Synthesis and evaluation of highly releasable and structurally stable antibody-SN-38-conjugates

Abstract Camptothecins, traditional chemotherapy drugs, have been clinically used in antibody-drug conjugates (ADCs), which refreshes the recognition that ADCs preferably incorporate highly potent payloads. However, SN-38, active metabolite of irinotecan from camptothecins, tended to be incorporated into ADCs with an unstable acid sensitive bond, not with the widely used Cathepsin B (CTSB) sensitive bond, which may pose the risk of off-target. Herein, we reported a novel strategy to construct highly releasable and structurally stable SN-38-conjugates, in which CTSB linkers directly connected to the 10-OH group through ether bond, not to the common 20-OH group of lactones of SN-38. In this paper, rapid release of SN-38 was skillfully demonstrated by utilizing the fluorescence properties of SN-38. The SN-38-ether-ADC displayed highly stable serum stability with the half-life over 10 days. Moreover, the drug-antibody-ratio (DAR) of ADC could be elevated to 7.1 through the introduction of polyethylene glycol (PEG) moieties without aggregation. The optimized ADC exhibited potent in vitro activities up to 5.5 nM, comparable to SN-38. Moreover, this ADC group significantly delayed tumor growth in vivo. In conclusion, the novel strategy has the potential to promote the development of SN38-ADCs and enrich the conjugation approaches for hydroxyl-bearing payloads.

However, of note that there were relatively few reports about SN-38 conjugates. Several scientific issues around two connection sites of SN-38 still need to be addressed. These ADCs adopting 20-OH group of SN-38 as connection site tended to show the instability in serum while others choosing 10-OH group could not release SN-38 in tumor cells timely. For example, IMMU-132 adopted the 20-OH group of SN-38 as the connection site with an acid-sensitive carbonate-based linker, which enhanced the stability of lactone to a great extent (Goldenberg et al., 2015). However, the ester bond connection strategy tended to cause instability of IMMU-132 (t 1/2 ¼ 23.98 h in human serum) and may pose the risk of off-target. Several attempts have been done for the improvement of the stability of SN-38 based ADCs. For example, CL2E-SN-38 adopted dipeptide-linker, such as valine-citrulline (VC), to connect 10-OH group of SN-38 through two spacers, p-aminobenzyloxycarbonyl (PAB) and ethylenediamine, which exhibited relatively stable in serum (Govindan et al., 2013). However, the extra incorporated ethylenediamine spacer may result in a relatively slow-release rate of SN-38 (Sharkey et al., 2012;Govindan et al., 2013), which may further affect the potency of its conjugates (Bargh et al., 2020;Dal Corso et al., 2020). Additionally, linkers sensitive to the b-glucuronidase in lysosomes have also been described in SN-38 based ADCs (Lau et al., 2018). However, the stereochemical complexity and potential overdependence on one specific enzyme of these linkers, as well as the unexplained low maximum tolerated dose of its conjugates may hinder its further clinical development (Burke et al., 2009;Bargh et al., 2019). So, these SN-38 ADCs still required for further improvements.
In this paper, we proposed a novel strategy to construct highly releasable and structurally stable SN-38-conjugates. The widely used Cathepsin B (CTSB) sensitive linker directly connect to the 10-OH group of SN-38 through ether bond. The ingenuity of this strategy lies in the first verification of dipeptide-phenolic ether fragment releasing payloads rapidly. Meanwhile, the acidic environment of lysosomes (pH ¼ 4.5-5.0) enables the lactone of SN-38 to transform from an open carboxylate, a less active form, into a closed form, thereby maintaining its strong cytotoxicity (Lau et al., 2018). In conclusion, this novel construction strategy of SN-38based ADCs has the potential to promote the development of SN-38 in ADCs and provide another choice for hydroxylgroup bearing payloads (Figure 1).

Materials
All chemical reagents were purchased from commercial sources (Aladdin, Acros, Alfa etc) and were used as received without any further purification. The anti-Her2 antibody Mil40 (a biosimilar of HerceptinV R ) was purchased from Hisun Pharmaceutical Co., Ltd. (Taizhou, China). Flash column chromatography was used for compound purification. 1 H NMR spectra of compounds synthesized were recorded at 400 MHz on JEM-ECA-400 magnetic resonance spectrometer using deuterated solvents with tetramethylsilane (TMS) as an internal standard. Chemical shifts were reported in ppm and coupling constants (J) in Hz. Mass spectra were recorded with API 3000 mass spectrometer. CTSB from human liver was purchased from Sigma and CTSB inhibitor E-64 from MedChemExpress. The fluorescence emission studies were performed on Microplate reader. All cells were purchased from American Type Culture Collection. Human ovarian carcinoma cell line SKOV-3, breast carcinoma cell line MCF-7 and MDA-MB-231 were cultured in high-glucose DMEM supplemented with 10% FBS, 0.1% penicillin-streptomycin and human breast carcinoma cell line Herceptin-resistant BT474 HerDR was cultured in RPMI-1640 medium containing 10% FBS, 0.1% penicillin-streptomycin.

Key intermediate compound 4
To the solution of compound 1 (1 g, 2.54 mmol) in anhydrous N, N-Dimethylformamide (DMF) (20 mL) under an ice bath was added slowly phosphorus tribromide (680 mg, 2.54 mmol) in anhydrous dichloromethane (30 mL). After 3 h, the reaction was completed and then the solution was transferred into ice water (100 mL). The dichloromethane layer was collected through multiple extractions and dried by anhydrous sodium sulfate overnight. After evaporation of the solvent under vacuum, the crude solid product compound 2 without being purified was obtained (815 mg).
To the solution of compound 2 (800 mg, 1.75 mmol) and SN-38 (689 mg, 1.75 mmol) in anhydrous DMF (25 mL) was added cesium carbonate (573 mg, 1.75 mmol). After being stirred in room temperature for 10 h, the reaction was transferred into the refrigerator for 1 h to precipitate cesium carbonate. After filtration of insoluble solid materials, the crude product was obtained by evaporation of the solution under vacuum and then purified through a silica gel column chromatography using methanol in dichloromethane as an eluent to yield compound 3 (862 mg, 64% Compound 3 (862 mg, 1.12 mmol) was dissolved in anhydrous dichloromethane containing 20% trifluoroacetic acid. After being stirred in room temperature 15 min, the crude product was obtained by evaporation of the solution under vacuum and then purified through a silica gel column chromatography using methanol (2%) in dichloromethane as an eluent to yield compound 4 (602 mg, 80.3%) as the light yellow solid.

Compound 5
To the solution of compound 4(100 mg, 0.15 mmol) in anhydrous DMF (15 mL) was added 6-Maleimidohexanoic acid Nhydroxysuccinimide ester (55.38 mg, 0.18 mmol) and N, N-Diisopropylethylamine (26 uL, 0.165 mmol). The reaction was stirred overnight under nitrogen protection. Next day, after evaporation of the solvent under vacuum, the crude product was purified through a silica gel column chromatography using methanol in dichloromethane from 1% to 2.5% as an eluent to yield compound 5 (98 mg, 76%) as the light yellow solid. 1

Compound 6
To the solution of compound 4(100 mg, 0.15 mmol) in anhydrousDMF (15 mL) was added N-Hydroxysuccinimidyl-(polyethylene glycol) 2-maleimide (76.5 mg, 0.18 mmol) and N, N-Diisopropylethylamine (26 uL, 0.165 mmol). The reaction was stirred overnight under nitrogen protection. Next day, after evaporation of the solvent under vacuum, the crude product was purified through a silica gel column chromatography using methanol in dichloromethane from 1% to 2.5% as an eluent to yield compound 6 (101 mg, 68.8%) as the light yellow solid. 1

Compound 7
To the solution of compound 4(100 mg, 0.15 mmol) in anhydrous DMF (15 mL) was added N-Hydroxysuccinimidyl-(polyethylene glycol) 4-maleimide (92.37 mg, 0.18 mmol) and N, N-Diisopropylethylamine (26 uL, 0.165 mmol). The reaction was stirred overnight under nitrogen protection. Next day, after evaporation of the solvent under vacuum, the crude product was purified through a silica gel column chromatography using methanol in dichloromethane from 1% to 2.5% as an eluent to yield compound 7 (86 mg, 53.8%) as the light yellow solid. 1

Compound 8
To the solution of compound 4(100 mg, 0.15 mmol) in anhydrous DMF (15 mL) was added N-Hydroxysuccinimidyl-(polyethylene glycol) 8-maleimide (123.72 mg, 0.18 mmol) and N, N-Diisopropylethylamine (26 uL, 0.165 mmol). The reaction was stirred overnight under nitrogen protection. Next day, after evaporation of the solvent under vacuum, the crude product was purified through a silica gel column chromatography using methanol in dichloromethane from 1% to 2.2% as an eluent to yield compound 8 (117 mg, 62.9%) as the light yellow solid. 1
Compound 9 (900 mg, 0.63 mmol) was dissolved in anhydrous dichloromethane containing 5% piperidine. After being stirred at room temperature for 2 h, the crude product was purified through flash chromatography using methanol in dichloromethane, ranging from 2% to 5%, to yield compound 10 (516 mg, 68.8%) as the light yellow solid. ESI m/z calculated for C 61 H 87 N 7 O 17 1189.62, found 1190.64.

Preparation of Mil40-SN38 conjugates
The storage buffer (20 mM L-Histidine, pH ¼ 5.5) and reaction buffer (20 mM L-Histidine, pH ¼ 7.2-7.5) of Humanized anti-Her2 IgG1 antibodies Mil40 were prepared. Compound 5, 6, 7, 8, 11 was dissolved in DMA as a stock solution, respectively. Mil40 in reaction buffer was treated with 2.4-3.0 mol equivalent TCEP to reduce the interchain disulfide bridges between the heavy-heavy and heavy-light chains for 2 h at 25 C. To the solution of the reduced Mil40 was added 8 mol equivalent of compound 5, 6, 7, 8, or 11 in DMA. The conjugation reaction was allowed to proceed at 25 C for 3 h with gentle mixing and the final concentration of Mil40 used in reactions were 5 mg/mL. After completion of conjugation, the reaction was quenched with excess Nacetylcysteine (NAC). Conjugates were exchanged into storage buffer using Sephadex-G25S to remove the free NAC-5, 6, 7, 8, or 11 and then were stored at À80 C before use for analysis and testing.
To further elevate the DAR values of conjugates, Mil40 in reaction buffer was treated with 8 mol equivalent TCEP to reduce the interchain disulfide bridges, then the solution of reduced Mil40 was added 15 mol equivalent of compound 5, or 11 in DMA. The conjugation reaction was allowed to proceed at room temperature for 24 h. Other operations were the same as the mentioned at first.

Quality control
DAR values of ADCs were analyzed by hydrophobic interaction chromatography (HIC), UPLC-Q-TOF-MS and the polymerization degree was determined by size exclusion chromatography. HIC analysis was carried out on a TSK-gel Butyl-NPR column (2.5 lm, 4.6 mm Â 3.5 cm, TOSOH Bioscience). Mobile phase A was 25 mM sodium phosphate and 1.8 M ammonium sulfate in water. Mobile phase B was 25 mM sodium phosphate in water containing 20% isopropanol. The gradient was 0% to 100% mobile phase B until 18.00 min, followed by re-equilibration from 18.00-18.01 min with 100% mobile phase A and isocratic step at 100% mobile phase A from 18.01-24.00 min, flow rate was 0.8 mL/min, and the UV detection wavelength was 280 nM. The ADC DAR was determined by calculating from the sum of the weighted ratios of the area under the peak multiplied by corresponding peak drug loading through HIC analysis. UPLC-Q-TOF-MS analysis was measured on Agilent 6550 LC-Q-TOF MS spectrometer with a TSK UP-SW3000 column (4.6 Â 150 mm, 2.5 lm). The column was eluted with an isocratic step at 0.2 mL/min with 200 mM ammonium acetate solution. The multiple charged peaks of ADCs were deconvoluted using the Agilent MassHunter Bioconfirm sofware (deconvolution for protein, Agilent technology).
The capacity of Mil40-11 binding to Her2 antigen compared with monoclonal antibody Mil40 was determined by ELISA. The detailed protocol of measuring the affinity of Mil40-11 to Her2 antigen was reported previously (Li et al., 2021;Xiao et al., 2021).

SN-38 released from phenolic ether
Released SN-38 was monitored by HPLC after the removal of Boc fragment of compound 20. Taking advantage of the selffluorescence of SN-38, the process of releasing payload from phenolic ether was studied. NAC-6, NAC-11, or NAC-21 was prepared by the reaction of compound 6, 11, or 21 with excessive NAC. The stock solution was prepared in triple distilled water containing DMSO (50%, v/v). Stock solution of E-64 of CTSB inhibitor was dissolved in triple distilled water. Stock solutions of CTSB with different concentrations were prepared in CTSB activity buffer (50 mM sodium acetate, 100 mM NaCl, 8 mM L-cysteine, 1 mM EDTA, pH 5.0). Fluorescence responses of NAC-21 (5 lM) in CTSB activity buffer containing DMSO (10%, v/v) to CTSB (3 UN/mL) were recorded after incubation at 37 C at different time points. Kinetics of fluorescence response of NAC-21 (5 lM) was recorded in the presence of varying concentrations of CTSB. The role of CTSB in releasing SN-38 was determined in the presence of CTSB inhibitor E-64.

Serum stability
To determine the stability of Mil40-11 in human plasma, SN-38 and 10-hydroxycamptothecin as an internal standard were diluted into 2 Â diluted human plasma to prepare the standard sample, and Mil40-11 sample preparation was the same as standards. Both samples were incubated at 37 C for a period of 10 days. The concentration of total free SN-38 released from the Mil40-11 conjugates samples was determined at various time-points according to the established standards using LC-MS. Chromatographic experiments were performed using a liquid chromatographic system consisting of Waters Alliance HPLC system equipped with a Sample Manager, a PDA detector and a QDa detector which is a compact single quad mass detector equipped with the electrospray ionization (ESI) interface (Waters, Czech Republic). Mobile phase A was 0.1% formic acid in water and mobile phase B was 0.1% formic acid in acetonitrile. A gradient elution was performed with 20% mobile phase B for 2 min, a linear increase to 100% mobile phase B until 3.0 min, followed by isocratic step at 100% mobile phase B from 3-5 min and re-equilibration from 5.1-8 min with 20% mobile phase B. The flow rate was set to 0.8 mL/min.

Flow cytometric analysis
Flow cytometry was applied to determine the targeting of Mil40-11 against SKOV-3 cells and MDA-MB-231 cells. SKOV-3 cells (5 Â 10 5 cells/sample) and MDA-MB-231 cells (5 Â 10 5 cells/sample) were harvested under the log phase of growth and suspended in FACS clean solution (PBS containing 2% fetal bovine serum). These cells were treated with Mil40-11 and Mil40 diluted in 100 lL PBS at the concentration of 5 ug/mL, or PBS as blank control at 4 C for 30 min, respectively. Then, the cells were washed twice with FACS clean solution and incubated with PE-conjugated anti-human IgG at 4 C for 30 min. Then, after being washed three times with FACS clean solution, cells were finally detected by flowed cytometry (BD Biosciences). The ratio of mean fluorescence intensity (rMFI) was used to reflect the targeting ability. Meanwhile, SKOV-3 cells were harvested and suspended in FACS clean solution. Experimental and control groups were all treated with Mil40, or Mil40-11 at 4 C for 1 h. After being washed twice with prechilled FACS clean solution, experimental groups were incubated at 37 C for 45 min, or 8 h, while the control groups were still incubated at 4 C for the set time points. All samples were washed three times with prechilled FACS clean solution when reaching the time points, and then incubated with PE-conjugated anti-human IgG at 4 C for 30 min. Flow cytometry was used to determine the rMFI, and internalization percentage was calculated with rMFI by the following formula: internalization rates (%) ¼ (100 -(rMFI of experimental groups/rMFI of control groups)) Â 100%.

In vitro cell imaging
SKOV-3 cells were seeded on chamber and incubated for 24 h at 37 C under 5% CO 2 and 95% air. After removal of medium, cells were treated with FITC labeled Mil40, or FITC labeled Mil40-11 in medium at 4 C for 30 min, or 37 C for 24 h. Groups incubated at 4 C were for observation of the affinity of Mil40, or Mil40-11 to SKOV-3 cells and groups incubated at 37 C for the observation of internalization of Mil40, or Mil40-11 into SKOV-3 cells. When reaching the time points, cells were washed with PBS and stained with Lysotracker Red at lysosome and DAPI at nucleus. After being washed three times with PBS, fluorescent images were taken under a fluorescent microscope.

In vivo antitumor activity in human ovarian xenograft model
All studies were conducted in compliance with accepted standards for the care and use of laboratory animals. Three BALB/c-nude mice (6-8 weeks old) were injected subcutaneously with SKOV-3 cells (1 Â 10 7 cells). When the volume of tumors reached approximate 500 mm 3 , the mice were killed and tumors were taken out, followed with the removal of envelope and necrotic tissues. The weighed 10 grams tumors were cut into a paste with a scalpel on ice and then wellmixed with 10 mL RPMI-1640 medium into tumor tissue suspension. Each mouse was subcutaneously inoculated with 0.2 mL tumor suspension. When the tumors reached a volume of approximately 100 mm 3 , nude mice were randomly assigned to seven treatment groups (n ¼ 7-8/group):control group (L-His buffer), three Mil40-11(DAR ¼ 7.1) treatment groups (5 mg/kg,10 mg/kg, and 20 mg/kg), two Mil40 treatment groups (10 mg/kg and 20 mg/kg) and the drug combination group (20 mg/kg Mil40-11 þ 0.66 mg/kg irinotecan). The seven treatment groups (250 lL) were injected by tail vein into mice at 0, 3, 7, 10 days. Subcutaneous tumors were measured using caliper every three or four days, and the tumor volume calculated as 1/2 Â length Â width Â width (mm). Animals were terminated when tumors grew to 2.5 cm 3 . Mice body weight was also monitored as a marker of toxicity. At the end of study, tumors of mice were collected and further analyzed by H&E staining.

Chemistry
The synthetic routes of Linker-SN-38, including compounds 5, 6, 7, 8, and 11, are shown in Scheme 1. The key intermediate compound 4 was obtained first in the following steps. Compound 1 was converted into compound 2 through bromination reactions. Because of the steric hindrance of a-hydroxy lactone, the hydroxy group of SN-38 does not require protection when preparing compound 3, which was obtained through a phenolic ether connection (Lau et al., 2018). The key intermediate compound 4 was finally obtained after removal of the Boc protection group and then reacted with maleimides containing different lengths of polyethylene glycol (PEG) to yield target compounds 5, 6, 7, and 8. Overall ADC incorporating the pendant PEG linker may exhibit the better hydrophilic properties than that of straight PEG linker. (Lyon et al., 2015;Tedeschini et al., 2021;Xiao et al., 2021) Compound 4 reacted with activated Fmoc-Lys (PEG 8 )-OH to provide compound 9. Fmoc deprotection followed by maleimide incorporation yielded the desired compound 11 containing the pendant PEG moiety. Detailed synthetic procedures are shown in the Experimental Section.

Antibody conjugation
Compounds 5, 6, 7, 8, and 11 were conjugated to the humanized monoclonal antibody Mil40. The well-studied Mil40, marketed biosimilar of Herceptin, targets the Her2 antigen. The intermolecular disulfide bonds of Mil40 was reduced with TCEP to expose the sulfhydryl groups, followed by in situ conjugation with maleimide-containing compounds 5, 6, 7, 8, or 11. The conjugation was terminated with excess of NAC.

In vitro SN-38 release
SN-38 release of compounds 20 or 21 was evaluated at the chemical or CTSB level (Figure 2(A)). Given that PAB-SN-38 is structurally unstable, a Boc protection group was introduced to the amino group of PAB-SN-38 to provide compound 20. Two consecutive steps were employed to determine the SN-38 release from compound 20. First, the Boc protection group was completely removed with trifluoracetic acid. Second, PAB-SN-38 and released SN-38 were monitored at set time points by HPLC after PBS was added to the reaction systems to initiate the self-elimination of PAB. PAB-SN-38 was found to release SN-38 with high efficiency with a halflife t 1/2 ¼ 36 min (Figure 2(B)), which is far shorter than that of N, N'-dimethyl ethylenediamine spacer of CL2E linker (Govindan et al., 2013). The results demonstrated that CTSB linker direct connection to phenolic hydroxyl group of SN-38 did have a higher efficiency of releasing payloads. Thus, the novel strategy of phenolic ether connection with higher efficiency of releasing payloads may exhibit the potent inhibitory activities comparable to SN-38. At the CTSB level, studies on SN-38 release from compound 21 with CTSB were carried out with the help of the self-fluorescence of SN-38.
The fluorescence spectra and UV-Vis absorption of free SN-38 were recorded. An emission maximum was centered at 550 nm, and UV absorption was centered at 373 nm ( Figure  S1) gradually increased over time, and the emission band at 427 nm decreased (Figure 2(C)). Under irradiation with light at k ¼ 365 nm, the physical appearance of the reaction solution observed by the naked eye changed from blue to pale yellow fluorescence (inset Figure 2(C)). Taken together, these findings suggested that free SN-38 could be released from compound 21 in the presence of CTSB. Furthermore, the fluorescence emission of compound 21 at 550 nm was also monitored in the presence of different concentrations (3 UN/ mL, 1 UN/mL, 0.3 UN/mL) of CTSB over time. The rate of fluorescence intensity enhancement exhibited dependent on CTSB concentrations (Figure 2(D)). In contrast, no significant fluorescence enhancement at 550 nm was observed after pretreatment with the CTSB inhibitor E-64 (4 lM) ( Figure  2(E)). These findings implied that CTSB played a key role in the cleavage of compound 21 and the release of SN-38.

In vitro antitumour activity of SN-38-based ADCs with DAR $ 4
Her2-positive cell lines SKOV-3 and BT474 HerDR, which are Herceptin-resistant, and Her2-negative cell lines MDA-MB-231 and MCF-7 were selected to evaluate the in vitro antitumour activities of SN-38-based ADCs with DAR $ 3.7. The Her2 status of these cell lines was previously determined with flow cytometry . As shown in Figure 2 (Table S1). However, after SN-38 was conjugated to Mil40 to provide ADCs, these conjugates did not show strong cytotoxicity, similar to SN-38, in Her2-positive cell lines. The activities of ADCs were significantly reduced by at least 8 times in SKOV-3 cells and at least 2 times in BT474 HerDR cells (Figure 3 and Table S1). It was speculated that the length of the PEG moiety or the efficiency of releasing SN-38 could influence the antitumour activities of ADCs (Dal Corso et al., 2020). Chemical structure analysis indicated that Mil40-11, a conjugate containing a pendant PEG 8 moiety, had better antitumour activities than other ADCs, including Mil40-6 with a PEG 2 moiety, Mil40-7 with a PEG 4 moiety, and Mil40-8 with a linear PEG 8 moiety. Notably, Mil40-5 without a PEG moiety also exhibited cytotoxicity similar to Mil40-11. Thus, Mil40-5 and Mil40-11 were preliminarily screened for further studies.
3.5. Preparation and characterization of Mil40-11 with DAR 5 7.1 Based on the in vitro antitumour activities of ADCs with DAR $3.7, Mil40-5 and Mil40-11 were further improved for better cytotoxicity. Previous studies indicated that elevating the DAR values would significantly improve the antitumour activities of ADCs, and this strategy was especially common within SN-38-based ADCs (Goldenberg et al., 2015). Thus, for better antitumour activity, ADCs with higher DAR values were obtained by improving conjugation conditions. HIC analysis showed that Mil40-11 and Mil40-5 were predominantly composed of species with a DAR value of 8. UPLC-Q-TOF-MS analysis determined that the percentage of DAR8 accounted for 78%, 51%, respectively, which indicated that we achieved the preparation of ADCs with high DAR values (Figure 4(A,B) and Figure S2). The specific DAR values of Mil40-5 (DAR ¼ 6.5) and Mil40-11 (DAR ¼ 7.1) were calculated on the UPLC-Q-TOF-MS analysis. However, according to the size exclusion chromatography (SEC) analysis, Mil40-5 (DAR ¼ 6.5) showed more aggregates with aggregation rate of 29% than that of Mi40-5 (DAR ¼ 3.7) with aggregation rate of 1% while Mil40-11 (DAR ¼ 7.1) had less aggregates with an aggregation rate of 2% (Figure 4(C) and Figure S3). These findings indicated that Mil40-11 had an acceptable low aggregation rate even with a high drug load, which could be attributed to the incorporated hydrophilic pendant PEG 8 moiety. In contrast, Mil40-5 had an unacceptably high aggregation rate after the DAR value was increased because of its hydrophobicity, which may contribute to its high clearance rate in vivo (Lyon et al., 2015). The S þ Sw value of compound 11 was 0.008, as predicted by the ADMET predicator 8.5 higher than that of compound 5 (0.005), which indicates that compound 11 may have better water solubility and is more suitable for construction into ADCs with a high drug load. Moreover, with the help of the self-fluorescence of SN-38, compounds 6 and 11 were selected to study the effect of the length of the PEG moiety on the SN-38 release efficiency of linker-SN-38. NAC-11 was found to release SN-38 with higher efficiency than NAC-6 ( Figure 4(D)), which implied that a longer PEG moiety seemed beneficial for SN-38 release. Taken together, these findings suggested that Mil40-11 (DAR ¼ 7.1) is worthy of further study.
The effect of increasing DAR values of Mil40-11 (DAR ¼ 7.1) on the binding affinity to Her2 antigen was determined using ELISA. The binding of Mil40-11 (DAR ¼ 7.1) and Mil40 was comparable, with EC 50 values of 0.17 mg/mL and 0.08 mg/ mL, respectively (Figure 4(E)). Thus, these results implied that Mil40-11 (DAR ¼ 7.1) retained a high affinity to Her2 antigen.
The stability of Mil40-11 (DAR ¼ 7.1) was evaluated in diluted human serum using LC-MS/MS analysis. As shown in Figure 4(F), approximately 0.6% ($2.33 nM) SN-38 released from Mil40-11 was detected after incubation for 10 days in 50% plasma, which indicated that Mil40-11 (DAR ¼ 7.1) was highly stable and that the application of the CTSB cleavable linker used for SN-38-based ADCs met the requirements for the design of ADCs.
The above data demonstrated that the strategy of elevating DAR values was feasible in terms of the construction and quality control of ADCs. Mil40-11 (DAR ¼ 7.1) exhibited significant advantages in reducing the aggregation rate and releasing SN-38 with high efficiency. Thus, Mil40-11 (DAR ¼ 7.1) was selected as the target ADC.
3.6. The targeting and internalization of Mil40-11 (DAR 5 7.1) Flow cytometry was applied to determine the targeting of Mil40-11 against Her2-positive SKOV-3 cell lines and Her2negative MDA-MB-231 cell lines. As shown in Figure 5(A), Mil40-11 showed similar targeting ability as Mil40. ADC was highly targeted against SKOV-3 cells, with a mean fluorescence intensity (rMFI) value of 42, while Mil40-11 rarely targeted MDA-MB-231 cells, which was consistent with the status of Her2 expression. The above results demonstrated that conjugating compound 11 onto Mil40 did not affect the targeting of Mil40-11 against Her2-positive cell lines. The internalization of Mil40-11 was also determined in SKOV-3 cells using flow cytometry. Mil40-11 and Mil40 possessed similar internalization efficiencies. After incubation in SKOV-3 cells for 8 h, the internalization percentage of Mil40-11 was 18%, and that of Mil40 was 11% ( Figure 5(B)). These findings indicated that conjugating compound 11 onto Mil40 did not affect the internalization of Mil40-11.
Meanwhile, the results of fluorescence microscopy showed that SKOV-3 cells incubated with FITC-stained Mil40 or FITCstained Mil40-11 at 4 C for 30 min showed a similar pattern, where they bound to the Her2 antigens expressed on the cell membrane (shown in green). After incubation at 37 C for 24 h, the green signal appeared within cells. Through the intracellular signal colocalization, lysosome sections were found to display yellow, which indicated that Mil40-11 was internalized and transported into lysosomes (Kalim et al., 2017) (Figure 5(C)).

3.7.
In vitro antitumour activity of Mil40-11 (DAR 5 7.1) As shown in Figure 2(B), the antitumour activity of Mil40-11 (DAR ¼ 7.1) was significantly enhanced in SKOV-3 cell lines compared with Mil40-11 (DAR ¼ 3.8), with IC 50 values of 22.2 nM and 86.3 nM, respectively. Moreover, Mil40-11 (DAR ¼ 7.1) exhibited comparable cytotoxicity to SN-38 in BT474 HerDR cell lines, with IC 50 values of 5.5 nM and 7.3 nM, respectively. Mil40-11 (DAR ¼ 3.8) showed relatively weak antitumour activity with an IC 50 value of 28.8 nM. Mil40 did not show cytotoxicity in SKOV-3 and BT474 HerDR cell lines. Based on the above data, Mil40-11 (DAR ¼ 7.1) exhibited strong cytotoxicity in Her2-positive cell lines, and its enhanced activity demonstrated that elevating the DAR values would significantly improve the antitumour activities of ADCs. This is observed in this case for in vitro antitumor activity and it is observed with compound 11 derived ADCs.
antitumour effects, although no significant tumor suppression (P > 0.05) was observed compared with the drug combination treatment group (Figure 6(A)).
During the whole treatment period, the body weights of mice treated with 20 mg/kg ADC were significantly higher than those of mice treated with buffer, which implied that ADC could improve the physical condition of tumor-bearing mice to avoid weight loss (Figure 6(B)). In addition, the tumor weight of the 20 mg/kg ADC treatment group was lower than that of the Mil40 group with the same dose and drug combination group (Figure 6(C)). H&E staining showed clear fibrosis of tumor tissue, nucleosome pyknosis and hyperchromatism of tumor cells in the ADC treatment group (Figure 6(D)).

Conclusions
We reported a novel SN-38-based ADC, Mil40-11, with a high DAR value of 7.1, the linker-drug part of which was constructed by directly connecting the phenolic hydroxyl group of SN-38 to the CTSB cleavable linker through ether bonds. In this study, the feasibility of directly connecting SN-38 to a Figure 6. In vivo efficacy and tolerability of Mil40-11 (DAR ¼ 7.1) in human tumor xenograft models. (A) Efficacy in the human ovarian cancer model SKOV-3. BALB/c-nude mice were implanted subcutaneously with SKOV-3 tumor tissues. When the size of tumors reached $100 mm 3 , tumor-bearing mice were treated with vehicle, Mil40, and Mil40-11 (DAR ¼ 7.1) on days 0, 3, 7 and 10, respectively. The results are shown as the mean ± SEM, n ¼ 7-8/group. To compare the difference in activity between Mil40 and Mil40-11 (DAR ¼ 7.1), unpaired two-tailed t tests were used, and statistical analysis was performed using GraphPad Prism 8.0. 2; ÃÃ p < .005 and ÃÃÃ p < .0005. (B) The body weight of mice did not show significant changes during treatment in the subcutaneous xenograft model. Statistical analysis of the difference in body weight was also performed using GraphPad Prism 8.0. 2; Ã p < .05. (C) Tumor weight of mice in the subcutaneous xenograft model. Statistical analysis of the difference in body weight was also performed using GraphPad Prism 8.0. 2; Ã p < .05. (D) Histological sections of SKOV3 tumor tissues with H&E staining. Red arrows: fibrosis of tumor. Blue arrows: nucleosome pyknosis. Green arrows: hyperchromatism. Scale bars: 100 lm.
VA-based linker through the ether bond was determined. With the advantages of the self-fluorescence of SN-38, the high SN-38 release efficiency of compound 21 was observed in the presence of CTSB. Upon exposure to CTSB, the emission maximum of compound 21 quickly shifted from 427 nM to 550 nM, characteristic of SN-38. After CTSB was pretreated with the CTSB inhibitor E-64, no emission maximum was observed. Thus, it was concluded that an ether bond connection could quickly release SN-38 after self-elimination of PAB and CTSB played a key role in this process.
We evaluated the antitumour activities of ADCs with DAR $ 3.7 in a series of Her2-positive tumor cells. Compared with SN-38 cells, ADCs displayed significantly lower activities in the SKOV-3 and BT474 cell lines. The strategy of elevating the DAR values of ADCs, especially common in SN-38-based ADCs, was adopted for better antitumour effects. Consequently, the increased hydrophilicity of compound 11 enabled Mil40-11 to be prepared with a high DAR value of 7.1 without a significant aggregation rate (2%), while the aggregation rate of compound 5-based ADCs with a DAR value of 6.5 reached 29% because of its hydrophobicity. In addition, the length of the PEG moiety could influence the efficiency of linker-SN-38 in the presence of CTSB. NAC-compound 11 with a pendant PEG 8 moiety released SN-38 faster than NAC-compound 6. Thus, Mil40-11 (DAR ¼ 7.1) was finally screened for further research.
Compared with Mil40, the targeting and internalization of Mil40-11 (DAR ¼ 7.1) was not affected during conjugation. Its stability in human serum demonstrated that the adopted phenolic ether connection addressed the instability of SN-38based ADCs and could be incorporated into the design of ADCs. Furthermore, elevating the DAR values enhanced the cytotoxicity of Mil40-11, which exhibited in vitro antitumour activities comparable to SN-38 and significantly delayed tumor growth.
In summary, the ether bond incorporated into SN-38based ADCs displayed high efficiency of releasing SN-38, stability in human serum and acceptable antitumour activities. These exciting findings demonstrated that the strategy of ether bond connection broadens the applications of SN-38 in ADCs and further provides another choice for CTSB cleavable triggers connecting other hydroxyl-bearing toxins.