128 average（NEX=128）.The total experiment time was about 5.12 min.

2.2.5Ex vivoImaging of ROS/RNS by19F MRI with Gd-DPBF Gd-DPBF（10 mmol/L）in 1×PBS solutions were subcutaneously injected to two spots of a piece of pork.H2O2（15 mmol/L）dissolved in PBS was subcutaneously injected into one of the two spots as indicated.The center frequency corresponding to the19F chemical shift atδ-62.6 was chosen for19F MRI.For acquiring1H MR images，a RARE sequence was used with the following parameters：TR/TE=1000 ms/8.5 ms，flip angle=180°，F(xiàn)OV=4 cm×4 cm，slice thickness=1 mm，matrix=256×256，average=4（NEX=4）.The total acquisition time for each time point was about 3.2 min.For acquiring19F MR images，A RARE sequence was used to acquire19F MR images with the following parameters：TR/TE=400 ms/8.1 ms，flip angle=90°，F(xiàn)OV=4 cm×4 cm，slice thickness=15 mm，Matrix=32×32 and 196 average（NEX=196）.The total experiment time was about 7.84 min.

2.2.6 Animal Ethics Animal experiments were conducted according to the protocols approved by the Institutional Animal Care and Use Committee of Xiamen University.

2.2.7 Establishment of Tumor-bearing Mouse Models Tumor inoculation was carried out by injecting 200μL of cell suspension containing 1×107U87 cells subcutaneously into the right forelimbs of male nude mice.The tumors were allowed to reach 100—200 mm3before experimentation.

2.2.8In vivoImaging of ROS with1H/19F MRI Allin vivo1H/19F MRI were acquired on a 9.4 T Bruker MRI scanner with commercially available19F/1H MRI coils.Two groups，each of which contains three male nude mice，were intramuscularly or intratumorally injected with Gd-DPBF（120μL，15 mmol/L in PBS）and subjected to1H/19F MRI as indicated.For acquiring19F MR images，a RARE sequence was used with the following parameters：TR/TE=400 ms/8.1 ms，flip angle=180°，F(xiàn)OV=4 cm×4 cm，slice thickness=10 mm，matrix=32×32，average=1960（NEX=1960）.The total acquisition time was about 78.3 min.For acquiring1H MR images，a RARE sequence was used with the following parameters：TR/TE=1000 ms/8.5 ms，flip angle=180°，F(xiàn)OV=4 cm×4 cm，slice thickness=1 mm，matrix=256×256，average=4（NEX=4）.The total acquisition time for each time point was about 3.2 min.

3 Results and Discussion

3.1 Synthesis of Gd-DPBF

Gd-DPBF was synthesizedviaa facile approach（Scheme S1，see the supporting information of this papaer）.Specifically，the amino group of dopamine was transformed to an azido group to afford compound 1，which was coupled with a previously reported Gd-DO3A derivative 2 with a terminal alkyneviacopper（I）-catalyzed azide-alkyne cycloaddition（CuAAC）to give Gd-DP（3）as the19F signal modulator.Gd-DPBF（4）was prepared by esterification（44%yield）between Gd-DP and 3，5-bis（trifluoromethyl）phenylboronic acid（BTFB），the latter of which serves as the fluorine-building block.The boronate ester linkage could be facilely cleaved in the presence of ROS.Synthetic details as well as characterization data of Gd-DPBF and important intermediates are included in the supporting information of this paper.

3.2 Measurements of 19F Relaxation Times

Longitudinal and transverse relaxation times（T1andT2）of19F in Gd-DPBF were assessed with Inversion Recovery（IR，forT1）and Carr-Purcell-Meiboom-Gill（CPMG，forT2）sequences on an NMR spectrometer（564 MHz for19F）at 25℃.As shown in Table 1，theT1andT2of19F in Gd-DPBF in PBS were significantly shortened（less than 3 ms）compared to those in BTFP in PBS（987 and 806 ms，respectively），which indicates the strong PRE effect between19F nuclei and unpaired electrons of Gd3+ions in intact Gd-DPBF.However，after the reaction of Gd-DPBF with 0.3 mmol/L H2O2in PBS，theT1andT2of19F were substantially extended（416 and 279 ms，respectively），which implicates the considerable weakening of the paramagnetic relaxation enhancement（PRE）effect.These results suggest that our design can adjust the relaxation times of19F in Gd-DPBF before and after incubation with H2O2，which permits its detection and imaging with this activatable19F MRI probe.

Table 1 T1 and T2 of 19F in 3,5-bis(trifluoromethyl)phenol(BTFP)and Gd-DPBF before and after the responsive process*

We also utilized the high-performance liquid chromatography（HPLC）to investigate the response of Gd-DPBF to H2O2.As shown in Fig.S1，two peaks（Gd-DP and BTFB）appeared in HPLC chromatogram of Gd-DPBF alone，indicating the hydrolysis of Gd-DPBF during the HPLC process.After incubation of Gd-DPBF with 0.4 mmol/L H2O2，several new peaks appeared，which correspond to be the products Gd-DP and BTFP.These results implicate the successful response of Gd-DPBF to H2O2，which is in accordance with our design.

3.3 In vitro Sensing of Generic ROS with Gd-DPBF Using 19F NMR

To investigate thein vitrogeneric detection of ROS with Gd-DPBF by19F NMR，0.2 mmol/L Gd-DPBF in PBS was incubated with various ROS（including KO2，H2O2，ROO·，ONOO-and HClO）as well as common biomolecules such as proteins in FBS，glutathione（GSH）and glucose（Glu）.As shown in the spectra of Fig.2（A），sharp19F NMR peaks atδ-62.6 could be seen for Gd-DPBF in the presence of many types of ROS（including KO2，H2O2，ONOO-and HClO），yielding high relative signal-to-noise ratios（SNRs＞25）［Fig.2（B）］.By comparison，no significant19F NMR signals（relative SNRs＜5）were observed after the exposure of Gd-DPBF to the other biomolecules，indicating the excellent selectivity of Gd-DPBF for generic ROS［Figs.2（A）and（B）］.These results illustrate that Gd-DPBF could respond to generic ROS with high selectivity，resulting in strong19F signals.

Fig.2 Representative 19F NMR spectra of 0.2 mmol/L Gd-DPBF treated with various analytes(0.4 mmol/L)for 1 h in PBS buffer.CF3COONa(atδ-75.4)was used as an internal reference of 19F chemical shift(A),SNR analysis corresponding to(A)(n=3).The SNR for the peak of Gd-DPBF in PBS was set as 1.0 and the other SNRs were normalized accordingly(B),representative 19F MR images of Gd-DPBF phantoms at different concentrations before and after specific activation toward 2.2,4.4,6.6,8.8 and 11 mmol/L H2O2(1 h incubation)respectively in PBS buffer at 37℃(C)and SNR analysis of Gd-DPBF before and after activation corresponding to(C)(n=3).The SNR for the peak of 2.0 mmol/L Gd-DPBF in PBS was set as 1.0 and the other SNRs were normalized accordingly(D)

3.4 In vitro Imaging of Generic ROS Using 19F MRI with Gd-DPBF

The imaging of generic ROS by19F MRI with Gd-DPBF using a rapid acquisition using relaxation enhancement（RARE）sequence（TR/TE=400 ms/8.1 ms）on a 9.4 T MRI scanner equipped with commercially available1H/19F MRI coils was also investigated.With the given pulse sequence and parameters，no apparent19F MRI signals were observed for intact Gd-DPBF because theT2of19F was substantially shortened.As expected，upon the exposure of Gd-DPBF to excessive H2O2，the19F MRI signals were significantly enhanced［Fig.2（C）］.Furthermore，the concentration-dependent imaging and relevant SNR analysis indicate the positive correlation the between the intensities of19F MRI signals and the concentrations of19F［Fig.2（D）］.These results demonstrate the practicability of visualizing H2O2（and other ROS）with Gd-DPBF by19F MRI.

3.5 Ex vivo Imaging of Gd-DPBF by 19F MRI with H2O2

Ex vivo19F MRI utilizing porcine tissues were then carried out.Gd-DPBF was subcutaneously injected into a piece of pork at two different spots.The bottom spot was further injected with excessive H2O2［Fig.3（A）］.The injection spots could be clearly identified using1H MRI because the Gd-chelate moiety in Gd-DPBF and Gd-DP could serve as aT1contrast agent that can significantly enhance the1H MRI signals of surrounding tissues［Fig.3（B）］.To confirm this observation，the longitudinal relaxivities（r1）of Gd-DPBF and Gd-DP for1H MRI were measured to be 5.92 and 5.79 mmol·L-1·s-1at 0.5 T，respectively，which are both higher than that of Gd-DOTA（4.77 mmol·L-1·s-1），facilitating the generation of strongT1-weighted1H MRI signals（Fig.S2，see the Electronic Supplementary Material of this paper）.Meanwhile，the tissues injected with both Gd-DPBF and H2O2［the bottom spot in Fig.3（B）］showed strong19F MRI signals，the intensity of which reached the apex at 30 min after injection and gradually decreased afterwards，as confirmed by SNR analysis［Fig.3（C）］.In contrast，no significant19F MRI signals were observed for the tissues injected with Gd-DPBF only［the top spot in Fig.3（B）］.These results implicate the feasibility of imaging H2O2in deep tissues with Gd-DPBF by19F MRI，illustrating the potential of Gd-DPBF as a19F MRI probe forin vivoimaging of generic ROS.

Fig.3 Schematic illustration showing the protocol for ex vivo 19F MRI.The center frequency corresponding to the 19F chemical shift atδ-62.6 was chosen for 19F MRI(A),representative 1H and 19F MR images of a piece of pork(43 mm×16 mm×31 mm)at indicated time points after hypodermic injection of Gd-DPBF solution(10 mmol/L,200μL)in 1×PBS alone(top,indicated by the white circles)and incubated with 15 mmol/L H2O2(bottom,indicated by the pink circles)for 3 h(B),SNR analysis corresponding to(B)(n=3).The SNR for the 19F MR image of Gd-DPBF with PBS at 0 min was set as 1.0 and the other SNRs were normalized accordingly(C)

3.6 In vivo Imaging of Generic ROS via 19F MRI with Gd-DPBF in Living Mice

Encouraged by theex vivoimaging experiments，thein vivo“hot-spot”19F MRI with Gd-DPBF was further explored to visualize endogenous ROS in tumor-bearing living mice.The biocompatibility of Gd-DPBF was assessed before experiments.As a result，Gd-DPBF did not reveal obvious cytotoxicity to HepG2 or L02 cells even at 15 mmol/L.Additionally，no conspicuous microscopic lesions were observed for hematoxylin and eosin（H&E）stained tissue section of all major organs collected from the mice at 3 d after intravenous injection of Gd-DPBF（Fig.S3，see the supporting information of this paper）.Table S1（see the supporting information of this paper）shows the results of liver and kidney function tests（including ALT，AST，ALP and BUN）of BALB/c mice after intravenous injection of 200μL Gd-DPBF（15 mmol/L）.All indicators of the experimental group at 24 h were not significantly deviated from the reference ranges.These results demonstrated the good biocompatibility of Gd-DPBF，which permits further imaging experiment on animals.The right forelimbs of healthy BALB/c mice were intramuscularly injected with Gd-DPBF and were further subjected to MRI.Bright1H MRI signals were observed in the tissues of right forelimbs due to the enhanced1H MRI signals resulting from the Gd-chelate moiety of Gd-DPBF.Concurrently，the bladder regions also showed strong1H MRI signals，which could be ascribed to the quick renal clearance of Gd-DPBF［Fig.4（A）］.As expected，no apparent19F MRI signals were detected in the forelimb and bladder regions，indicating that Gd-DPBF is still intact under these conditions.However，when injected Gd-DPBF intratumorally to U87 tumor-bearing mice，evident19F MRI signals were observed in both the tumor and bladder regions along with strong1H MRI signals serving for co-localization analysis［Fig.4（B）］.This observation could be attributed to the releasing of BTFP and Gd-DP from Gd-DPBF due to excess endogenous ROS in the tumor，which are subsequently accumulated in the bladderviarapid renal clearance，leading to significant19F and1H MRI signals in both the tumor and bladder regions.These results demonstrate the feasibility of imaging intratumoral endogenous ROS in living subjects with the probe using19F MRI.

Fig.4 In vivo 19F MRI using Gd-DPBF

4 Conclusions

A small molecular19F MRI probe（Gd-DPBF）for detecting and imaging of generic ROS was developed，which consists of a fluorine-containing moiety and a paramagnetic Gd chelateviaa ROS-responsive boronate linkage.The feasibility of using Gd-DPBF for in-depth and real-time detection and imaging of ROSvia19F NMR/MRI have been clearly demonstrated byin vitro，ex vivo，andin vivoexperiments.Interestingly，Gd-DPBF and its cleaved products could undergo rapid renal clearance，which is beneficial for further biomedical applications.In addition，according to the flexibility in the constructing strategy of this probe，imaging targets can be easily extended to other biomolecules in different biological systems.Though there are still some challenges needed to be overcome for further clinical applications，we believe that Gd-DPBF and related19F MRI probes constructed according to our strategy，could serve as promising diagnostic agents for various diseases in the future.

The supporting information of this paper see http：//www.cjcu.jlu.edu.cn/CN/10.7503/cjcu20220545.