Trim-Away in adult animals through Nano-ERASER and its application in cancer therapy

Trim-Away is a versatile intracellular protein degradation pathway that has been extensively explored in vitro. However, the in vivo application of Trim-Away is limited at oocyte and zygote stages due to the lack of an in vivo practical approach for intracellular antibody delivery. To broaden the application of Trim-Away, especially for clinical use, we developed a nanogel-based Nano-ERASER system. Here, we demonstrated that the intracellular delivery of anti-programmed cell death ligand 1 (PD-L1) antibody through Nano-ERASER could effectively deplete PD-L1 in triple negative breast cancer (TNBC) cells and induce cancer cell death. Furthermore, with the help of a tumor tissue-targeted nanogel, anti-PD-L1 antibody-loaded Nano-ERASER effectively inhibited tumor progression in a TNBC mouse model. These results confirmed that Nano-ERASER realized Trim-Away in adult animals for the first time, which could be an effective tool for disease treatment and studying gene/protein function both in vitro and in vivo.


Introduction 64
Trim-Away is a recently discovered endogenous protein degradation mechanism that 65 can selectively degrade proteins in mammalian cells by intracellularly delivering 66 antibodies. 1 It utilizes an intrinsic cellular self-defense machinery, which involves the 67 intracellular supply of an antibody, binding of the antibody to its target protein, 68 formation of tripartite motif containing-21 (TRIM21) antibody/target protein complex 69 and ubiquitination, and proteasome-mediated degradation of the complex. Attributed 70 to its merit in highly selective and rapidly degrading intracellular protein, A broad 71 spectrum of proteins have been successfully down-regulated in various cells with Trim-72 Away. 1-7 73 Despite being broadly explored in vitro, the in vivo application of Trim-Away is 74 limited in oocyte and zygote stages due to the lack of a practical approach to delivering 75 antibodies in animal models, not to mention in humans for treating diseases. [2][3][4][5][6][8][9][10][11] By 76 far, the intracellular delivery of antibodies in most Trim-Away applications was realized 77 through micro-injection and electroporation, which are impractical for an in vivo setting. 78 Given the challenge and the potential benefit of targeting undruggable targets, there is 79 an urgent need for a clinical translational approach to enable Trim-Away-based protein 80 degradation for in vivo application. 81 Our group recently developed a Nano-ERASER technology, which realizes Trim-Away 82 in cells without microinjection or electroporation through intracellular delivery and 83 tracelessly releasing antibodies. 12 Nano-ERASER is an antibody-conjugated polymer 84 nanogel (NG) system that can effectively intracellularly deliver and release antibodies 85 and induce the degradation of a specific endogenous protein. The potential therapeutic 86 efficacy of Nano-ERASER has been validated by depleting COPZ1 protein and 87 causing the death of cancer cells. 12 88

Generation of an aPDL1-loaded tumor-targeted nanogel (TN-PDL1) 116
To enable Nano-ERASER based Trim-Away for cancer therapy, we generated TN-117 PDL1 by functionalizing Nano-ERASER with an ASGPR receptor targeting ligand, 118 lactobionic acid (LBA) (Supplementary Scheme 1). 27 TN-PDL1 was fabricated 119 following our published protocol except replacing the previous antibody and targeting 120 ligand with aPDL1 and LBA through amide linkage (Fig. 1a), 12 respectively. Agarose 121 gel electrophoresis confirmed that aPDL1 could be encapsulated into the nanogel and 122 the encapsulated aPDL1 can be successfully released under a reducing environment 123 mimicking intracellular conditions (Fig. 1c). Dynamic light scattering revealed that the 124 size distribution of the nanogel was 78.6 ± 35.6 nm (PDI: 0.2, Fig. 1d); transmission 125 electron microscope found that the nanogel had a spherical shape (Fig. 1e). Zeta sizer 126 shown that TN-PDL1 had a surface charge of -18±1.19 mV (Supplementary Fig. 1a). 127 To evaluate the targeting effect of the LBA ligand for TNBC, Cy5 labeled-aPDL1 was 128 adopted. Nanogels with LBA modification (TN-PDL1) and without LBA modification (N-129 PDL1) were incubated with 4T1 cells. After three hours of co-culture, the cellular uptake 130 of the nanogels was observed by confocal microscope and quantified by flow cytometry. 131 Fig. 1b showed that the functionalization of LBA greatly 132 improved the cellular uptake of TN-PDL1, evidenced by the boosted red signals inside 133 the cells. Similar results were also confirmed by flow cytometry (Fig. 2b-c). As 134 expected, only a weak red signal was observed in the free aPDL1 treated cells, 135

Fig. 2a-c and Supplementary
suggesting the necessity of an effective tool to transfer aPDL1 into the cell. Fig. 2c  136 revealed that there was a greater than seven-fold increase in fluorescent intensity in 137 TN-PDL1 treated cells than that of free aPDL1 treated cells. 138

Subcellular localization and lysosomal escaping of TN-PDL1 139
To study the lysosomal escaping of an antibody inside the cells, Cy5 labeled IgG  Cy5) was employed, due to its inert effect, as a model antibody to yield TN-IgG-Cy5 141 nanogel. It was reported that copper ions facilitate the lysosomal escaping of 142 chelators. 28 4T1 cells were incubated with TN-IgG-Cy5 for 3 h with or without the 143 addition of 10 µM CuCl2, and observed with confocal microscopy. The abundant red 144 signals inside TN-IgG-Cy5 treated cells proved that targeted NGs could effectively 145 enter the cells (Fig. 2d). It was noticed that most red signals in the TN-IgG-Cy5 treated 146 cells overlapped with the green signals of Lysotraker, indicating the slow escaping of 147 the TN-IgG from the lysosomes. In contrast, the addition of copper ions facilitated the 148 lysosomal escaping of TN-IgG, evidenced by more standing alone red spots in the  IgG-Cy5/Cu treated cells (Fig. 2d). Since copper ions has been reported for promoting 150 lysosomal escaping, 28 we attribute the quick lysosomal escaping of the NGs to the 151 positive charges of the copper ions. 152

Intracellular release of antibody from nanogel 153
To monitor the intracellular release of antibody, which is a prerequisite of Trim-Away, 154 Förster resonance energy transfer (FRET) technology was employed. Due to its inert 155 bioactive property, control IgG was selected as a model antibody. Cy5-labeled IgG and 156 Cy3-labeled PDA-PEG polymer were used to yield Cy3/Cy5 dual-labeled NGs. After 4 157 h of incubation with TN-IgG-Cy3-Cy5, the cells were excited with a 555 nm laser. 158 Strong Cy5 (red, emission 640-700 nm) signals accompanied by weak Cy3 (green, 159 emission 560-600 nm) signals were observed in the 4T1 cells (Fig. 2e), indicating an 160 apparent FRET phenomenon due to the closely associated polymer (Cy3) and 161 antibody (Cy5). At 8 h post-incubation, the increased Cy3 signals and declined Cy5 162 signals were observed, indicating a partial liberation of Cy5-labelled antibody from the 163 NGs. The subsequent fading of red signals at 24 h suggests the almost full release of 164 antibodies inside the cell. These results suggest targeted nanogels (TN) can effectively 165 deliver an antibody into cells and release it in a timed manner. 166

TN-PDL1 depletes both total PD-L1 and membrane PD-L1 in 4T1 cells 167
For the success of Trim-Away, the availability of intracellular TRIM21 is required. 168 Supplementary Fig. 3 revealed that 4T1 cells express a high level of TRIM21, which 169 satisfies the prerequisite of Trim-Away. Since TN-IgG can effectively carry IgG into the 170 cells and release IgG intracellularly (Fig. 2d), we first probed whether TN-IgG could 171 induce the degradation of PD-L1. Supplementary Fig. 2 found that the intracellular 172 delivery of IgG did not affect the expression of PD-L1. However, TN-PDL1 effectively 173 reduced PD-L1 level in 4T1 cells at the dose of 100 ng/ml after 6 h of treatment, 174 suggesting the quick action of Trim-Away ( Supplementary Fig. 3a-b). Furthermore, 175 TN-PDL1 exhibited higher efficacy in degrading PD-L1 than N-PDL1, attributed to its 176 enhanced cellular uptake. There was no significant difference between the treatments 177 with or without the addition of CuCl2 (Supplementary Fig. 3a-b). Interestingly, after 24 178 h of treatment, N-PDL1 and TN-PDL1 coupled with CuCl2 induced more than 40% and 179 60% PD-L1 reduction ( Fig. 3a-b), respectively. In contrast, TN-PDL1 alone did not 180 cause significant PD-L1 change in the same period, suggesting insufficient TN-PDL1 181 was released inside the cytoplasm due to its relatively poor lysosomal escaping 182 capacity (Fig. 2d). By combining the results in Supplementary Fig. 2 and Fig. 3a- we can conclude that the reduction of PD-L1 in TN-PDL1 treated cells was due to the 184 specific degradation effect of TN-PDL1, not the carrier system. 185 To further probe the effect of TN-PDL1 on the expression of membrane PD-L1, 186 immunocytochemistry was employed. Fig. 3c-d revealed that the degradation of 187 intracellular PD-L1 also resulted in reduced PD-L1 expression on the cell membrane 188 of 4T1 cells. This suggests that TN-PDL1 may exhibit immunotherapeutic effect by 189 preventing the binding of PD-1 on T cells to the PD-L1 on the TNBC cells.

TN-PDL1 inhibits the proliferation of cancer cells 191
The cell proliferation inhibitory effect of TN-PDL1 on 4T1 cells was evaluated by MTT 192 assay. Since our previous study found that the combination of copper ions and 193 polymer-carriers kills cancer cells, 29 polymer control at the corresponding dose of its 194 nanogel counterpart was included. Without the addition of CuCl2, none of the 195 treatments showed apparent toxicity (Fig. 3e). Interestingly, PDL1 at the dose of 500 ng/ml effectively killed 4T1 cells when 10 µM CuCl2 was added 197 (Fig. 3f), attributing to the function of copper ions in facilitating the nanogel escape 198 from lysosomes, 28 which is also evidenced in Fig. 2d. In contrast, both aPDL1 and 199 polymer/copper at the same dose did not kill cancer cells. Since copper ions boost the 200 efficacy of TN-PDL1, 10 µM CuCl2 was added in the following in vitro experiments. 201 Furthermore, TN-IgG, loaded with a control antibody (isotype IgG), did not induce cell 202 death ( Fig. 3e-f), suggesting the safety of the carrier system. 203 In addition, TN-PDL1 also exhibited a cell-killing effect for NCI/ADR-Res, PANC-1, and 204 Fig. 4), indicating the broad spectrum of its 205 therapeutic effect. In contrast, TN-PDL1 is nontoxic for normal fibroblasts, NIH-3T3 206 cells (Supplementary Fig. 4). Since TN-PDL1 treatment downregulates the membrane 207 expression of PD-L1 in cancer cells, we further evaluated the potential 208 immunotherapeutic effect of TN-PDL1 in vitro by co-culturing 4T1 cells together with 209 mouse peripheral blood mononuclear cells (PBMCs). It was revealed that the 210 combination of TN-PDL1 and PBMCs exhibited a much stronger cell-killing effect than 211 TN-PDL1 alone (Fig. 3g). This enhanced effect is even more significant when copper 212 ions were included (Fig. 3h), suggesting a boosted immunotherapeutic effect. 213

TN-PDL1 prevents the activation of STAT3 and inhibits the repair of DNA damage 214
To explore the mechanism of TN-PDL1 in killing cancer cells, we first investigated the 215 that the depletion of intracellular PD-L1 resulted in the reduction of both STAT3 and 217 pSTAT3, which is critical for the proliferation and migration of cancer cells. 30 It was also 218 found that the TN-PDL1-induced protein reduction for STAT3, pSTAT3, PD-L1, and 219 TRIM21 was dose-dependent, while the addition of CuCl2 further boosted its efficacy 220 ( Supplementary Fig. 3b-c). It has been reported that silencing PD-L1 prevents the 221 repair of DNA damage. 31 The accumulation of γ-H2AX, a marker for DNA damage, 222 was observed in TN-PDL1 treated cancer cells (Fig. 3l-

TN-PDL1 inhibits the migration and invasion of TNBC cells 227
Recent research found that intracellular PD-L1 promotes epithelial-mesenchymal 228 transition (EMT) in cancer cells. 25 To probe the effect of TN-PDL1 on the EMT of 4T1 229 cells, western immunoblotting was employed to evaluate the expression of E-Cadherin, 230 a marker for EMT, after TN-PDL1 treatment. Transwell invasion assay, respectively. As expected, both N-PDL1 and TN-PDL1 236 effectively inhibited the migration (Fig. 4c-d) and invasion ( Fig. 4e-f) of 4T1 cells. In 237 contrast, free aPDL1 only showed a slightly inhibitory effect. 238

TN-PDL1 promotes the disassemble of tumor spheroid and the penetration of 239
PBMCs 240 To study if the functionalization of LBA could promote the penetration of nanogels into 241 the tumor mass, tumor spheroid was employed. Bovine serum albumin (BSA) was adopted as a model protein to eliminate potential interference due to the activity of 243 antibodies. Compared with free BSA and N-BSA, the ligand functionalized TN-BSA 244 penetrated much deeper into the tumor spheroid (Fig. 4g). Encouraged by the 245 outstanding tumor infiltrating capacity of the targeted nanogel, we further investigated 246 the cell-killing effect of TN-PDL1 in a 3D tumor spheroid model. Without the 247 supplementation of CuCl2, none of the treatments induced the death of 4T1 cells (Fig.  248   4h). Interestingly, the addition of CuCl2 resulted in the disassembly of the tumor 249 spheroid (Supplementary Fig. 5a) and the death of the 4T1 cells ( Fig. 4i), attributed to 250 the better lysosomal escaping of TN-PDL1 when coupled CuCl2. 251 Since TN-PDL1 effectively attenuated the membrane expression of PD-L1 ( Fig. 3c-d), 252 we further probed if TN-PDL1 treatment could promote the penetration of PBMC into 253 the tumor mass. Fig. 4j found that TN-PDL1 treatment alone induced the assembly of 254 dye-prelabeled PBMCs surrounding the tumor spheroid. Interestingly, Fig. 4k revealed 255 that the combination of TN-PDL1 and CuCl2 effectively promoted the penetration of 256 PBMCs deep into the tumor spheroid, attributed to the reduced PD-L1 expression on 257 the cell membrane (Supplementary Fig. 5b). 258

TN-PDL1 biodistribution in a TNBC orthotopic tumor mouse model 259
To investigate the biodistribution of the targeted nanogel, Cy5-labeled aPDL1 was 260 employed. N-PDL1 and TN-PDL1 nanogels were i.v. injected into BALB/c mice bearing 261 orthotopic 4T1 TNBC tumor established following our published protocol and detected 262 with an IVIS imaging system. 32 It was revealed that free aPDL1 and N-PDL1 were 263 quickly cleared from the circulation system, as evidenced by the strong fluorescence 264 signal in the bladder 4 h post injection (Fig. 5a). In contrast, a weak fluorescence signal 265 was observed in the bladder of TN-PDL1 treated mice, suggesting extended circulation 266 time for TN-PDL1. The ex vivo image shown in Fig. 5b revealed that TN-PDL1s were 267 selectively retained in the tumor tissue, as evidenced by the brightest fluorescence signal, which was much stronger than that from the N-PDL1 treated group, indicating 269 that the functionalization of LBA did boost nanogel targeting TNBC tumor. The overall 270 stronger fluorescence signals found in other organs further confirmed that LBA 271 modification prolonged the circulation of TN-PDL1, which subsequently resulted in its 272 more efficient tumor-targeting effect. 273

TN-PDL1 inhibits the growth of TNBC 274
The tumor growth inhibitory effect of the TN-PDL1 was evaluated in BALB/c mice 275 inoculated with 4T1-Luc cells orthotopically. Treatments were given via tail vein 276 injection on day 10, 13, and 16 post-cell inoculation at an aPDL1 equivalent dose of 277 2.5 mg/kg with/without the i.p. administration of copper gluconate at the dose of 2 278 mg/kg (Fig. 5c). For aPDL1 treatment, only the first two doses were given to four mice, 279 because the other three mice immediately died upon a third dose, possibly due to 280 intrinsic toxicity and/or a xenogeneic hypersensitive response to rat-derived aPDL1. 33 281 To study the benefit and potential side effects of long-term treatment, one group of 282 mice received an additional three doses of TN-PDL1 therapy on day 21, 28, and 32. 283 and Supplementary Fig. 6-7). It is worth noting that the addition of copper did not result 288 in a notable difference in the treatments involved N-PDL1 and TN-PDL1, which is 289 different from the in vitro results shown in Fig. 3e-h. We postulate this may be due to 290 the elevated copper concentration in the tumor tissue, 34 which negates the necessity 291 of copper supplement in vivo. Furthermore, six doses of TN-PDL1 treatment exhibited 292 a more significant inhibitory effect on tumor progression than three doses ( Fig. 5d-f). 293

TN-PDL1 inhibits the metastasis of TNBC without inducing side effects
The decrease in tumor nodules in the lungs of TN-PDL1 treated mice revealed that 295 TN-PDL1 also attenuated the metastasis of the TNBC to the lung ( Fig. 5h and 296 Supplementary Fig. 7), which may be attributed to the anti-migration and anti-invasion 297 effect of TN-PDL1 ( Fig. 4c-f). In contrast to the TNBC tumor metastasis to the liver in 298 other treatment groups, an attenuated liver metastatic effect was also observed in TN-299 PDL1 treated mice (Fig. 5i). More importantly, animals in all treatment groups did not 300 see apparent weight loss except the control group, in which mice lost about 13% weight 301 at the end of the study (Fig. 5g). In addition, no abnormal structure was observed in 302 the major organs among all groups (Supplementary Fig. 8). 303

TN-PDL1 effect on the expression of PD-L1 and pSTAT3 in the tumor tissue 304
To investigate how TN-PDL1 treatment affects the expression of PD-L1, γ-H2AX, 305 pSTAT3, and STAT3 in vivo, two mice were sacrificed 3 days after receiving the 306 second treatment dose. Western immunoblotting found that TN-PDL1 treatment 307 reduced the PD-L1, pSTAT3, and STAT3 expression in the tumor tissue to 32%, 30%, 308 56% ( Fig. 6a, 6c-d), respectively, proving the success of the application of Trim-Away 309 in vivo. As expected, the expression of γ-H2AX was upregulated by 15.5 folds in the 310 tumor after TN-PDL1 treatment (Fig. 6b). 311 This study showed a successful in vivo Trim-Away treatment in adult animals for the 343 first time and proved that Nano-ERASER is an effective tool for cancer treatment. 344

TN-PDL1 regulates the infiltration of T cells inside the tumor tissue
Furthermore, due to the interchangeable nature of the antibody and targeting ligand in 345 the Nano-ERASER, we expect many diseases caused by the malfunction of proteins, including undruggable targets, could be effectively controlled through Trim-Away. 347 Furthermore, the success of in vivo Trim-Away through Nano-ERASER technology 348 would broaden the application for antibodies and provide a highly effective and specific 349 tool for studying gene/protein function in adult animals.

477
The lysosomes were stained with LysoTracker Green. The cell nuclei were stained with DAPI.                 To track nanoparticles in vivo distribution in the above prepared 4T1 orthotopic tumor-bearing 782 BALB/c mice, Cy5 was loaded in the nanogels as a fluorescence probe by conjugating with aPDL1. Cy5 labeled aPDL1, N-PDL1, and TN-PDL1 were injected intravenously at a Cy5 dose 784 of 0.5 mg/kg. Three hours post-injection, the mice were anesthetized and imaged using an 785 IVIS Lumina III imaging system (excitation: 620 nm; emission: 690 nm). After that, the animals 786 were sacrificed, and the major organs, including heart, liver, spleen, lung, kidney, and tumor, 787 were collected for ex vivo imaging to investigate the tissue distribution of aPDL1 and aPDL1 788 loaded nanogels.  The harvested tumor tissues and major organs were fixed with 4% paraformaldehyde for 24 h, 802 followed by 15% sucrose (24h)/30% sucrose (24h), and processed by OCT embedding and 803 cryosectioning. Tissue sections (heart, liver, spleen, and kidney) and lung sections were 804 hematoxylin and eosin (H&E) stained for acute toxicity evaluation and lung metastasis 805 evaluation. The tissue sections were attached to poly-L-lysine-coated slides and warmed to 806 RT for 30 min. After that, the OCT was washed with PBS, and the slices were dried at RT 807 before blocked with 5% BSA for 1 h. The primary antibodies of CD3 (Proteintech, cat. no.