Tailoring renal-clearable zwitterionic cyclodextrin for colorectal cancer-selective drug delivery

Although cyclodextrin-based renal-clearable nanocarriers have a high potential for clinical translation in targeted cancer therapy, their designs remain to be optimized for tumour retention. Here we report on the design of a tailored structure for renal-clearable zwitterionic cyclodextrin for colorectal cancer-selective drug delivery. Twenty cyclodextrin derivatives with different charged moieties and spacers are synthesized and screened for colloidal stability. The resulting five candidates are evaluated for biodistribution and an optimized structure is identified. The optimized cyclodextrin shows a high tumour accumulation and is used for delivery of doxorubicin and ulixertinib. Higher tumour accumulation and tumour penetration facilitates tumour elimination. The improved antitumour efficacy is demonstrated in heterotopic and orthotopic colorectal cancer models. Optimizing the retention of drug delivery nanocarriers for improved cancer therapy has the potential to improve clinical outcomes. Here the authors screen 20 renal-clearable zwitterionic cyclodextrin-based nanocarriers for optimized biodistribution and tumour retention, demonstrating application in colorectal cancer models.

Despite the remarkable progress in nanotechnology and biomedical engineering, only a few injectable medical nanocarriers (NCs) are currently available for cancer diagnosis and treatment. This is mainly owing to potential toxicity concerns caused by off-target accumulation and long-term retention of NCs in normal organs [1][2][3][4][5][6] . Uncontrolled distribution profiles may result from the non-renal-clearable particle size (>10 nm), which is a primary prerequisite of conventional NCs utilizing the enhanced permeability and retention effect 1,2,5-7 . With a limited clinical translation rate, this old-school NC design faced a paradigm shift towards a renal-clearable ultrasmall nanoplatform [8][9][10][11][12][13][14][15] , which has several advantages compared with its non-renal-clearable counterpart: (1) rapid elimination of NCs distributed to off-target tissues via urinary excretion, (2) high tumour vascular permeability, and (3) highly reduced uptake by the mononuclear phagocyte system 9,16,17 . These favourable features allow the high-dose administration of cytotoxic cargo molecules with increased tumour selectivity, resulting in enhanced efficacy and minimal adverse effects 9,16,18-22 . However, renal-clearable NCs are not free from drawbacks that hamper their clinical translation. One major drawback is the residual toxicity caused by trace amounts of injected NCs, which may face inadvertent chronic accumulation 11,23 . Notwithstanding the short-term biocompatibility, long-term or lifetime exposure to undegradable foreign materials, such as silica, metal and synthetic polymer nanoparticles, discourages their clinical use 24,25 . To address this issue, renal-clearable NCs, namely, H-Dots, composed of biodegradable organic materials including β-cyclodextrin (CD)-grafted polylysine, were reported, demonstrating efficient tumour-targeted drug delivery with low background tissue retention 21,26 . However, the tumour selectivity of these NCs is mainly due to their significantly reduced residence time in normal tissues, not the augmented tumour-homing ability. This suggests that by extending the tumour retention of renal-clearable organic NCs, much higher tumour selectivity can be achieved. Although targeting ligands can be introduced into the NC structure 27,28 , this so-called active targeting strategy does not always ensure improved selectivity, unexpectedly inducing opsonization and consequent non-target accumulation, primarily via uptake by the mononuclear phagocyte system [29][30][31][32][33][34] . Therefore, delicate tailoring of the optimal NC structure is required for clinical success, balancing the targeting efficiency and off-target accumulation of renal-clearable NCs.

Biodistribution of CD derivatives
Given high affinity to the hydrophobic cavity of CD, adamantyl sulfocyanine 7 (ACy7) was successfully loaded onto CD 16-20 by inducing host-guest interactions under aqueous conditions 41,42 , resulting in stable CD/ACy7 inclusion complexes. Although CD 11 and CD 14 had decent colloidal stability (Fig. 1b), gradual precipitation was observed in the presence of ACy7 ( Supplementary Fig. 12b), discouraging further evaluation of these two compounds. As a negative control molecule, (ZW)-CD, lacking PBA in its structure, was synthesized and the corresponding ACy7 inclusion complex was prepared ( Supplementary Fig. 13).
Next, free ACy7, (ZW)-CD/ACy7 and CD 16-20/ACy7 inclusion complexes were intravenously administered to a HT-29 tumour-bearing mice, and ex vivo imaging was performed 24 h post-injection (Fig. 1d). Compared with free ACy7, which showed non-selective organ distribution, (ZW)-CD/ACy7 (zwitterionic) and CD 16/ACy7 (zwitterionic) showed significantly reduced lung and liver accumulation with increased kidney selectivity, implying that zwitterionic CD derivatives may reduce the off-target distribution or retention of cargo molecules and facilitate renal excretion. A similar phenomenon was observed in the case of CD 17/ACy7, another zwitterionic inclusion complex, wherein the average fluorescence intensity in normal tissues were almost at the same levels as those of CD 16/ACy7. However, the tumour distribution efficiency of CD 17/ACy7 was 2.02-fold higher than that of CD 16/ACy7 (Fig. 1e), even with similar physicochemical properties (Fig. 1c). This result could be attributed to the insertion of the PEG 4 spacer between the PBA moiety and glutamic acid linker, reducing the possible interference of sulfobetaine in the PBA-sialic acid interaction (Fig. 1f). Furthermore, the tumour accumulation of CD 17/ACy7 was 14.8-fold higher than that of (ZW)-CD/ACy7 (Fig. 1e), demonstrating the potential of PBA as a potent tumour-targeting moiety. Notably, (ZW)-CD/ACy7 showed a significantly lower level of tumour accumulation compared with free ACy7, suggesting that the lack of the PBA moiety resulted in (ZW)-CD promoting the clearance of cargo molecules from both tumours and normal organs. Although CD 18/ ACy7, a negatively charged inclusion complex, showed an overall high tumour-to-background ratio (TBR) comparable to that of CD 17/ACy7, the tumour distribution efficiency was 1.46-fold lower (Fig. 1e). Conversely, CD 19/ACy7 (positively charged) and CD 20/ACy7 (neutral) showed comparable tumour distribution efficiencies to CD 17/ACy7, but their TBR values were approximately 2-fold lower in all organs (Fig. 1g). Based on high TBR and tumour distribution efficiency, CD 17, hepatkis-(6-deoxy-6-((phenylboronic acid-tetraethyleneglycol-l-glutamic acid N α -sulfobetaine)-octaethyleneglycol-caproamide))-β-cyclodextrin (PBA-(ZW)-CD), was selected as an optimized renal-clearable CD derivative for further investigations.
Colorectal cancer (CRC) is the third most commonly diagnosed cancer with the second highest mortality rate 35 . Adjuvant chemotherapy (ACT) after surgical resection is the current mainstay treatment for stage II and stage III CRC 36 . However, the severe toxicity of chemotherapeutic agents often limits the effectiveness of ACT 37,38 , even with combination therapy using small doses of various drugs 39 . This situation calls for innovative modalities to improve or replace existing ACT regimens 40 . In this Article, we report the development of renal-clearable CD nanoplatforms that can promote tumour retention, without causing any off-target accumulation of imaging and therapeutic agents in CRC treatment.

Design, synthesis, and optimization of CD derivatives
Twenty CD derivatives (CD 1-20) were designed by modifying the charged moiety linked to the amine group of the glutamic acid linker (P1) and spacers between cyclodextrin and glutamic acid (P2), and glutamic acid and phenylboronic acid (PBA) (P3) (Fig. 1a). The detailed synthetic schemes are described in Methods and Supplementary Figs. 1 and 2. Syntheses of the intermediates were confirmed using proton nuclear magnetic resonance ( 1 H NMR) spectroscopy and mass spectrometry  and the introduction of PBA to CD intermediates was confirmed via the alizarin red S test (Supplementary Fig. 8). The fluorescamine assay and 1 H NMR analysis revealed that charged moieties were successfully conjugated to CD intermediates to yield CD 16-20 (Supplementary Figs. 9 and 10).
Next, we screened out CD derivatives with poor colloidal stability in phosphate-buffered saline (PBS; high ionic strength condition) or fetal bovine serum (FBS; bloodstream condition) based on turbidimetric analysis ( Supplementary Fig. 11a, summarized in Fig. 1b). Although the CD derivatives with spacers (PEG 8 or PEG 12 ) at P2 were dissolved in PBS and FBS, the A 2 module (compound 4)-conjugated CD derivative (CD 12) precipitated in both media. Thus, the A 1 module (compound 2) was selected as the zwitterionic functional group. Next, the length of the PEG spacer was extended to further improve the colloidal stability. For CD 16, PBA and A 1 were directly conjugated to glutamic acid and a long PEG 12 -caproic acid spacer was inserted at P2. In contrast, CD 17-20 were prepared by inserting a PEG 8 -caproic acid spacer at P2 and the PEG 4 spacer at P3, hypothesizing that nearby sulfobetaine could hinder the sialic acid-targeting ability of PBA. As expected, CD 16-20 were dissolved clearly in both media and showed no significant turbidity changes for 48 h, suggesting high colloidal stability after intravenous administration ( Supplementary Fig. 11b). Moreover, the hydrodynamic diameters (HDs) of CD 16-20 were of renal-clearable size (5-6.6 nm), and the zeta potential of each CD derivative corresponded to the decorated charged moieties. Various spacers (E and F n ) were introduced between the CD ring and the glutamic acid linker (P2) and glutamic acid and PBA moiety (P3). Design strategies and magnified images of CD 1-20 structures can be found in Supplementary  Information. b, Relationship between CD structure and colloidal stability. A turbidimetry-based solubility test was performed on CD derivatives dispersed in FBS or PBS at 1 mmol l -1 concentration. CD derivatives with no spacers (that is, zero length) at P2 and P3 (CD 1-5) exhibited precipitation when dispersed in PBS and FBS. The introduction of a short PEG 4 spacer at P2 was insufficient to secure colloidal stability in PBS (CD 6, 7, 9 and 10); CD 8-10 were readily dissolved in FBS, whereas CD 6 and CD 7 showed severe aggregation. These results can be attributed to the strong inter-and intramolecular interactions of the zwitterionic moieties rather than the inherent hydrophobicity of the CD molecules 74 . CD derivatives with longer spacers (PEG 8 ) at P2 (except CD 12) were dissolved in PBS and FBS. c, Physicochemical properties of five candidate structures with high colloidal stability. d, Biodistribution of ACy7 inclusion complexes prepared with CD 16-20 or (ZW)-CD. ACy7 was synthesized as a fluorescence probe for the NIRF imaging-assisted biodistribution study of CD derivatives ( Supplementary  Fig. 12a). Ex vivo NIRF images of tumours and major organs resected from HT-29 tumour-bearing mice 24 h after the intravenous injection of free ACy7 or inclusion complexes. H, heart; Lu, lungs; Li, liver; Sp, spleen; Mu, muscle; Tu, tumour; SI, small intestine; LI, large intestine; Ki, kidneys. e, TBRs in each organ. Radiant efficiency values in tumour tissues are presented in the inset. Data are presented as mean ± s.d. of four biologically independent mice per group. ACy7 versus (ZW)-CD/ACy7, **P = 0.00254; (ZW)-CD/ACy7 versus CD 17/ACy7, ****P < 0.00001; CD 16/ACy7 versus CD 17/ACy7, ****P = 0.00001; CD 17/ACy7 versus CD 18/ACy7, **P = 0.00159. f, Addition of the PEG 4 spacer in the P3 position significantly improved the tumour accumulation of zwitterionic CD derivatives, presumably reducing the interference of sulfobetaine in PBA-target interactions. g, Among four differently charged CD derivatives, only the zwitterionic CD derivative (CD 17, PBA-(ZW)-CD) exhibited enhanced tumour retention and reduced off-target accumulation simultaneously.

Tumour targetability and biodistribution of PBA-(ZW)-CD/ACy7
The structural components of PBA-(ZW)-CD and in vivo distribution profiles of PBA-(ZW)-CD/ACy are shown in Fig. 2a,b. Based on our design strategies, we hypothesized that PBA-(ZW)-CD/ACy7 could be rapidly distributed throughout the body and selectively retained in tumour tissues via PBA-sialic acid interactions. In addition, the zwitterionic sulfobetaine moiety could reduce non-specific interactions between PBA-(ZW)-CD and normal tissues; thus, the clearance of ACy7 may be accelerated via the renal route, further enhancing tumour selectivity.
Before verifying the targetability of PBA-(ZW)-CD in vivo, enhancement in the cellular uptake of PBA-(ZW)-CD/ACy7 via PBA-sialic acid interaction was confirmed in HT-29 cells by fluorescence microscopy ( Supplementary Fig. 14a). The specificity of PBA-(ZW)-CD binding to cancer cells overexpressing sialic acid was further confirmed in four cell lines (HT-29, HCT-116, NIH3T3 and MDCK) with varying levels of sialic acid expression ( Supplementary Fig. 14b,c).
The tumour-targeting ability of PBA-(ZW)-CD/ACy7 was investigated in HT-29 tumour xenograft mice using real-time near-infrared fluorescence (NIRF) imaging (Fig. 2c). In the early phase (~2 h), intravenously administered ACy7 and PBA-(ZW)-CD/ACy7 were both distributed throughout the body with no tumour selectivity. However, after rapid renal excretion of PBA-(ZW)-CD, which was verified in a preliminary study with sulfocyanine 7-conjugated PBA-(ZW)-CD ( Supplementary Fig. 15), the PBA-(ZW)-CD/ACy7 inclusion complex-treated mice showed distinct fluorescence signals in the tumour and kidney regions in whole-body imaging (Fig. 2c). Furthermore, the TBR of PBA-(ZW)-CD/ACy7 exceeded 2, the standard cut-off value for tumour detection 43 , at 8 h post-injection and continuously increased for the next 16 h, whereas free ACy7 showed TBR values less than 2 throughout the experimental period (Fig. 2d). The high TBR of PBA-(ZW)-CD/ACy7 can be attributed to prolonged tumour retention and low background tissue retention, as observed in the biodistribution study (Fig. 1d). From a clinical perspective, determining the tumour margin is of paramount importance for the accurate surgical resection of CRC, especially in laparoscopic or robot-assisted operations [43][44][45] .
To evaluate the applicability of PBA-(ZW)-CD/ACy7 in image-guided surgery, PBA-(ZW)-CD/ACy7 was intravenously injected into an orthotopic CRC model. At 24 h post-injection, PBA-(ZW)-CD/ACy7 showed high TBR values of approximately 4 against adjacent normal tissues, including the small intestine, colon and muscle, providing accurate tumour margins for resection (Fig. 2e,f). The cumulative urinary excretion of PBA-(ZW)-CD/ACy7 was 2.3-fold higher than that of free ACy7, which suggests that PBA-(ZW)-CD may facilitate the renal excretion of cargo molecules (Fig. 2g). The shorter elimination half-life (t 1/2β ) of PBA-(ZW)-CD/ACy7 observed in pharmacokinetic study implies that PBA-(ZW)-CD reduced non-specific binding of ACy7 in normal tissues ( Supplementary Fig. 16), also supporting low background tissue retention of PBA-(ZW)-CD.
PBA-(ZW)-CD/DOX had a mean HD of 13.04 ± 3.85 nm at pH 7.4 (Fig. 3e). The slight increase in HD compared with that of the blank PBA-(ZW)-CD could be attributed to concentration-dependent (that is, reversible) coordination between the primary amine of DOX and boronic acid of PBA-(ZW)-CD 53 . As the coordination weakens at acidic pH, the mean HD of PBA-(ZW)-CD/DOX decreased to 7.09 ± 0.26 nm at pH 5.5, which is almost the same value of blank PBA-(ZW)-CD HD (6.58 ± 0.27 nm). However, PBA-(ZW)-CD/UXT showed comparable HD distributions at pH 5.5 and pH 7.4, showing mean HDs of 6.30 ± 0.11 nm and 6.30 ± 0.09 nm, respectively, which implies a lesser influence of acidity. PBA-(ZW)-CD/drugs showed pH-dependent release profiles, which could translate into selective drug exposure to tumour tissues, considering the acidic tumour microenvironment (Fig. 3f). The zeta potential of both inclusion complexes increased at acidic pH due to the ionization of the guest drug while maintaining zwitterionic properties at pH 7.4 (Fig. 3e).

Biodistribution of PBA-(ZW)-CD/drug inclusion complexes
The biodistribution of PBA-(ZW)-CD/drug inclusion complexes was assessed in subcutaneous HT-29 tumour xenograft mice in comparison with that of free drugs. The off-target exposure of DOX to normal tissues was significantly reduced in the PBA-(ZW)-CD/DOX-treated group compared with the DOX solution group during the distribution phase (Fig. 4a). Considering that cardiotoxicity and skeletal muscle atrophy are the major side effects of DOX that limit its application to CRC 54-57 , the improved biodistribution profile via PBA-(ZW)-CD complexation could enhance the chances of future clinical translation. Despite the observed reduction in DOX distribution in heart and muscle by PBA-(ZW)-CD, the levels of DOX in the lung and kidney remained relatively high. This outcome could be a result of the highly adsorptive property and tissue-selective affinity of DOX [58][59][60] . During circulation, DOX released from PBA-(ZW)-CD may have primarily distributed to the lung and kidney, as these organs show relatively high tissue-to-plasma partition coefficients (that is, K p values) compared with other tissues 61 .
In the tumour tissue, although both free DOX and PBA-(ZW)-CD/DOX showed similar exposure at 0.5 h, PBA-(ZW)-CD/DOX had prolonged tumour retention compared with free DOX, resulting in 2.0-fold higher DOX accumulation at 8 h (P < 0.005).
Unlike DOX, which non-specifically binds to various biomacromolecules, UXT showed the expected biodistribution profile when formulated as PBA-(ZW)-CD/UXT (Fig. 4b). PBA-(ZW)-CD/UXT showed rapid clearance in the normal tissues, with UXT amounts remaining in the heart, lungs, kidneys and liver less than 0.65 μg per g tissue within 2 h, which were 8.0-, 17.0-, 5.6-and 13.6-fold lower values than those of free UXT solution, respectively. In the tumour tissues, however, PBA-(ZW)-CD/UXT showed a 2.1-fold higher accumulation of UXT than free UXT at 8 h post-injection (P < 0.05). These findings clearly indicate that PBA-(ZW)-CD can enhance the tumour retention of guest drugs and reduce their off-target accumulation in normal tissues.
Homogeneous drug distribution throughout the tumour mass is also of importance to enhance therapeutic efficacy of drugs [62][63][64] . PBA-(ZW)-CD/drug inclusion complexes could evade reduced convection issue of conventional NCs given their ultrasmall size and zwitterionic surface charge [65][66][67][68] . As observed in the ex vivo matrix-assisted laser desorption ionization-mass spectrometry imaging (MALDI-MSI) (Fig. 4c), the intratumoural DOX (m/z 544.6) and UXT (m/z 434.7) signals in the inclusion complex-treated group showed a comparable spatial intensity distribution (as the coefficient of variation) to those of the solution-treated group, suggesting that PBA-(ZW)-CD/drug was as highly tumour-penetrable as a free drug solution (Fig. 4d). In addition, the average intensity values of both drugs were significantly higher in the PBA-(ZW)-CD/drug groups than in the corresponding solution groups (Fig. 4e).
The unique pharmacokinetic properties of PBA-(ZW)-CD, which are rarely observed in other clinical reports 69 , are summarized as follows: (1) PBA-(ZW)-CD may control the biodistribution of cargo molecules in an opposite manner depending on the deposition site, showing decreased and increased clearance in tumour and normal tissues, respectively; and (2) PBA-(ZW)-CD/drug inclusion complexes may penetrate the tumour microenvironment as readily as free drug solutions, promoting homogeneous drug delivery throughout the tumour mass.

PBA-(ZW)-CD/drug-assisted combination therapy
Encouraged by the remarkable biodistribution profiles and negligible toxicity of PBA-(ZW)-CD ( Supplementary Fig. 17a,b), the in vivo antitumour efficacy of each PBA-(ZW)-CD/drug inclusion complex was evaluated in the heterotopic HT-29 xenograft mouse model. PBA-(ZW)-CD/ drug-assisted single-drug therapies (using DOX or UXT) showed a significant improvement in tumour growth inhibition at the therapeutic dose (5 mg per kg) compared with the drug solution group (Supplementary Figs. 18 and 20). Notably, PBA-(ZW)-CD significantly reduced DOX toxicity to normal organs and immune cells ( Supplementary Fig. 19), which corresponded well with the improved biodistribution profiles of PBA-(ZW)-CD/DOX. As both inclusion complexes had marginally higher cytotoxicity than the corresponding free drug ( Supplementary  Fig. 21), their highly enhanced antitumour efficacy in vivo could mainly result from improved biodistribution.
Next, we performed a combination therapy with PBA-(ZW)-CD/ DOX and PBA-(ZW)-CD/UXT in the same mouse model. As observed in the tumour growth profile (Fig. 5a), the combination of DOX and UXT solutions showed enhanced antitumour efficacy on average when compared with monotherapy with each drug, but no statistically significant difference was found between the DOX + UXT solution and DOX solution groups (P > 0.05). However, the combination of inclusion complexes showed markedly improved tumour growth inhibition compared with the other groups, of which the average tumour volume on day 14 was 2.1-fold smaller than that in the DOX + UXT solution group (P < 0.05). A similar tendency was observed in the dissected tumour weight data (Fig. 5b), where the PBA-(ZW)-CD/DOX + PBA-(ZW)-CD/ UXT group showed the lowest tumour weight compared with the other groups, showing a 2.9-fold lower average tumour weight on day 14 compared with that of the DOX + UXT solution group (P < 0.05). The superior antitumour efficacy of the PBA-(ZW)-CD/drug combination could be explained by the improved drug distribution to the tumour tissue spatially and temporally enough to induce synergistic effects between DOX and UXT. Meanwhile, the mice in all groups did not lose body weight significantly during the treatment (Fig. 5c) and showed no abnormal changes in blood biochemistry ( Supplementary Fig. 22), indicating negligible systemic toxicity. Data are presented as mean ± s.d. of three biologically independent mice per group. Statistical significance is calculated via a two-tailed Student's t-test. At 0.5 h: heart, ****P = 0.00016; lung, **P = 0.0098; kidney, ***P = 0.0020. At 2 h: heart, ****P = 0.00053; lung, ***P = 0.0046; kidney, ***P = 0.00042. The amount of DOX in the heart 2 h after administration was 6.5-fold lower in the inclusion complex group compared with the free drug group (P < 0.001). In muscle tissue, DOX was not detected (<200 ng per g tissue) in the inclusion complex group even at 0.5 h post-injection, while the free DOX group exhibited 2.73 ± 0.30 μg per g tissue. In tumour: at 2 h, *P = 0.0206; at 8 h, ***P = 0.0047. a Below the detection limit (<200 ng per g tissue). b, UXT distribution in normal tissues, including mononuclear phagocyte system-rich organs, and tumour after the intravenous administration of UXT solution and PBA-(ZW)-CD/UXT at a UXT dose of 5 mg per kg. Data are presented as mean ± s.d. of three biologically independent mice per group. Statistical significance is calculated via a two-tailed Student's t-test. In all tissues, ****P < 0.0001. At 0.5 h: heart, *P = 0.0222; lung, *P = 0.0156. At 2 h: liver, ***P = 0.0018. At 8 h: tumour, *P = 0.0263. c, MALDI-MSI study of tumour tissues dissected at 30 min post-injection of drug solution or PBA-(ZW)-CD/drug inclusion complex (n = 3). Intratumoural mass spectrometry signals corresponding to DOX (m/z 544.6, pseudocoloured in red) and UXT (m/z 434.7, pseudocoloured in green) are presented. Scale bar, 1 mm. d,e, The spatial signal distributions as coefficient of variation (mean ± s.d., n = 3 biologically independent samples) (d) and area-averaged signal intensity (mean ± s.d., n = 3 biologically independent samples) (e) were also evaluated via image analysis. Data are presented as mean ± s.d. of three biologically independent samples per group. Statistical significance is calculated via a two-tailed Student's t-test.

PBA-(ZW)-CD/drug combination therapy in orthotopic CRC model
The antitumour efficacy of the PBA-(ZW)-CD/drug-assisted combination therapy was evaluated in an orthotopic CRC model ( Supplementary  Fig. 23). Bioluminescence imaging revealed that both solution-and inclusion complex-based combination therapies significantly reduced tumour growth compared with no intervention (Fig. 5d). Notably, PBA-(ZW)-CD/drug combination therapy had a much higher tumour inhibitory efficacy than solution therapy, showing a 4.2-fold lower bioluminescence intensity on day 14 (P < 0.005) (Fig. 5e). In addition, a decreased frequency of multiple signal foci was observed in the group receiving the combination of PBA-(ZW)-CD/drug inclusion complexes, compared with the other groups ( Fig. 5d and Supplementary Fig. 24), suggesting a reduced development of peritoneal metastasis [70][71][72][73] .
Unlike the heterotopic model prepared by the subcutaneous injection of HT-29 cell suspension, the orthotopic CRC model prepared via laparotomy showed lower tolerability to solution-based combination therapy, with the average body weight reduced to 88.9% of the non-intervention (PBS) group (P < 0.01) (Fig. 5f) group, implying that PBA-(ZW)-CD reduced the systemic toxicity of ACT after surgical resection. At the end of the efficacy test, no detrimental changes were observed in normal tissues (Fig. 5g), indicating that PBA-(ZW)-CD can minimize the dose-limiting toxicity of anticancer agents. Terminal deoxynucleotidyl transferase dUTP nick-end labelling staining was also performed on the tumour tissues (Fig. 5g), which revealed that PBA-(ZW)-CD/drug combination therapy induced a higher degree of apoptosis than solution-based combination therapy. Statistically significant differences were determined using ANOVA and Tukey's test. ***P = 0.0022, ****P < 0.0001. f, The body weight change of each mouse was recorded (mean ± s.d., n = 5). Statistically significant differences were determined using ANOVA and Tukey's test. **P = 0.0043. g, Hematoxylin and eosin (tumour and organs) and terminal deoxynucleotidyl transferase dUTP nick-end labelling (TUNEL, tumour) staining images of tissues dissected on day 14. Scale bar, 200 μm. The experiment was repeated three times and comparable results were obtained.
Article https://doi.org/10.1038/s41565-023-01381-8 The toxicity of PBA-(ZW)-CD/drug combination was evaluated in the orthotopic CRC model after applying the combination regimen as used in the antitumour efficacy study. As expected from the toxicity study of PBA-(ZW)-CD/DOX monotherapy, PBA-(ZW)-CD significantly reduced the organ toxicity, inflammatory responses and systemic immunosuppression associated with ACT ( Supplementary Fig. 25).

Conclusions
We present a delicately tailored design of renal-clearable zwitterionic CD, namely, PBA-(ZW)-CD, for CRC-selective imaging and drug delivery. Twenty CD derivatives were prepared by modifying the charged moieties and spacers to improve their colloidal stability and in vivo pharmacokinetic behaviour. CRC targetability, organ distribution and renal clearance of CD derivatives were screened in the form of ACy7 inclusion complexes, from which PBA-(ZW)-CD was selected as the optimized structure with enhanced tumour selectivity and reduced off-target accumulation. Given its high biocompatibility, PBA-(ZW)-CD/ drug inclusion complexes of DOX and UXT were fabricated and applied in combination therapy for CRC. Notably, PBA-(ZW)-CD enhanced the tumour retention of drugs, while facilitating their elimination in normal tissues. In addition, the improved antitumour efficacies of PBA-(ZW)-CD/drug-assisted single-drug and combination therapies were demonstrated in various CRC models. Therefore, PBA-(ZW)-CD may be used as a promising CRC-targeting nanoplatform with high potential for clinical translation.

Online content
Any methods, additional references, Nature Portfolio reporting summaries, source data, extended data, supplementary information, acknowledgements, peer review information; details of author contributions and competing interests; and statements of data and code availability are available at https://doi.org/10.1038/s41565-023-01381-8.
Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
Synthesis of compounds 13-20. Compounds 13 and 14 were prepared by conjugating heptakis (6-deoxy-6-amino)-β-cyclodextrin heptahydrochloride and t-Boc-N-amido-PEG n -acid (n = 4 and n = 8, respectively), using the general procedure for HBTU coupling. Boc groups of 13 and 14 were removed according to the general procedure for Boc deprotection of CD intermediates to yield TFA salts of 15 and 16, respectively, which were used in the next step without further purification. Compounds 17 and 18 were prepared by conjugating 6 with 15 and 16, respectively, using the general procedure for HBTU amide coupling. Boc groups of 17 and 18 were deprotected using the method described above, resulting in compounds 19 and 20, respectively.

Synthesis of compounds 21-26.
Compound 21 was prepared by conjugating heptakis (6-deoxy-6-amino)-β-cyclodextrin heptahydrochloride and 6-(Boc-amino)hexanoic acid, using the general procedure for HBTU amide coupling. Then, 22 was obtained by deprotecting the Boc groups of 21, following the general procedure for Boc deprotection of the CD intermediates. t-Boc-N-amido-PEG n -acid (n = 8 and n = 12) was conjugated to 22 using the general procedure for HBTU amide coupling to yield 23 and 24. Compounds 25 and 26 were prepared by the removal of Boc groups from 23 and 24, respectively, according to the general procedure for Boc deprotection of CD intermediates.
Then, the Boc groups of 27 and 29 were removed using the general procedure for Boc deprotection of CD intermediates to yield 28 and 30, respectively.
Turbidimetric analysis of CD derivatives. The colloidal stability of CD derivatives in PBS or FBS was evaluated using a turbidity assay. Briefly, each CD derivative powder sample was added to PBS or FBS to prepare a 1 mmol l -1 solution. The degree of aggregation was evaluated by measuring the absorbance at 600 nm (OD 600 ) using a ultraviolet-visible (UV-vis) spectrophotometer (EMax Precision Microplate Reader, Molecular Devices) 79 . The OD 600 values of the PBS and FBS solutions were normalized to that of the DDW/DMSO (50:50, v/v) solution of the corresponding CD derivatives to correct the molecular absorbance.

Residual amines after charged moiety conjugation.
Each CD derivative solution in PBS (1 mg ml -1 , 200 μl) was vortex-mixed with the same volume of fluorescamine solution in DMSO (2.5 mg ml -1 ) for 1 h. l-Serine dissolved in PBS (5-100 nmol ml -1 ) was prepared as a standard sample. Similarly, each l-serine solution was vortex-mixed with the same volume of fluorescamine solution for 1 h. Fluorescence intensity of each sample was measured using a SpectraMax M5 multimode microplate reader (λ ex /λ em = 395 nm/495 nm; Molecular Devices).

Biodistribution studies of CD derivatives
Preparation of NIRF dye-loaded inclusion complexes. ACy7 was synthesized by adding 1-adamantylamine (1.1 mg, 0.007 mmol) and TEA (1 μl, 0.007 mmol) to SCy7-NHS (5 mg, 0.0056 mmol) dissolved in anhydrous DMSO. The mixture was monitored by ESI-MS and stirred at r.t. until SCy7-NHS was completely consumed, followed by 48 h of lyophilization to remove excess volatile impurities including Nature Nanotechnology Article https://doi.org/10.1038/s41565-023-01381-8 1-adamantylamine and TEA. The lyophilized product was washed with acetonitrile and dried under vacuum to yield ACy7 (82% yield). ACy5 was synthesized using the same method as described above, except that SCy5-NHS was used. To prepare ACy7 inclusion complexes, CD derivatives (1 μmol) and ACy7 (1.2 μmol) were agitated in DDW for 24 h, followed by dialysis against DDW in a dialysis membrane (MWCO 3.5 kDa; Cellu·Sep, Membrane Filtration Products) for 24 h to remove excess ACy7. The dialysis product was lyophilized to obtain the CD/ ACy7 powder. The absorbance of the CD/ACy7 solution was measured at 740 nm to calculate ACy7 content.
The expression level of sialic acid in the cell suspensions of HT-29, HCT-116, NIH3T3 and MDCK was determined using the Sialic Acid Assay Kit (MAK314; Sigma-Aldrich) as per the manufacturer's protocol. The total protein concentration of each cell suspension was quantified using the BCA protein assay kit (Thermo Fisher Scientific) after lysis with RIPA buffer containing protease and phosphatase inhibitors. The expression of sialic acid in each cell line was expressed as relative sialic acid content (μg of sialic acid per 1 mg of protein).
For flow cytometry analysis, HT-29, HCT-116, NIH3T3 and MDCK cells were seeded in 6-well plates at a density of 1 × 10 6 cells per well, with the same treatment schedule as that used in the fluorescence microscopy assay. To assess the interaction between PBA and sialic acid, ACy5, PBA-(ZW)-CD/ACy5, (ZW)-CD/ACy5 and CD 16/ACy5 were compared, and free PBA (500 μmol l -1 ) or sialic acid (500 μmol l -1 ) was used as competitive inhibitors. The competitive inhibitor was incubated with the cells for 15 min before PBA-(ZW)-CD/ACy5 treatment. The ACy5 fluorescence signal of the cells (10,000 events within the total population) was analysed using FACSCalibur (BD Biosciences).
Animal models. Male BALB/c nude mice (5 weeks old; Nara Biotech) were maintained in a light-controlled room at a temperature of 22 ± 2 °C and relative humidity of 55 ± 5% (Animal Center for Pharmaceutical Research, College of Pharmacy, Seoul National University, Seoul, Republic of Korea). The animal study protocol (SNU-200630-1-1) was approved by the Animal Care and Use Committee of Seoul National University. In this study, HT-29 was utilized because this cell line represents the traits of human CRC, including the high expression of sialyl Lewis A, a well-known target of PBA 80 . A heterotopic HT-29 tumour xenograft mouse model was established by injecting HT-29 cells (2 × 10 6 cells per mouse) into the right hind limb region of the mice. HT-29 cells (Korean Cell Line Bank) were cultured using the method described above. The tumour volume (V, mm 3 ) was calculated using the following formula: V (mm 3 ) = 0.5 × longest diameter × (shortest diameter) 2 . The orthotopic xenograft model was established according to a previously reported method 81 . Briefly, under isoflurane anaesthesia, the peritoneal cavity of the mice was opened carefully to expose the caecum. HT-29 cell suspension (2.0 × 10 6 cells, 20 μl) was inoculated into the subserosal layer of the caecum walls. After inoculation, continuous suturing was performed to close the peritoneum and abdominal wall. Further details regarding the surgical procedure and model validation can be found in Supplementary Fig. 23.

NIRF imaging studies of CD/ACy7 inclusion complexes in HT-29
tumour xenograft model. NIRF imaging was performed using a heterotopic xenograft model. After the tumour volume exceeded 200 mm 3 , free ACy7 or ACy7 inclusion complexes of CD 16-20 dissolved in normal saline were injected into the tail vein of mice at an ACy7 dose of 1 μmol per kg. Whole-body scanning was performed at 0.17, 0.5, 1, 2, 4, 8 and 24 h post-injection using an IVIS Spectrum In Vivo Imaging System (PerkinElmer), followed by ex vivo imaging of the major organs and tumour tissues using the same instrument. Ex vivo imaging studies were performed using an orthotopic xenograft model. Four weeks after inoculation, the same dose of free ACy7 or ACy7 inclusion complexes of (ZW)-CD or CD 16-20 was administered intravenously. The mice were killed 24 h post-injection, from which the muscle, small intestine, colon and caecum with tumours were dissected for ex vivo NIRF imaging.

Pharmacokinetic evaluation of CD/ACy7 inclusion complexes.
BALB/c nude mice (5 weeks old; Nara Biotech) were intravenously injected with ACy7 or PBA-(ZW)-CD/ACy7 (as ACy7, 1 μmol per kg). Blood samples (~50 μl) were collected from the retro-orbital vein using heparinized capillary tubes (Fisher Scientific) at 1, 5, 15, 30, 60, 120 and 240 min post-injection under isoflurane anaesthesia (staggered sampling). Blood samples were centrifuged at 16,000 × g for 5 min, and plasma samples (supernatant, 20 μl) were transferred to microsample tubes. The fluorescence intensities of the plasma samples were measured using an NIRF imaging system (VISQUE InVivo Smart-LF equipped with sCMOS Image Sensor, Vieworks). The pharmacokinetic parameters of ACy7 were calculated by fitting a two-compartment model using WinNonlin software (version 5.0.1, Pharsight). For the evaluation of urinary excretion, ACy7 or PBA-(ZW)-CD/ACy7 was injected using the same protocol described above. Urine samples were collected and weighed at 4 h and 8 h post-injection, followed by centrifugation at 16,000 × g for 5 min. Aliquots of the supernatant (100 μl) were loaded into a 96-well plate, and ACy7 absorbance was measured at 740 nm using a UV-vis spectrophotometer (EMax Precision Microplate Reader, Molecular Devices).

Preparation of PBA-(ZW)-CD/Drug Inclusion Complexes.
PBA-(ZW)-CD/drug inclusion complexes were fabricated using a freeze-drying method 82 . As both drugs were poorly soluble in aqueous media, a co-solvent of DMSO and water was used to solubilize the drugs with PBA-(ZW)-CD to induce host-guest interactions. The tubes were immersed in release media (PBS; pH 5.5 and 7.4; 10 ml) and agitated at 50 rpm at 37 °C. Aliquots (200 μl) of the release medium were collected at 2, 4, 6, 8, 12 and 24 h, and the same volume of fresh medium was replenished at each time point. The amount of drug released was determined using high-performance liquid chromatography (HPLC). The amount of DOX in the sample was quantified using a Waters HPLC system equipped with a reverse-phase C18 column (Gemini, 250 × 4.6 mm, 3 μm; Phenomenex), a separation module (Waters e2695) and a fluorescence detector (Waters 2475). The mobile phase consisted of 10 mmol l −1 phosphate buffer (pH 2.5, adjusted with phosphoric acid) and acetonitrile with 0.1% TEA (70:30, v/v). The excitation and emission wavelengths were 470 nm and 565 nm, respectively. The injection volume and flow rate were set at 20 μl and 0.8 ml min -1 , respectively. The DOX retention time was 7.3 min. The amount of UXT in the sample was determined using a Waters HPLC system equipped with a Fortis C18 column (250 × 4.6 mm, 5 μm; Fortis Technologies), autosampler (Waters 717), binary pump (Waters 1525) and UV detector (Waters 2487). The mobile phase was composed of a phosphate buffer (pH 2.5, adjusted with phosphoric acid) and acetonitrile (60:40, v/v). The eluent was monitored at an absorption wavelength of 255 nm. The injection volume and flow rate were 20 μl and 1 ml min -1 , respectively. The UXT retention time was 6.5 min.
Biodistribution in heterotopic HT-29 xenograft model. The heterotopic HT-29 xenograft mouse model was prepared as described in the 'Biodistribution studies of CD derivatives' section. When the tumour volume reached approximately 200 mm 3 , the mice were randomly divided into four groups (DOX, UXT, PBA-(ZW)-CD/DOX and PBA-(ZW)-CD/UXT) and intravenously administered each intervention at a dose of 5 mg per kg. The mice were killed at 0.5, 2 and 8 h post-injection, and tumour and normal tissues were excised and weighed, followed by tissue homogenization in 1% Triton X-100 (25%, w/v). DOX or UXT was extracted from tissue homogenate (50 μl) using acidified isopropanol (200 μl, containing 0.5 mol l -1 HCl) or acetonitrile (200 μl, containing 100 ng ml -1 carbamazepine as an internal standard), respectively. The mixture was vortexed for 15 min and centrifuged at 16,000 × g for 5 min at 4 °C. The amount of DOX in the tissue homogenate was quantified using the HPLC method described in 'In vitro drug release'. The UXT amount was determined by using an LC-MS/MS system equipped with a reverse-phase C18 column (Synergi Max-RP 80 Å, 75 × 4.6 mm, 4 μm; Phenomenex), a 1260 Infinity HPLC system (Agilent Technologies) and 6430 Triple Quad LC/MS system (G6430A, Agilent Technologies), and the gas temperature, gas flow rate, nebulizer pressure and capillary voltage were manually optimized to be 300 °C, 11 l min -1 , 15 psi and 4,000 V, respectively. The mobile phase consisted of 10 mmol l -1 ammonium formate buffer and acetonitrile (20:80, v/v). The m/z values of precursor/product ions in multiple reaction monitoring mode were set at 433.1/262.1 for UXT, and 237.0/194.0 for carbamazepine. The fragmentor voltage, collision energy and cell accelerator voltage for UXT and carbamazepine were set to 89 V/15 eV/2 V and 126 V/18 eV/2 V, respectively. The retention times of UXT and carbamazepine were 2.3 min and 2.1 min, respectively.
Ex vivo MALDI-MSI. Intratumoural drug distribution was assessed using ex vivo MALDI-MSI. When the tumour volume reached approximately 200 mm 3 , DOX, UXT and PBA-(ZW)-CD/DOX (10 mg per kg drug) were intravenously injected into HT-29 tumour-bearing mice. Mice were killed at 30 min, and the resected tumours were immediately frozen and sectioned at 10 μm thickness (CM3050S; Leica Biosystems). Tissue sections were placed onto indium tin oxide-coated glass slides and coated with 10 mg ml -1 DHB in a co-solvent of acetonitrile and DDW with 0.1% TFA (1:1, v/v). MS imaging of the tumour tissues was performed using a RapifleX MALDI Tissuetyper (Bruker Daltonics).
In vitro cytotoxicity. HT-29 cells were cultured as described in the 'Biodistribution studies of CD derivatives' section. HT-29 cells were seeded in a 96-well plate at a density of 5.0 × 10 3 cells per well and incubated for 24 h at 37 °C. The free drug or PBA-(ZW)-CD/drug inclusion complex was treated at various concentrations (0.005-5 μmol l -1 and 0.005-50 μmol l -1 for DOX and UXT, respectively), and cell viability was measured after 48 h of incubation using a colorimetric method (CellTiter 96 Aqueous One Solution Cell Proliferation Assay, Promega). Similarly, the cytotoxicity of blank PBA-(ZW)-CD was evaluated in HT-29 cells at PBA-(ZW)-CD concentrations of 1-500 μg ml -1 . After incubation for 24 h, 48 h and 72 h, cell viability was assessed using the method described above.
Complete blood count. Male BALB/c mice (5 weeks old; Nara Biotech) were intravenously injected with an excess amount of PBA-(ZW)-CD (270 mg per kg; >5-fold dose) once or three times daily. Blood samples (0.5 ml) were collected 24 h post-injection, and a complete blood count test was performed using Scil Vet abc Plus (HORIBA). The white blood cell count, percentage of lymphocytes, monocytes and eosinophils, red blood cell count, haematocrit, mean corpuscular volume, platelet count and mean platelet volume were recorded.
Single-drug therapy in heterotopic HT-29 xenograft model. The antitumour efficacy of PBA-(ZW)-CD/drug inclusion complexes was evaluated in subcutaneous HT-29 tumour xenograft mice, which were prepared according to the protocol described in the 'Biodistribution studies of CD derivatives' section. For DOX monotherapy, the mice were randomized into four groups: PBS, PBA-(ZW)-CD (blank carrier), DOX solution and PBA-(ZW)-CD/DOX, and intravenously administered each intervention at a dose of 5 mg per kg every three days. The tumour volume and body weight were monitored every other day. The efficacy test for UXT monotherapy was performed using the same protocol, except that UXT solution and PBA-(ZW)-CD/UXT were adopted.
Combination therapy in heterotopic HT-29 xenograft model. Subcutaneous HT-29 tumour-bearing mice were prepared as described in the 'Biodistribution studies of CD derivatives' section. To evaluate the antitumour efficacy of the combination therapy, the mice were randomly divided into six groups: PBS, PBA-(ZW)-CD (blank carrier), DOX solution (3 mg per kg), UXT solution (3 mg per kg), DOX + UXT solutions (3 mg per kg each), and PBA-(ZW)-CD/DOX + PBA-(ZW)-CD/UXT (as a drug, 3 mg per kg each), and each intervention every three days. The tumour volume and body weight were recorded every other day.
Evaluation of therapeutic efficacy in orthotopic colorectal tumour model. To establish an orthotopic colorectal tumour model, luciferase-expressing HT-29 (HT-29/Luc) (2.0 × 10 6 cells, 20 μl) were inoculated into the caecum of male BALB/c nude mice (5 weeks old; Nara Biotech) under isoflurane anaesthesia. Seven days after HT-29/Luc inoculation, d-luciferin (150 mg per kg) was injected intraperitoneally, and bioluminescence was measured with the IVIS Spectrum In Vivo Imaging System (PerkinElmer) to confirm tumour formation in the caecum region. The mice were randomly divided into three groups and intravenously administered PBS, DOX + UXT solutions (3 mg per kg for each), or PBA-(ZW)-CD/DOX + PBA-(ZW)-CD/UXT (as a drug, 3 mg per kg for each). Tumour growth (via bioluminescence) and body weight were measured on days 0, 4, 8, 11 and 14. On day 14, the organs and tissues of interest, including the tumour, heart, lungs, liver, kidneys and spleen, were dissected and stained with haematoxylin and eosin. Tumour tissues were subjected to terminal deoxynucleotidyl transferase dUTP nick-end labelling staining. Tissue images were obtained using Vectra (v3.0.5).

Blood biochemistry analyses.
On the last day of the antitumour efficacy studies, blood samples (0.5 ml) were collected from the left ventricle and centrifuged at 16,000 × g for 5 min to obtain plasma Nature Nanotechnology Article https://doi.org/10.1038/s41565-023-01381-8 samples. The levels of aspartate transaminase, alanine transaminase, alkaline phosphatase, blood urea nitrogen, creatinine, total protein, albumin, creatine kinase, creatine kinase MB isoenzyme and lactate dehydrogenase were analysed using Fuji Dri-Chem 3500 s (Fujifilm Holdings). The level of interleukin-6, the inflammatory cytokine, was determined using enzyme-linked immunosorbent assay kit (RAB0308; Sigma-Aldrich) according to the manufacturer's protocol.

Reporting summary
Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.
Corresponding author(s): Jae-Young Lee & Dae-Duk Kim Last updated by author(s): Mar 15, 2023 Reporting Summary Nature Portfolio wishes to improve the reproducibility of the work that we publish. This form provides structure for consistency and transparency in reporting. For further information on Nature Portfolio policies, see our Editorial Policies and the Editorial Policy Checklist.

Statistics
For all statistical analyses, confirm that the following items are present in the figure legend, table legend, main text, or Methods section.

n/a Confirmed
The exact sample size (n) for each experimental group/condition, given as a discrete number and unit of measurement A statement on whether measurements were taken from distinct samples or whether the same sample was measured repeatedly The statistical test(s) used AND whether they are one-or two-sided Only common tests should be described solely by name; describe more complex techniques in the Methods section.
A description of all covariates tested A description of any assumptions or corrections, such as tests of normality and adjustment for multiple comparisons A full description of the statistical parameters including central tendency (e.g. means) or other basic estimates (e.g. regression coefficient) AND variation (e.g. standard deviation) or associated estimates of uncertainty (e.g. confidence intervals) For null hypothesis testing, the test statistic (e.g. F, t, r) with confidence intervals, effect sizes, degrees of freedom and P value noted

Software and code
Policy information about availability of computer code Data collection In vivo and ex vivo images were acquired using the PerkinElmer IVIS Spectrum Optical Imaging System. NMR spectra were obtained using a nature portfolio | reporting summary

Authentication
The cell lines were authenticated by periodic morphology checks using a microscope.

Mycoplasma contamination
All cell lines were regularly tested for mycoplasma contamination, and no evidence of mycoplasma contamination was found.

Commonly misidentified lines (See ICLAC register)
No commonly misidentified cell lines were used in this study.

Animals and other research organisms
Policy information about studies involving animals; ARRIVE guidelines recommended for reporting animal research, and Sex and Gender in Research Laboratory animals All animals used in this study were male and 5 weeks old at the start of the experiments. The Balb/c and Balb/c nude strains were used. The mice were all purchased from Nara Biotech (Seoul, Republic of Korea). The mice were housed in a polycarbonate cage with a controlled environment of temperature (20-25 °C) and humidity (40-45%), with a 12:12 h light/dark cycle.

Wild animals
No wild animals were used in this study.

Reporting on sex
The sex of the animals was not considered in this study.
Field-collected samples The study did not involve samples collected from the field.

Ethics oversight
All animals were maintained and used in accordance with the Guidelines for the Care and Use of Laboratory Animals of the Institute of Laboratory Animal Resources, Seoul National University (Seoul, Republic of Korea; approved animal experimental protocol number: SNU-200630-1-1).
Note that full information on the approval of the study protocol must also be provided in the manuscript.

Flow Cytometry
Plots Confirm that: The axis labels state the marker and fluorochrome used (e.g. CD4-FITC).
The axis scales are clearly visible. Include numbers along axes only for bottom left plot of group (a 'group' is an analysis of identical markers).
All plots are contour plots with outliers or pseudocolor plots.
A numerical value for number of cells or percentage (with statistics) is provided.