Low-density lipoprotein receptor-targeting chimeras for membrane protein degradation and enhanced drug delivery

22 Antibody-based therapeutics encompass diverse modalities for targeting tumor 23 cells. Among these, antibody-drug conjugates (ADCs) and extracellular targeted protein 24 degradation (eTPD) specifically depend on efficient lysosomal trafficking for activity. 25 However, many tumor antigens exhibit poor internalization, limiting ADC effectiveness. 26 To address this, we developed low-density lipoprotein receptor -targeting chimeras 27 (LIPTACs), leveraging the constitutive endocytic and recycling activity of the LDLR to 28 enhance lysosomal delivery. LIPTACs enable efficient and selective degradation of 29 diverse extracellular membrane proteins. Additionally, by coupling LIPTACs with cytotoxic 30 payloads to generate degrader -drug conjugates, we can achieve superior intracellular 31 delivery and enhanced cytotoxicity compared to conventional ADCs. The dual modality 32 addresses key challenges of inadequate internalization in conventional ADCs and 33 cytotoxic potency for current eTPD strategies. Our findings demonstrate that LDLR -34 mediated trafficking can enhance eTPD and ADCs , providing a hybrid blueprint for 35 developing next-generation antibody therapeutics with broader utility and improved 36 efficacy in cancer treatment. 37 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted June 9, 2025. ; https://doi.org/10.1101/2025.06.06.658366doi: bioRxiv preprint 2

38 Extracellular and membrane -associated proteins represent approximately one-39 third of all protein -coding genes and are key targets for antibody -based therapeutics1. 40 Antibodies provide diverse mechanisms of action in cancer therapy, including receptor 41 inhibition or activation2, immune cell recruitment 3, antibody drug conjugates (ADCs) for 42 targeted toxin delivery 4, and extracellular targeted protein degradation (eTPD) for 43 proteolytic degradation5. Among these, ADCs combine tumor-targeting antibodies with 44 cytotoxic drugs to achieve selective tumor cell killing 6. However, the efficacy of ADCs 45 hinges on the efficient internalization of the antibody -antigen complex to facilitate 46 intracellular drug release in the lysosome . Not all surface antigens undergo productive 47 internalization upon antibody binding, posing a significant limitation to ADC effectiveness7. 48 Recent innovations in bispecific antibodies have sought to overcome this hurdle by 49 enhancing receptor internalization through receptor clustering , such as biparatopic 50 ADCs8,9, or by targeting fast-internalizing receptors in dual -antigen strategies for more 51 efficient lysosomal trafficking10,11. 52 In parallel, eTPD has emerged as a promising therapeutic approach that co -opts 53 natural endolysosomal pathways to selectively degrade membrane -bound and soluble 54 extracellular proteins. Unlike intracellular targeted protein degradation (iTPD) that is 55 driven by recruiting the proteasome for degradation, eTPD mostly shuttles extracellular 56 proteins to the lysosome for degradation. Furthermore, whereas iTPD generally uses only 57 two different widely expressed E3 ligases12 cereblon and von Hippel–Lindau (VHL), eTPD 58 utilizes a wide array of cell surface degrader systems including: neonatal Fc receptor 59 (FcRn)13, glycan binding receptors 14,15, transmembrane E3 ligases 16–18, cytokine 60 receptors19, integrins20, and transferrin receptors 21. Increasing the optionality of cell 61 surface degraders offers greater opportunity for cell specific eTPD. Moreover, ADCs with 62 cleavable linkers require the same lysosomal trafficking system as eTPD. To this end, we 63 wondered if it was possible to develop degrader–drug conjugates (DDCs)—a new class 64 of bifunctional therapeutics that intentionally hybridizes eTPD with ADC for greater 65 efficiency of drug payload delivery. 66 To implement this approach, we targeted the robust recycling low -density 67 lipoprotein receptor (LDLR) as a lysosomal trafficking effector. LDLR naturally internalizes 68 LDL via clathrin-mediated endocytosis22,23 and facilitates its delivery to the lysosome at 69 low pH24. LDLR is upregulated in proliferating cancer cells25,26 and activated T cells27. It is 70 one of the fastest and most efficient internalizers that recycles through the lysosome 71 every 12 minutes28. These features make the LDLR an attractive candidate for eTPD we 72 refer to as LDLR-targeting chimeras (LIPTACs). We show that LIPTACs mediate selective 73 and efficient lysosomal degradation for multiple membrane proteins . Furthermore, by 74 conjugating cytotoxic payloads to LIPTACs or cytokine receptor -targeting chimeras 75 (KineTACs)19, we show these DDCs can boost the potencies of conventional ADCs by up 76 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted June 9, 2025. ; https://doi.org/10.1101/2025.06.06.658366doi: bioRxiv preprint 3 to 20-fold. This work highlights LIPTACs as a versatile platform to enhance payload 77 delivery and broaden the therapeutic utility of antibody-based modalities. 78 79

80 Selection and characterization of LDLR antibodies 81 The extracellular portion of the LDLR contains the ligand binding domain, the 82 epidermal growth factor (EGF)-like domain, and the O-link sugar domain29. We and others 83 have observed that the ligand -binding domain of the LDLR can be shed to varying 84 degrees in cells30,31 transformed with KRAS(G12V) or HER2. This prompted us to select 85 for antibodies against the membrane proximal EGF-like domain of the LDLR in order to 86 preserve its ligand ability for LDL uptake and to enable recruitment of both full-length and 87 cleaved forms (cLDLR) for eTPD. After four rounds of phage selection, we plated 96 88 single phage colonies for screening using an enzyme-linked immunosorbent assay 89 (ELISA) (Extended Data Fig.1a,1b). 90 Phage that passed initial screening were expressed recombinantly as monoclonal 91 fragment antigen-binding (Fab) antibodies for further characterization. The ELISA binding 92 assay against the cLDLR or full -length (flLDLR) identified multiple high affinity binding -93 affinity clones ( Fig. 1a , Extended Data Fig.2a ) such as 142F1 and 142F6. Flow 94 cytometry confirmed that all Fabs bound to LDLR+ MDA-MB-231 cells (Fig. 1b). None of 95 the Fabs exhibited cross-reactivity with the other members of the LDLR family, including 96 the LDLR -related protein 2 (LRP2), LRP8 and VLDLR (Extended Data Fig.2a ). 97 Additionally, the Fabs displayed minimal polyreactive binding across the nonspecific 98 antigen panels 32 (Extended Data Fig.2b ). Epitope binning experiments by b iolayer 99 interferometry (BLI) showed Fabs 142F1 and 142F6 could bind simultaneously, indicating 100 two distinct and non -competitive epitopes ( Fig.1c, Extended Data Fig.3 ). BLI 101 experiments showed that 142F1 and 142F6 bound to cLDLR with affinities of 5.8 nM and 102 16 nM, respectively (Fig.1d). 103 Given that LDLR is essential for regulating plasma cholesterol levels, we then 104 investigated whether our LDLR Fabs affected LDL uptake. We used an Incuyte -based 105 uptake assay with fluorescently labeled (pHrodo) LDL to monitor trafficking to low pH 106 vesicles. We serum starved Hela cells, treated them with LDLR Fabs, heparin, or 107 unlabeled competitor LDL, and then added pHrodo red dye-labeled LDL. As expected, 108 pHrodo red dye -labeled LDL trafficked robustly in both the presence and absence of 109 LDLR Fabs, but was inhibited upon addition of unlabeled LDL and heparin 33 (Fig.1e). 110 These data suggest the LDLR-mediated LDL trafficking remains intact upon Fab binding. 111 112 Design of LIPTAC degraders for EGFR degradation 113 The design of LIPTACs involved a bispecific antibody, with one arm targeting a 114 protein of interest (POI) and the other arm recruiting LDLR to bring the POI and LDLR in 115 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted June 9, 2025. ; https://doi.org/10.1101/2025.06.06.658366doi: bioRxiv preprint 4 close proximity for lysosomal trafficking and eTPD. As a proof-of-concept, we generated 116 LIPTACs to degrade EGF receptor (EGFR), a receptor tyrosine kinase that plays a critical 117 role in the development and progression of various types of cancers34,35. The therapeutic 118 anti-EGFR Cetuximab (Ctx) and anti -LDLR antibody (142F1 or 142F6) were fused to 119 heterodimeric Fc domains respectively, with T350V/L351Y/F405A/Y407V mutations in 120 chain A and T350V/T366L/K392L/T394W mutations in chain B 36. To eliminate effector 121 function for macrophage and NK cell recruitment, we introduced the L234A/L235A/P329G 122 mutations (LALAPG)37 in both Fc chains. To avoid heavy and light chain mispairing, one 123 arm was designed as single chain variable fragment (scFv) and the other in Fab format 124 (Fig.2a). We produced four Ctx-LIPTAC formats as shown in Extended Data Fig.4a and 125 evaluated their degradation efficiency in Hela cells. After 24 h treatment with 50 nM of 126 each LIPTAC, levels of EGFR were quantified by western blotting. We found that the 127 LIPTAC with LDLR antibody as Fab and Ctx in scFv exhibited better EGFR degradation 128 compared to the LDLR scFv format ( Extended Data Fig.4b ). LIPTAC1 containing the 129 higher affinity Fab toward cLDLR showed the highest degradation level, suggesting that 130 the recycling arm with higher affinity possesses better internalization efficiency. 131 The Ctx -KineTAC utilizing the CXCL12 cytokine efficiently degrades EGFR on 132 HeLa cells using the CXCR7 recycling receptor19. Thus, we compared the KineTAC and 133 LIPTAC1. We found the LIPTAC showed somewhat higher degradation efficiency 134 (Fig.2b), indicating that LDLR might internalize and recycle faster than CXCR7. A dose -135 response curve for LIPTAC1 at 24h revealed that LIPTAC1 retained potent EGFR 136 degradation activity at concentrations as low as 5 nM (Fig.2b). 137 Given the tumor -associated nature of LDLR, we further tested degradation in 138 multiple cancer cells. LIPTACs exhibited efficient EGFR degradation on triple -negative 139 breast cancer cell line HCC1143 (Extended Data Fig.4c), the pancreatic cancer cell line 140 PANC-1 (Extended Data Fig.4d), as well as the non-small cell lung cancer cell line NCI-141 H1975 (Extended Data Fig.4e ). Flow cytometry demonstrated that both LIPTAC1 and 142 the CXCL12 KineTAC efficiently degraded surface EGFR on MDA-MB-231 cells (Fig.2c). 143 Moreover, LDLR levels did not change, indicating it was not consumed in the process 144 (Extended Data Fig.4f). Similar findings were observed by western blotting ( Extended 145 Data Fig.4g), suggesting that LDLR was recycled back. Treatment with either arm of the 146 LIPTACs individually at 50 nM did not affect EGFR levels, indicating both targets must be 147 brought together to cause degradation (Fig.2d). Additionally, LIPTAC1 efficiently 148 degraded EGFR with a maximal percent degradation (Dmax) of 86%, on the EGFR high 149 expressing epidermoid carcinoma cell line A431 (Fig. 2e,2f). To assess the specificity of 150 LDLR-mediated protein degradation, we treated LDLR knockout 38 (KO) and control 151 HCC1143 cells with LIPTAC 1. EGFR degradation was less efficient in LDLR KO cells 152 compared to control Cas 9 cells (Fig.2g), indicating the requirement of LDLR for 153 degradation. 154 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted June 9, 2025. ; https://doi.org/10.1101/2025.06.06.658366doi: bioRxiv preprint 5 To further investigate the impact of LIPTAC treatment on the proteome, we 155 conducted quantitative mass spectrometry analysis of surface -enriched lysates 39 156 following LIPTAC1 treatment in MDA-MB-231 cells . We found o nly a few of surface 157 proteins showed changes in abundance (Extended Data Fig.5 ), with EGFR exhibiting 158 the most significant reduction, supporting the high selectivity of LIPTAC1 (Fig.2h). 159 Interestingly, two t etraspanin (TSN) proteins were downregulated upon treatment 160 (Extended Data Fig.5). TSN has been reported to interact with ligand -bound EGFR40,41, 161 suggesting that TSN11 and TSN14 are potential neighbors of EGFR that can also be 162 degraded. Interestingly, LDLR levels increased approximately two -fold (Fig.2h), 163 potentially due to partial competition between the 142F1 antibody and the p roprotein 164 convertase subtilisin/kexin type 9 (PCSK9) binding (Extended Data Fig. 6), which may 165 inhibit PCSK9-mediated LDLR degradation 42,43. The abundance of other LDLR family 166 members observed in our data set, including LRP1, LRP6, and LRP8 , remained 167 unchanged. This data further supports the selectivity of the LDLR antibodies. 168 Immunofluorescence microscopy revealed virtually complete removal of EGFR 169 from the cell surface following 24 h of LIPTAC treatment compared to treatment with PBS, 170 Ctx, or 142F1 clone , further highlighting that LIPTACs induce robust internalization of 171 target proteins (Fig.2i). The colocalization of lysosomal LAMP1 with intracellular EGFR 172 further supported lysosomal shuttling by LIPTACs. Together, our findings demonstrate 173 that LIPTAC-mediated targeted protein degradation is efficient, selective, and dependent 174 on LDLR. 175 176 LIPTAC-mediated degradation of multiple membrane proteins 177 We sought to determine whether the LIPTAC could degrade other therapeutically 178 relevant cell surface proteins. First, we targeted PD-L1, an immune checkpoint expressed 179 in tumor microenvironment that suppresses cytotoxic T cell function 44. We generated a 180 PD-L1-targeting LIPTAC by incorporating anti -PD-L1 Atezolizumab 45 (Atz) Fab with 181 142F1 scFv. Atz -LIPTAC efficiently degraded PD -L1 in MDA -MB-231 cells after 24 h 182 treatment (Fig.3a, 3b). Additionally, to determine degradation mechanisms, cells were 183 pre-treated with Bafilomycin A ( an inhibitor of lysosome acidification 46) or MG132 (a 184 proteasome inhibitor 47) prior to LIPTAC1 treatment. Bafilomycin A inhibited PD -L1 185 degradation, while MG132 did not ( Fig.3c), suggesting that LIPTAC -mediated protein 186 degradation occurs predominantly by delivery to lysosome. 187 Next, we targeted human epidermal growth factor receptor 2 (HER2), which is 188 frequently upregulated in cancer and linked to breast cancer invasiveness and tumor 189 progression48. We generated a HER2 -targeting LIPTAC by incorporating anti -HER2 190 trastuzumab (Traz) scFv with our LDLR Fab. Treatment of MCF7 cells with the Traz -191 LIPTAC resulted in efficient HER2 degradation ( Fig.3d). We then sought to evaluate 192 degradation of multi-pass transmembrane proteins, such as G protein-coupled receptors 193 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted June 9, 2025. ; https://doi.org/10.1101/2025.06.06.658366doi: bioRxiv preprint 6 (GPCRs). We targeted CXCR4, a chemokine receptor i nvolved in tumor growth and 194 metastasis49. The CXCR4 -targeting LIPTAC was built using a CXCR4 antagonizing 195 nanobody (Nb) 50 on one arm and the LDLR 142F1 Fab on the other arm of the Fc. 196 Western blotting showed significant CXCR4 degradation after 24 h treatment of Nb -197 LIPTAC, and not by the Nb monomer (Fig.3e). The antigen-targeting arm of the LIPTAC 198 can be flexibly incorporated into various antibody formats, including Fab fragments, scFvs, 199 and Nbs. 200 Additionally, we investigated CUB domain-containing protein 1 (CDCP1), which is 201 highly overexpressed in RAS -driven cancers and undergoes ectodomain cleavage by 202 extracellular proteases on cancer cells but not healthy cells51–53. We previously developed 203 two antibody clones: 4A06, which targets both full-length and cleaved forms of CDCP1, 204 and CL03, which selectively recognizes the cleaved form for enhanced tumor specificity54. 205 We found that the 4A06 IgG was efficiently internalized, leading to the downregulation of 206 both full-length and cleaved CDCP1 (Fig.3f). In contrast, CL03 IgG alone did not induce 207 CDCP1 degradation. However, the CL03 -based LIPTAC selectively and efficiently 208 degraded cleaved CDCP1, without affecting the full-length form, in PL45 cells (Fig.3f, 3g). 209 Overall, these results highlight the versatility of the LIPTAC platform in selectively 210 degrading a range of extracellular membrane proteins. 211 212 Improved modality of target cell killing via degrader drug conjugates 213 Both toxin delivery by ADCs and protein degradation by eTPD require 214 internalization, delivery, and proteolysis in the lysosome. We hypothesized that eTPD 215 coupled to an ADC could be used to enhance the potency of an ADC in situations where 216 the target of the ADC was not optimized for lysosomal trafficking (Fig.4a). Our data here 217 show that the Ctx-IgG only partially induces EGFR degradation (Fig. 2e). Given that Ctx-218 LIPTAC efficiently degrades EGFR, we selected Ctx ADC as a test case as a DDC. 219 To evaluate the drug release efficiency of DDCs, we conjugated a cathepsin -220 dependent fluorescent probe, LysoLight Deep Red (LLDR), to monitor the lysosomal 221 catabolism of internalized proteins 55. The antibody -probe conjugate was linked via a 222 valine-citrulline ( Vc) linker and remains non-fluorescent until cleavage by lysosomal 223 cathepsins, that generates a bright fluorescent signal and serves as a surrogate for an 224 ADC. We found that Ctx -DDC based conjugates, either in a KineTAC or LIPTAC form, 225 showed much more efficient internalization and lysosomal trafficking compared to Ctx -226 IgG ADC (Fig.4b, Extended Data Fig.7a), suggesting enhanced drug delivery efficiency 227 for the DDC. 228 Next, we evaluated the toxicity of Ctx -DDC by conjugating them to VcMMAE as 229 the cytotoxic payload. MMAE is an auristatin derivative that inhibits tubulin polymerization 230 with a cleavable VC linker56. To compare the cell killing potency of DDCs and Ctx ADC, 231 we conjugated VcMMAE to monomeric 142F1, LIPTAC1, LIPTAC3, KineTAC, and Ctx 232 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted June 9, 2025. ; https://doi.org/10.1101/2025.06.06.658366doi: bioRxiv preprint 7 IgG respectively. The average drug –antibody ratio (DAR) was 2. 6, as estimated using 233 2,4-dinitrophenol (DNP)-PEG4 conjugation with the same method (Extended Table 1). 234 First, we confirmed that antibodies after drug conjugation retained binding to EGFR+ cells 235 (Extended Data Fig.7b). Next, we evaluated the potencies of the DDCs in EGFRhigh A431 236 cells. The Ctx ADC exhibited an IC 50 of 0.9 nM, while the LIPTAC and KineTAC DDCs 237 significantly enhanced the potency by 18 -fold and 7-fold, respectively (Fig.4c). We then 238 assessed the cytotoxicity in EGFR medium MDA-MB-231 cells, which others have shown 239 are relatively insensitive to Ctx-VcMMAE due to the low receptor expression57. After 72 h 240 of treatment, the Ctx ADC induced only moderate cytotoxicity, even at the highest ADC 241 concentration ( Fig.4d). The single arm anti -LDLR 142F1 -VcMMAE also exhibited cell 242 killing weak potency, suggesting that LIPTAC binding is primarily driven by the higher -243 affinity Ctx arm (Fig.4d, Extended Data Fig. 7b). Remarkably, all bispecific DDCs, 244 including LIPTAC1 -VcMMAE, LIPTAC3 -VcMMAE, and KineTAC -VcMMAE, potently 245 killed the target cells (Fig.4d), with IC50 values of 23 pM, 38 pM, and 59 pM, respectively. 246 We observed similar DDC -mediated killing potency on PANC -1 cells which 247 expressed similar levels of EGFR as MDA-MB-231 (Extended Data Fig.7c). None of the 248 antibodies without payload were significantly toxic to cells over three days of treatment 249 (Extended Data Fig. 7d), suggesting that the cytotoxicity is mediated by internalization 250 and release of the degrader conjugated MMAE. The cytotoxicity of ADC and DDCs was 251 minimal on EGFR negative MCF7 cells (Extended Data Fig.7e). 252 We also evaluated the lysosomal dependency of cell killing using an incucyte -253 based killing assay. Accumulation of cell death was still detectable when cells were 254 treated with as low as 0.4 nM of DDCs, while the cell death was not detectable with Ctx -255 ADC or 142F1-ADC (Fig.4e, Extended Data Fig.7f). Bafilomycin A treatment abolished 256 the cytotoxicity of both ADC and DDCs, further indicating the requirement of lysosomal 257 release of cytotoxic payload (Fig.4e). We further treated cells with chemical inhibitors of 258 several endocytic pathways. Inhibitors of proteasomes (MG132) or caveolar -mediated 259 endocytosis (Nystatin) 58 did not significantly affect the cytotoxicity of the DDCs, 260 highlighting receptor-mediated lysosomal delivery (Fig. 4f). Together, these data suggest 261 that DDCs improve lysosomal delivery of cytotoxic payload and enhance the cytotoxic 262 potency. 263 264 Degrader-drug conjugates improved cytotoxicity on cells with moderate antigen 265 expression levels 266 To evaluate whether DDCs can enhance the drug delivery efficiency of these 267 ADCs, we selected four clinically relevant ADC targets: anti-B cell maturation antigen 268 (BCMA) Belantamab59,60, anti-CD19 Loncastuximab61, anti-trophoblast cell surface 269 antigen 2 (TROP2) Sacituzumab62, and anti-HER2 Trastuzumab63. We engineered these 270 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted June 9, 2025. ; https://doi.org/10.1101/2025.06.06.658366doi: bioRxiv preprint 8 therapeutically relevant antibodies into corresponding LIPTAC formats and conjugated 271 them to VcMMAE as similar DARs to EGFR-targeting DDCs. 272 BCMA is highly expressed on malignant plasma cells and represents a validated 273 target in multiple myeloma . Belantamab mafodotin (Blenrep) is a novel ADC composed 274 of an anti -BCMA antibody conjugated to MMAF via a non -cleavable linker 64. Our 275 Belantamab-LIPTAC demonstrated more efficient internalization than Belantamab in 276 RPMI-8226 cells (Fig.5a). Both the conjugated ADCs and DDCs retained binding affinity 277 comparable to their unconjugated counterparts (Extended Data Fig.8a). In the luciferase-278 based viability assay, LIPTAC DDCs improved potency in RPMI -8226 cells by 4 -fold, 279 reducing the EC₅₀ from 85 nM to 22 nM. The effect was not observed in MM1.S cells with 280 higher BCMA expression (Extended Data Fig.8b). Notably, by Annexin V and propidium 281 iodide staining, LIPTAC DDCs significantly enhanced cell death by 18-fold in RPMI-8226 282 and 5-fold in MM1.S cells ( Fig.5b, Extended Data Fig.8c). This difference may be due 283 to incomplete killing by DDCs in BCMA+ cells, with Annexin V staining captur ing early 284 apoptotic events that are not detected by viability assays. Overexpression of BCMA in 285 RPMI-8226 cells rendered both LIPTAC DDC and Belantamab ADC comparably potent 286 (Extended Data Fig.8d , 8e), suggesting that LIPTAC -mediated trafficking offers the 287 greatest advantage in cells with moderate antigen expression. 288 Next, we investigated CD19, a protein strictly and ubiquitously expressed on B 289 cells across multiple developmental stages. Loncastuximab tesirine is an anti-CD19 ADC 290 where the payload is a PBD dimer that inhibits DNA metabolic processes via a cleavable 291 dipeptide linker61. CD19 has rapid internalization kinetics, with colocalization of anti-CD19 292 antibodies and lysosomal compartments observed as early as 15 minutes post -293 treatment65. CD19-LIPTAC did not enhance internalization (Fig.5c, Extended Data 294 Fig.9a). Although the LIPTAC improved CD19 degradation, Loncastuximab (Lonca) itself 295 efficiently internalized and degraded CD19 (Fig.5d). Consequently, the cytotoxicity of the 296 LIPTAC DDC was comparable to Lonca -ADC (Fig.5e, Extended Data Fig. 9b,9c), 297 indicating minimal benefit from the LIPTAC format in this context. 298 Similarly, Trastuzumab is known to induce HER2 internalization and 299 downregulation upon binding 66,67. Although HER2 -LIPTAC modestly improved 300 internalization (Extended Data Fig.10a), it did not enhance the cytotoxicity in HER2high 301 BT474 cells or HER2low T47D cells (Extended Data Fig.10b,10c). A similar outcome was 302 observed for TROP2 -targeting Sacituzumab ADCs. TROP2 undergoes rapid 303 internalization upon antibody binding, with a reported internalization half -life (t₁ /₂) of 304 approximately 30 minutes 68. Despite moderate improvements in internalization with 305 TROP2-LIPTAC ( Extended Data Fig. 10d), no significant increase in cytotoxicity was 306 observed in TROP2high A431 or TROP2medium PL45 cells (Extended Data Fig.10e-10g). 307 Interestingly, LIPTAC DDC did show greater potency compared to monomeric 308 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted June 9, 2025. ; https://doi.org/10.1101/2025.06.06.658366doi: bioRxiv preprint 9 Sacituzumab ADC (Extended Data Fig. 10e-10g), suggesting that increased avidity or 309 receptor clustering could contribute to improved lysosomal delivery for the ADC. 310 Overall, these results indicate that LIPTAC-based DDCs can enhance cytotoxicity 311 in specific contexts, particularly in target-low settings or when the parental ADC is limited 312 by suboptimal internalization. When the ADC is in high abundance and a good recycler 313 the benefit of a DDC is clearly less. These findings underscore the importance of target 314 selection and receptor biology in guiding DDC design. 315 316

317 TPD has emerged as an exciting new modality for potent degradation of 318 functionally important proteins inside and outside the cell. Whereas iTPD has principally 319 focused on two intracellular E3 ligases, due to limitations of chemical materials that can 320 bind cereblon and VHL, eTPD can readily access scores of extracellular recycling 321 receptors with routine antibody generation and established bispecific constructs or 322 conjugations. Thus, eTPD allows for greater tissue selectivity by proper choice of the 323 degrader system that matches the expression of the POI on disease versus healthy 324 tissues. The LDLR-based LIPTAC expands opportunities for eTPD. LDLR is an attractive 325 recruiter for catalytic and durable degradation due to its rapid recycling kinetics . It can 326 complete one endocytic cycle approximately every 12 minutes and recycle up to 150 327 times69. Here, we show that LIPTAC enables efficient, selective, and LDLR -dependent 328 lysosomal degradation of a diverse range of therapeutically relevant membrane proteins, 329 including both single- and multi-pass transmembrane proteins, as well as the cleaved, 330 tumor-specific neoepitopes such as CDCP1. 331 We further demonstrate that eTPD technologies such as LIPTAC and KineTAC 332 can complement ADCs by improving intracellular trafficking and lysosomal delivery, which 333 are critical determinants of ADC efficacy 70. Recent biparatopic and bispecific ADC 334 designs ( e.g., HER2×CD63 10, HER2×PRLR 71, and TROP2×APLP2 11) have leveraged 335 receptor clustering and dual -antigen engagement to enhance lysosomal targeting 9. 336 Building on this, we show that antibody -based degraders can similarly drive efficient 337 endolysosomal delivery and serve as potent payload vehicles. In head -to-head 338 comparisons, EGFR -targeting Ctx-DDCs outperformed Ctx-based ADCs in lysosomal 339 trafficking and cytotoxicity. Tumor selectivity was conferred by the high-affinity Ctx arm72, 340 as LDLR-targeting alone exhibited weak activity. LIPTAC-DDC exhibited minimal toxicity 341 to cells without EGFR expression, further highlighting that LIPTAC specificity stems from 342 the tumor antigen -binding domain. The bispecific nature that binds the POI and the 343 degrader system can allow for greater tumor specificity by choosing degrader systems 344 matched to the tumor and not healthy cells. The LDLR is significantly upregulated in many 345 cancer cells relative to normal cells providing an advantage in this regard. 346 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted June 9, 2025. ; https://doi.org/10.1101/2025.06.06.658366doi: bioRxiv preprint 10 The therapeutic potency of ADCs depends on antigen abundance, receptor 347 internalization, and payload potency. Our results suggest that LIPTACs can enhance 348 payload delivery particularly in contexts where conventional ADCs are limited, such as 349 receptors with moderate expression or poor internalization (e.g., BCMA on RPMI -8226 350 cells). However, this advantage diminished in BCMA -overexpressing cells and was not 351 observed for efficiently internalizing targets like CD19 or TROP2, where LIPTAC and 352 KineTAC formats performed comparably to approved ADCs (e.g., Loncastuximab, 353 Sacituzumab). Reduced avidity and lower binding affinities of LIPTACs also likely 354 contributed to these differences. 355 The DDC format may expand druggability to challenging targets with poor 356 internalization kinetics or low surface abundance, such as post -translational 357 modifications73,74, tumor-specific neo -antigens75, or glycosylphosphatidylinositol (GPI)-358 anchored proteins 76. Future optimization of antigen -targeting arms and use of 359 heterobifunctional degraders like LIPTAC could reduce DAR requirements, minimize off-360 target effects, and broaden the therapeutic window. 361 Finally, a major limitation of eTPD approaches is incomplete degradation, which 362 may be insufficient to fully suppress oncogenic signaling. The DDC format overcomes 363 this by delivering cytotoxic payloads directly to tumor cells, independent of full protein 364 clearance. This modular platform can be expanded to deliver diverse therapeutic cargos, 365 including radionuclides, kinase inhibitors, or proteolysis-targeting chimeras (PROTACs). 366 In summary, the DDC modality represents a promising next -generation hybrid strategy 367 for targeted cancer therapy. 368 369

370 We thank Dr. Brandon Holmes, Dr. Jon Ostrem and Dr. Rohit Bhadoria for their 371 assistance with discussions, and the Wells Lab broadly for helpful discussions and 372 expertise. We thank Paul Burroughs for providing the BCMA overexpressing RPMI8226 373 cells. We are grateful to generous support from NIH -1R01CA248323-01(J.A.W), NIH -374 R35GM122451 (J.A.W. ), the Hind Professorship in Pharmaceutical Sciences (J.A.W ), 375 and R01CA276207 (J.A.O.). K.S. is supported by a Helen Hay Whitney Foundation 376 Fellowship. Z.Y. is supported by a National Institute of General Medical Sciences F32 377 Postdoctoral Fellowship. K.K. is supported by a graduate fellowship funded by the 378 National Science Foundation. 379 380 Author Contributions 381 F.Z. and J.A.W. conceived and designed the study. F.Z. performed phage display, 382 antibody screening and characterization experiments. F.Z., Y.W., Y.Z., S.G. and K.K. 383 cloned and expressed the recombinant proteins. F.Z. and Z.Y. performed the western 384 blotting experiments. F.Z., Y.W., and K.K. labeled the antibodies with dyes or payloads 385 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted June 9, 2025. ; https://doi.org/10.1101/2025.06.06.658366doi: bioRxiv preprint 11 and performed the internalization and cytotoxicity assays. K.S. and T.M.P.C. performed 386 the proteomics and processed the data. K.M. provided the confocal microscopy data. A.I 387 and J.A.O. provided the LDLR KO cells. F.Z. and J.A.W. wrote the manuscript and all 388 authors reviewed and edited the manuscript. 389 390 Declaration of interests 391 F.Z. and J.A.W. have filed patent applications relating to the LDLR-targeting chimeras 392 and the degrader -drug conjugates. J.A.W. is a founder of EpiBiologics and K.K. is a 393 founding advisor. Both J.A.W. and K.K. hold stock in the company. 394 395

396 Plasmid construction 397 All the IgGs were constructed in a pcDNA3.4 vector that expresses the light chain and 398 heavy chain, respectively for mammalian expression. For generating bispecific antibody, 399 the heavy chain variable regions or scFvs were cloned into zymework -A mutant Fc and 400 zymework-B Fc sequences respectively. Antigen was cloned into pFuse vector with IL-2 401 signal peptide followed by t obacco etch virus (TEV) protease, Fc, and Avitag sequence 402 in the C-terminus. All the Fabs were constructed in a dual-expression pBL347 vector that 403 expresses the light chain and the heavy chain with the pelB and the stII signal peptides, 404 respectively, for the periplasm expression. 405 406 Cell lines 407 HEK293T, HeLa, MDA-MB-231, PANC -1 and A431 were cultured in DMEM 408 (ThermoFisher Scientific) with 10% fetal bovine serum (FBS) and 1% penicillin –409 streptomycin. M CF7 cells were cultured in DMEM with 10% FBS, 1% penicillin -410 streptomycin, 1% sodium pyruvate, and 1% non -essential amino acid. NCI -H1975, 411 HCC1143, PL45, BT474, Raji, Ramos, RPMI8226, and MM1.S cells were cultured in 412 RPMI-1640 (ThermoFisher Scientific) with 10% FBS and 1% penicillin –streptomycin. 413 HCC1143 LDLR KO cells 38 were provided by Dr. James Olzmann and were 414 supplemented with 100 μg/mL hygromycin ( Sigma-Aldrich). of Expi293F cells were 415 cultured in FreeStyle 293F medium. 416 417 Phage display 418 Phage selection was done as described previously77. In brief, library E and UCSF library 419 were incubated with streptavidin-coated magnetic beads pre-conjugated with biotinylated 420 Fc protein to remove nonspecific binders. Unbounded phages were then incubated with 421 streptavidin-coated magnetic beads pre -conjugated with biotinylated cLDLR -TEV-Fc 422 antigens. After 4 washes, antigen -bound phages were eluted from beads by incubating 423 with 1 μM TEV protease for 20 min. In total, four rounds of selections were performed 424 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted June 9, 2025. ; https://doi.org/10.1101/2025.06.06.658366doi: bioRxiv preprint 12 with a decreasing concentration of cLDLR-antigen (1000, 50, 20, 10 nM). From round 3, 425 the phage library was first enriched by protein A magnetic beads to deplete nondisplayed 426 or truncated Fab phage before each round of the selection. 427 428 Phage ELISA 429 384-well Maxisorp plates were coated with Neutravidin (10 μg/mL) overnight at 4 °C and 430 subsequently blocked with BSA (2% w/v) for 1 h at RT. 20 nM biotinylated cLDLR antigens 431 were captured on the NeutrAvidin-coated wells for 30 min followed by the addition of 1:5 432 diluted single-colony phage for 1 h. The secondary antibodies were either a horseradish 433 peroxidase (HRP)-conjugated anti-M13 phage antibody (Sino Biological) for phage ELISA 434 or an anti-human IgG antibody (Sigma-Aldrich) for recombinant protein ELISA. The ELISA 435 plates were washed three times after each incubation, and antibody binding was detected 436 by TMB substrate (VWR) and read at 450 nm. 437 438 Protein expression 439 IgGs and antigen were expressed in Expi293F cells in a 30 mL scale. In brief, 24 μg of 440 DNA was added to 3 mL of OptiMEM, followed by 24 μL of FectoPro transfection reagents. 441 After 10 min of incubation, 27 mL of Expi293F cells at 3 millions/mL were added and 442 shake at 37°C. On the second day, 300 μL of 300 mM vaporic acid and 270 μL of 45% 443 glucose were added to the cells. After 5 days of transfection, cells were harvested, spun 444 down at 4,000 g for 20 min and filtered by 0.45 μm steri-flip. Supernatants were then 445 incubated with Sepharose A resin for 2 h, proteins were then eluted by 0.1 M acetic acid 446 and neutralized by Tris pH 11. Proteins were buffer changed 3 times in PBS in amicon 447 tubes. Fabs were expressed in Escherichia coli C43 (DE3) Pro+ grown in an optimized 448 TB autoinduction medium at 37 °C for 6 h, cooled to 30 °C for 18 h. Cells were harvested 449 by centrifugation and lysed using B-PER lysis buffer. The lysate was incubated at 60 °C 450 for 20 min and centrifuged to remove the inclusion body. The Fabs were purified by 451 Sepharose A resin via affinity chromatography and buffer exchanged in PBS for further 452 characterization. Purity and integrity of all proteins were assessed by SDS–PAGE. 453 454 Recombinant protein ELISA 455 384-well Maxisorp plates were coated with Neutravidin (10 μg/mL) or anti-histag antibody 456 (Invitrogen, 2 μg/mL) overnight at 4 °C and subsequently blocked with BSA (2% w/v) for 457 1 h at RT. 20 nM of antigens were captured onto pre -coated wells for 1h. Recombinant 458 full-length LDLR, LRP2, LRP8, and VLDLR proteins were purchased from 459 ACROBiosystems. For polyspecificity ELISA, autoantigens cardiolipin (Sigma, 50 μg/mL), 460 insulin (Sigma, 1 μg/mL), l ipopolysaccharid (LPS, InvivoGen, 10 μg/mL), and single -461 stranded DNA (ssDNA, Sigma, 1 μg/mL), were directly coated onto plates overnight 4 °C. 462 After three times of wash, serially diluted Fabs were added to the plates and incubated 463 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted June 9, 2025. ; https://doi.org/10.1101/2025.06.06.658366doi: bioRxiv preprint 13 for 1 h at RT. After three times of wash, 1:5000 diluted peroxidase-anti-human IgG (H+L) 464 (Jackson ImmunoResearch) were added to the plates and incubated for 30 min. After 465 three times of wash, antibody binding was detected by TMB substrate (VWR), quenched 466 by 1 M phosphoric acid, and read at 450 nm. 467 468 Flow cytometry 469 Cells were collected by centrifugation at 400 g for 5 min. Pellets were washed once with 470 PBS + 1% BSA. Cells were incubated with fluorophore-conjugated antibodies in PBS + 1% 471 BSA for 15 min at RT or 30 min at 4 °C. Cells were washed three times and resuspended 472 in cold PBS for flow analysis. Antibodies used included APC anti-LDLR (Invitrogen, Cat# 473 MA5-40994, 1:400), PE anti -human EGFR (Invitrogen, Cat# MA5-28544,1:400), APC 474 anti-human CXCR4 (Biolegend, Cat#306509, 1:400), Alexa fluor 647 goat anti-human IgG 475 (H+L) ( Invitrogen, Cat# A-21445, 1:1000), APC Annexin V ( BioLegend, Cat# 640920, 476 1:500), Alexa fluor 647-conjugated protein A (Invitrogen, Cat# P21462, 1:1000). Dead cell 477 staining included p ropidium Iodide (Biolegend, Cat#421301, 1:250), and LIVE/DEAD™ 478 fixable violet dead cell stain kit (Invitrogen, Cat# L34964). Flow cytometry was performed 479 using a CytoFLEX cytometer (Beckman Coulter, v.2.3.1.22) and Cyt oExpert software 480 (v.2.3.1.22). Data were analyzed with FlowJo (v.10.8.0). 481 482 Biolayer interferometry 483 BLI experiments were performed at room temperature using an Octet RED384 instrument 484 (ForteBio). 20 nM biotinylated antigens were immobilized to an optically transparent SA 485 biosensor (ForteBio). Different concentrations of antibodies in kinetics buffer (PBS, 0.05% 486 Tween-20, 0.2% BSA) were used as the analyte in a 384 -well microplate (Greiner Bio -487 One). Affinities (KDs) were calculated by a global fit analysis and by a 2:1 heterogeneous 488 ligand model using the Octet RED384 Data Analysis HT software. 489 490 Epitope binning by BLI 491 Anti-LDLR antibodies were binned into epitope specificities using an Octet RED384 492 system. 20 nM of biotinylated cLDLR-Fc antigens were captured using streptavidin 493 biosensors (Fortebio). After antigen loading, a saturating concentration of antibodies (200 494 nM) was added for 10 min. Competing concentrations of antibodies (40 nM) were then 495 added for 5 min to measure binding in the presence of saturating antibodies. All incubation 496 steps were performed in PBS/0.05% Tween-20/0.2% BSA. For PCSK9 epitope binning, 497 all incubation steps and protein dilution were performed at acidic endosomal pH to 498 increase PCSK9 binding affinity78. 200 nM PCSK9 D374Y (AcroBiosystems) was added 499 for 10 min after antigen loading. Then 40 nM of PCSK9 D374Y, 142F1, or 142F6 was 500 added for 5 min. 501 502 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted June 9, 2025. ; https://doi.org/10.1101/2025.06.06.658366doi: bioRxiv preprint 14 LDL uptake assay 503 LDL uptake was measured by the Image-iT™ pHrodo™ Red Low Density Lipoprotein 504 Uptake Kit (Thermo Scientific, Cat#I34360) following manufacturer’s instructions. Briefly, 505 HeLa cells were seeded at 5000 cells/well on a 96-well polystyrene tissue culture treated 506 plate (Corning, Cat#3596). The next day, media was removed and cells were serum 507 starved for 12 h. Next, cells were pretreated with unlabeled LDL, heparin, and anti-LDLR 508 Fabs for 30 min at 37  °C respectively, followed by pHrodo™ red-labeled LDL treatment. 509 Cells were then imaged by Incucyte (Sartorius) every 1 to 2 h. I nternalization was 510 calculated by total integrated intensity (ROCU x μm2/image) on the Incucyte software. 511 512 Degradation experiments 513 Cells were plated in 6- or 12-well plates and grown to ~70% confluency before treatment. 514 On the next day, cell culture medium was aspirated, various concentrations of antibodies 515 in 1 mL of culture medium were then added to each well. Cells were incubated for 24 h 516 at 37 °C for flow cytometry or western blotting experiments. 517 518 Western blotting 519 Cells were lifted with PBS+0.05% EDTA, transferred to Eppendorf tubes, spun down at 520 500 g for 4 min, and wash 2 times with PBS. Then cells were lysed with 1× RIPA lysis 521 buffer (EMD Millipore) with cOmplete mini protease inhibitor cocktail (Sigma -Aldrich) at 522 4 °C for 20 min. Lysates were centrifuged at 20,000g for 10 min at 4 °C. Protein amounts 523 were quantified by Rapid Gold BCA Protein Assay Kit (Pierce). Lysates were mixed with 524 4× Nupage LDS Sample Buffer ( Invitrogen) and 2-mercaptoethanol, and then run on 525 NuPAGE™ 4-12% Bis Tris Protein Gels (Thermo Fisher Scientific ). Proteins were 526 transferred to polyvinylidene difluoride membranes using the iBlot2 Western Blotting 527 Transfer System ( Thermo Scientific). Membranes were blocked with TBS + 5% BSA + 528 0.5% Tween for 1h, and stained with primary antibodies overnight. After three times of 529 washing, membranes were stained with secondary antibodies for 1 h at RT. After three 530 times of washing, membranes were imaged with a LICOR imager or the ChemiDoc MP 531 imaging system (BioRad). Antibodies used included rabbit anti -human EGFR ( Cell 532 Signaling Technology , Cat# 4267S, 1:1000), rabbit anti-human PD-L1 (Cell Signaling 533 Technology, Cat#13684S, 1:1000), rabbit anti-human CXCR4 (Cell Signaling Technology, 534 Cat#64837S, 1:1000), rabbit anti-human ERBB2 (Cell Signaling Technology, Cat#4290S, 535 1:1000), rabbit anti-human CDCP1 ( Cell Signaling Technology, Cat# 13794S, 1:1000), 536 rabbit anti-human CD19 (Cell Signaling Technology, Cat# 90176T, 1:1000), mouse anti-537 human β-tubulin (Cell Signaling Technology, 3873S, 1: 3000), goat anti-human LDLR 538 (R&D Systems , Cat#AF2148, 1:1000) , IRDye 800CW goat anti -rabbit IgG ( LI-COR 539 Biosciences, Cat# 926-32211), IRDye 680RD goat anti-mouse IgG (LI-COR Biosciences, 540 Cat# 926 -68070, 1:5000 ), IRDye 800CW donkey anti-goat IgG (LI -COR Biosciences, 541 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted June 9, 2025. ; https://doi.org/10.1101/2025.06.06.658366doi: bioRxiv preprint 15 Cat# 926 -32214, 1:5000), p eroxidase goat a nti-rabbit IgG (H+L) (Jackson 542 ImmunoResearch, Cat# 111-035-144, 1:5000). 543 544 Confocal microscopy 545 HeLa cells were plated on the chambered coverslip (Ibidi, 8-well uncoated) and incubated 546 for 24 h at 37°C. Cells were then treated with 50 nM bispecific or control antibodies in 547 complete growth medium. After 24 h of incubation at 37°C, medium was aspirated, and 548 cells were washed with PBS. Cells on the coverslips were fixed with paraformaldehyde 549 (PFA) for 15min at RT, then permeabilized with 0.1% Triton -X in PBS for 10 min at RT. 550 After washed 3 times by PBS, the resulting sample were stained with anti-LAMP1 rabbit 551 antibody (Cell Signaling Technology, Cat# 9091T), anti-EGFR mouse antibody (Thermo 552 Scientific, Cat# MA5-13070), and DAPI (Cell Signaling Technologies) . Goat anti -rabbit 553 IgG 488 (Invitrogen, Cat#A -11008) and goat anti -mouse IgG 647 (Invitrogen, Cat#A -554 21240) were stained for visualization. Samples were imaged using a Nikon Ti Microscope 555 with a Yokogawa CSU-22 spinning disk confocal and a 60x objective lens; 405-, 488- and 556 647-nm lasers were used to image DAPI, LAMP1 and EGFR, respectively. Images were 557 deconvoluted and processed using NIS -Element (v5.21.03) and Fiji software (v2.1.0) 558 packages. 559 560 Cell culture/stable isotope labeling using amino acids in cell culture (SILAC) 561 labeling 562 MDA-MB-231 cells were grown in DMEM for SILAC (Thermo Fisher) with 10% dialyzed 563 FBS (Gemini). Medium was also supplemented with either light L-[12C6,14N2]-lysine/l-564 [12C6,14N4]-arginine (Sigma) or heavy L-[13C6,15N2]-lysine/L-[13C6,15N4]-arginine 565 (Cambridge Isotope Laboratories). Cells were maintained in SILAC medium for five 566 passages to ensure complete isotopic labeling. Heavy-labeled cells were treated with 567 PBS control and light-labeled cells were treated with 50 nM bispecific LIPTAC for 48 h 568 before cells were collected. Cells were then used to prepare surface -proteome 569 enrichment. 570 571 Mass spectrometry 572 For proteomic analysis, cells were processed following established cell surface capture 573 methods39. Approximately 2 million SILAC-labeled cells were first washed in PBS (pH 6.5) 574 before the glycoproteins were oxidized with 1.6 mM sodium periodate (Sigma) in PBS 575 (pH 6.5) for 20 min at 4 °C. Cells were then biotinylated via the oxidized vicinal diols with 576 1 mM biocytin hydrazide (Biotium) in the presence of 10 mM aniline (Sigma) in PBS (pH 577 6.5) for 90 min at 4 °C. Cell pellets were lysed with a 2× dilution of commercial RIPA 578 buffer (Millipore) supplemented with 1× protease inhibitor cocktail (Sigma) and 2 mM 579 EDTA (Sigma) for 10 min at 4 °C. Cells were further disrupted with probe sonication (20% 580 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted June 9, 2025. ; https://doi.org/10.1101/2025.06.06.658366doi: bioRxiv preprint 16 amplitude, 5 min, 4 °C), followed by cell debris removal (20,000xg, 10 min, 4 °C), and 581 the clarified cell lysates were then incubated with 50 µL of high-capacity NeutrAvidin-582 coated agarose beads (Thermo) in Poly-Prep chromatography columns (Bio-Rad) for 2 h 583 at 4 °C to isolate biotinylated glycoproteins. To enrich for biotinylated proteins, the resin 584 was washed sequentially with 5 mL of 1× RIPA (Millipore) plus 1 mM EDTA, 5 mL high-585 salt PBS ( 20 mM phosphate (pH 7.4) with 1 M NaCl (Sigma)) and 5 mL of denaturing 586 urea buffer (50 mM ammonium bicarbonate and 2 M urea). All wash buffers were heated 587 to 42 °C before use. Proteins on the beads were next reduced, carbidomethylated, 588 digested and desalted using the Preomics iST mass spectrometry sample preparation kit 589 (Preomics) per the manufacturer’s recommendations. After desalting, samples were dried, 590 resuspended in 0.1% formic acid and quantified using the Pierce peptide quantification 591 kit (Thermo Scientific) before liquid chromatography–tandem mass spectrometry analysis. 592 Liquid chromatography –tandem mass spectrometry was performed using a Bruker 593 NanoElute chromatography system coupled to a Bruker timsTOF Pro mass spectrometer. 594 Peptides were separated using a prepacked IonOpticks Aurora (25 cm × 75 μm) C18 595 reversed-phase column (1.6-µm pore size, Thermo) fitted with a CaptiveSpray emitter for 596 the timsTOF Pro CaptiveSpray source. For all samples, 200 ng of resuspended peptides 597 was injected and separated using a linear gradient of 2 –23% solvent B (solvent A: 0.1% 598 formic acid and 2% acetonitrile; solvent B: acetonitrile with 0.1% formic acid) over 90 min 599 at 400 µl min–1 with a final ramp to 34% B over 10 min. Separations were performed at 600 a column temperature of 50 °C. Data -dependent acquisition was performed using a 601 timsTOF PASEF tandem mass spectrometry method (TIMS mobility scan range of 0.70–602 1.50 V•s cm –2, mass scan range of 100 –1,700 m/z, ramp time of 100 ms, 10 PASEF 603 scans per 1.17 s, active exclusion of 24 s, charge range of 0 –5 and minimum MS1 604 intensity of 500). The normalized collision energy was set at 20. 605 606 Mass spectrometry data analysis 607 LC-MS-MS data was analyzed using PEAKS online Xpro 1.6 (Bioinformatics Solutions 608 Inc.; Ontario, Canada). Spectral searches were performed in PEAKS Q (de novo assisted 609 Quantification) mode. The precursor mass error tolerance was set to 20 ppm, and the 610 fragment mass error tolerance was set to 0.5. Peptides containing 6 and 45 amino acids 611 in length were then searched in a semi -specific trypsin/LysC digest mode against a 612 proteome file that contains human cell surface proteins79. Carbidomethylation (+57.0214 613 Da) on cysteines was a set static modification; methionine oxidation (+15.994), and the 614 isotopic labels (13C(6)15N(2); 13C(6)15N(4)) were set variable modifications. Quantified 615 peptides were matched between three experimental replicates and peptide enrichments 616 were normalized based on the total ion chromatograph (TIC). SILAC -labeled protein 617 ratios were further analyzed if the proteins were identified by more than one peptide and 618 present in at least two experimental replicates. A p-value of 0.05 and two-fold protein ratio 619 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted June 9, 2025. ; https://doi.org/10.1101/2025.06.06.658366doi: bioRxiv preprint 17 differences were set as cut -offs to determine if protein abundance differences were 620 significant between vehicle treated cells and LIPTAC -treated cells. Proteomic data is 621 available on PRIDE with PXD064642 accession code. 622 623 Lysolight deep red assay 624 Antibodies were labeled using the L ysoLight Antibody Labeling Kits (Invitrogen, Cat# 625 L36003) following manufacturer’s instructions. Briefly, antibodies were labeled with LLDR 626 with a molar ratio of 1:6 in the presence of 100 mM sodium bicarbonate (pH 8.4) for 2 h 627 at RT. Antibodies were then purified with 7k Zeba dye and biotin removal columns 628 (Thermo Scientific, Cat# A44297). Cells were seeded at 5000/well on a 96-well 629 polystyrene tissue culture treated plate (Corning, Cat# 3596). The next day, media was 630 removed, treated with LLDR-labeled antibodies, and then imaged on the Incucyte every 631 2 h for 72 h. 632 633 In vitro ADC assays 634 IgGs were labeled with NHS ester -PEG4-ValCit-PAB-MMAE (BroadPharm, Cat# BP-635 25503) with 1:6 or 1:10 molar ratio at RT for 2 h with 100 mM sodium bicarbonate. 636 Antibodies were then desalted using the Pierce Zeb a desalt spin columns (Thermo 637 Scientific). For adherent cells, 5000 cells/well were seeded on a 96-well polylysine-coated 638 white plate (Corning, Cat#3917). The next day, media was aspirated, MMAE-labeled IgGs 639 were added and incubated for 72 h. For suspension cells, 6000 cells/well were incubated 640 with ADCs or DDCs for 96 hours . Viability was measured using CellTiter -Glo Reagent 641 (Promega). For incucyte-based killing assay, the cells were treated with ADCs or DDCs, 642 cytotoxic green dye (Sartorius, Cat#4633, 1:10000) or p ropidium Iodide (Biolegend, 643 Cat#42130, 2 µg/mL). For compound inhibitor assay, 0.8 nM of DDCs were incubated 644 with 1 µg/mL of p ropidium Iodide with or without 50 nM Bafilomycin A ( Santa Cruz 645 Biotechnology, Cat#sc-201550A), 50 nM MG132 ( Selleck Chemicals, Cat#S2619), or 1 646 µM Nystatin (MedChem Express, Cat# HY-17409), respectively. 647 648 Characterization of DAR 649 IgGs were side-by-side labeled with DNP-PEG4-NHS ester (MedChem Express, Cat# 650 HY-140614) with the same molar ratio as MMAE conjugation, and incubated at RT for 2 651 h with 100 mM sodium bicarbonate. Antibodies were then desalted using the Pierce Zeba 652 desalt spin columns (Thermo Scientific). The absorbance of conjugated antibodies at 280 653 nm and 360 nm was measured by UV-Vis spectrophotometer. The correction factor (CF) 654 was determined by measuring A280 and A360 of the pure 100 µM DNP-PEG4-NHS ester 655 solution. 656 DAR = !"#$ %&'( ÷ ( !)*$ +(!"#$ × /0) %2345678 ) 657 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted June 9, 2025. ; https://doi.org/10.1101/2025.06.06.658366doi: bioRxiv preprint 18 Figures and Legends 658 659 660 Fig.1. Characterization of cLDLR -specific antibodies. a, ELISA binding of 661 recombinant Fabs against the cLDLR antigen. Absorbance was read at 450 nm. An anti-662 CDCP1 Fab 4A0654 was used as a negative isotype control. Each sample was tested in 663 biological duplicate and error bars represented standard deviations. b, Flow cytometry of 664 different Fabs binding to LDLR + MDA-MB-231 cells. 50 nM of each Fab was incubated 665 with cells for 30 min and washed twice, followed by AF647 conjugated goat anti -human 666 IgG (H+L) antibody staining for 15 min. c, Epitope binning of two anti-LDLR Fabs, 142F1 667 and 142F6 revealed two different epitopes on cLDLR. Biotinylated cLDLR was captured 668 using a streptavidin biosensor and indicated antibodies at a concentration of 200 nM were 669 incubated for 10 min followed by incubation with 50 nM of the second competing 670 antibodies for 5 min. d, BLI analysis of 142F1 and 142F6 Fabs to estimate their affinities 671 to cLDLR. Biontinylated cLDLR was immobilized via the streptavidin biosensor and 672 varying concentrations of each Fab was injected. Black lines were the experimental trace 673 obtained from the BLI experiments and colored lines were the global fits. e, Internalization 674 of pH-sensitive Phrodo red-labeled LDL on HeLa cells after 30 min pretreatment with LDL, 675 heparin, or each Fab, respectively. Total integrated intensity is calculated by ROCU x 676 μm2/image on the Incucyte software. Each sample was tested in biological triplicate and 677 error bars represent standard deviations. Statistics were calculated by one-way ANOVA 678 and Holm-Sidak multiple comparisons test. 679 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted June 9, 2025. ; https://doi.org/10.1101/2025.06.06.658366doi: bioRxiv preprint 19 680 681 Fig.2. Generation of LIPTACs for degradation of EGFR. a, Schematic illustration of 682 LDLR-Ctx LIPTAC bispecific constructs. b, Western blot showing degradation of total 683 EGFR on HeLa cells following 24 h of treatment with Ctx-LIPTAC or 50 nM Ctx-KineTAC, 684 and 50 nM Ctx control IgG. Data represented three biological replicates. Percent EGFR 685 levels were quantified by ImageJ relative to PBS control. c, Changes in surface EGFR 686 based on flow cytometry analysis on MDA -MB-231 cells following 24 h of 50 nM 142F1 687 isotype IgG, Ctx IgG, Ctx -LIPTAC, or Ctx -KineTAC treatment. Percent EGFR was 688 determined by median fluorescence intensity (MFI) of the PE fluorescence channel of live 689 cells. Each sample was tested in biological triplicate and error bars represented standard 690 deviation. Statistics were calculated by unpaired two -tailed student t test. ***P < 0.001. 691 ****P < 0.0001. d, Western blot showing degradation of total EGFR on MDA-MB-231 cells 692 after 24 h treatment of 50 nM Ctx, KineTAC, monomeric LDLR isotype, or varying 693 concetrations of LIPTACs . Data represented three biological replicates e, EGFR 694 degradation in A431 cells following 24 h of Ctx -LIPTAC1 treatment . Data represented 695 three biological replicates . f, Flow cytometry analysis showing degradation of surface 696 EGFR on A431 cells following 24 h of Ctx -LIPTACs, 50 nM Ctx IgG, and 50 nM Ctx-697 KineTAC treatment . Each sample was tested in biological triplicate and error bars 698 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted June 9, 2025. ; https://doi.org/10.1101/2025.06.06.658366doi: bioRxiv preprint 20 represented standard deviations. Statistics were calculated by unpaired two -tailed 699 student t test. ***P < 0.001. ****P < 0.0001. g, Western blot analysis of EGFR and LDLR 700 in LDLR knockout and control HCC1143 cells after 24 h of 5 nM LIPTAC treatment. Data 701 represented two biological replicates. h, Fold-change in surface protein abundance in 702 MDA-MB-231 cells following 48 h of treatment with or without 50 nM Ctx -LIPTAC1, as 703 measured by quantitative proteomics analysis. N-linked cell surface glycoproteins were 704 captured by the cell -surface capture technology 39 and enriched by biocytin hydrazide. 705 Surface proteins were annotated using the SURFY database 79. i, Confocal microscopy 706 images of HeLa cells treated with 50 nM of indicated bispecific antibodies or isotype 707 controls for 24 h. Scale bar, 10 μm. 708 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted June 9, 2025. ; https://doi.org/10.1101/2025.06.06.658366doi: bioRxiv preprint 21 709 710 Fig.3. LIPTAC-mediated degradation of multiple extracellular proteins. a,b, 711 Schematic illustration of PD -L1 targeting LIPTAC, and western blot showing total 712 degradation of PD-L1 on MDA-MB-231 cells following 24 h treatment of 50 nM monomeric 713 anti-PD-L1 parent, Atz, or LIPTAC containing Atz. Percent PD-L1 levels were quantified 714 by ImageJ relative to PBS control . Each sample was tested in biological triplicate and 715 error bars represented standard deviations. Statistics calculated by unpaired two -tailed 716 student t test. *P < 0.05. **P < 0.01. c, Western blot analysis showing lysosome-dependent 717 PD-L1 degradation on MDA -MB-231 cells. Cells were pretreated with either 500 nM 718 Bafilomycin A (BafA) or 500 nM MG132 for 1 h followed by 24 h treatment with 50 nM 719 LIPTAC. d, Schematic illustration of HER2 -targeting LIPTAC and western blot analysis 720 showing total HER2 degradation on MCF7 cells following 24 h treatment of monomeric 721 Traz, LIPTAC and KineTAC. Percent ERBB2 (HER2) levels were quantified by ImageJ 722 relative to PBS control . Data represented for at least two independent experiments. e, 723 Schematic illustration of CXCR4 -targeting LIPTAC and western blot analysis showing 724 total CXCR4 degradation on HeLa cells following 24 h treatment of Nb monomer, 725 monomeric 142F1, or LIPTAC. f,g, Schematic illustration of cleaved CDCP1 -targeting 726 LIPTAC and degradation of CDCP1 in PL45 cells following 24 h treatment of 50 nM 727 cleaved CDCP1 (cCDCP1) binder CL03 IgG, pan-CDCP1 binder 4A06 IgG, or cCDCP1-728 specific LIPTAC. Each sample was tested in biological triplicate and error bars 729 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted June 9, 2025. ; https://doi.org/10.1101/2025.06.06.658366doi: bioRxiv preprint 22 represented standard deviations. Statistics were calculated by unpaired two -tailed 730 student t test. *P < 0.05. **P < 0.01. ns, not significant. 731 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted June 9, 2025. ; https://doi.org/10.1101/2025.06.06.658366doi: bioRxiv preprint 23 732 Fig.4. Development of degrader-drug conjugates (DDC) for potent target cell killing. 733 a, Schematic illustration of LIPTAC -DDCs where the membrane POI is internalized by 734 endocytosis and the cytotoxic payload is released from cleavable linkers in the lysosome. 735 b, A payload cleavage assay in EGFR-expressing A431 cells. Cells were treated with 25 736 nM of antibodies conjugated with lysolight deep red (LLDR) dyes. Images were captured 737 every 2 h for 72 h on the Incucyte. Total integrated intensity was calculated by NIRCU x 738 μm2/image. Error bars represented standard deviations for four biological replicates. 739 Statistics were calculated by one-way ANOVA and Holm-Sidak multiple comparisons test. 740 c,d, Cytotoxicity of Ctx-ADC, monomeric 142F1-ADC, and Ctx-DDC either in LIPTAC or 741 KineTAC formats on A431 and MDA -MB-231 cells, respectively. After 72 h incubation, 742 cell viability was measured using the CellTiter-Glo Reagent. EC50 values were calculated 743 using “One-Site Fit LogIC50” regression in GraphPad Prism 10.2. e, Cytotoxicity of ADCs 744 and corresponding DDCs on A431 cells in the presence and absence of the lysosomal 745 inhibitor BafA. Dead cells were labeled by cytotoxic green dye after 48 h of treatment. 746 Dashed lines represented 50 nM BafA treatment together with antibodies. Total integrated 747 intensity was calculated by GreenCU x μm2/image on the Incucyte. f, Cytotoxicity of 0.8 748 nM DDCs on A431 cells in the presence of 50 nM BafA, 50 nM MG132, and 1 µM Nystatin 749 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted June 9, 2025. ; https://doi.org/10.1101/2025.06.06.658366doi: bioRxiv preprint 24 respectively. Cells treated with inhibitors alone showed no cytotoxicity at the 750 concentrations used in DDC treatments . Dead cells were labeled by 1 µg/mL propidium 751 iodide after 48 h of treatment. Statistics were determined by one-way ANOVA. Error bars 752 represented standard deviations of four biological replicates. Statistics were calculated 753 by unpaired two-tailed student t test. *P < 0.05. ****P < 0.0001. ns, not significant. 754 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted June 9, 2025. ; https://doi.org/10.1101/2025.06.06.658366doi: bioRxiv preprint 25 755 756 757 Fig.5. Comparison of clinically approved ADCs and the DDC counterparts . a, 758 Antibody internalization assay in RPMI8226 cells, a multiple myeloma cell line, treated 759 with 100 nM of LLDR dye -labeled antibodies. An anti -SARS-CoV-2 spike antibody 760 CC12.1 was used as the isotype control 80. Images were captured every 2 h for 72 h on 761 the Incucyte. Each sample was tested in biological triplicate, and error bars represent the 762 standard deviations. b, Cytotoxicity of anti -BCMA Belantamab (Belan) ADC or the 763 corresponding LIPTAC-DDC after 4 days incubation with RPMI8226 and MM1.S cells. To 764 monitor cell death, cells were stained with APC-annexin V and propidium iodide (PI) and 765 analyzed by flow cytometry. Each sample was tested in biological triplicate and error bars 766 represent standard deviations. c, Antibody internalization assay in Raji cells, a B -cell 767 lymphoma cell line, treated with 100 nM of LLDR dye -labeled antibodies. Images were 768 captured every 2 h for 72 h on the Incucyte. Each sample was tested in biological triplicate 769 and error bars represent ed the standard deviations. d, Western blot analysis showing 770 total degradation of CD19 on Ramos and Raji cells after 24 h treatment of 50 nM 771 antibodies. Percent CD19 levels were quantified by ImageJ relative to PBS control . e, 772 Cytotoxicity of an anti -CD19 Loncastuximab (Lonca) ADC or LIPTAC -DDC after 4 days 773 incubation with Ramos cells. Cell viability was measured using the CellTiter-Glo Reagent. 774 Each sample was tested in biological triplicate and error bars represent ed standard 775 deviations. EC50 values were calculated using “One -Site Fit LogIC50” regression in 776 GraphPad Prism 10.2. 777 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted June 9, 2025. ; https://doi.org/10.1101/2025.06.06.658366doi: bioRxiv preprint 26

778 1. Uhlén, M. et al. Tissue-based map of the human proteome. Science 347, 1260419 779 (2015). 780 2. Mayes, P. A., Hance, K. W. & Hoos, A. The promise and challenges of immune 781 agonist antibody development in cancer. Nat. Rev. Drug Discov. 17, 509 –527 782 (2018). 783 3. Goebeler, M.-E. & Bargou, R. C. T cell -engaging therapies - BiTEs and beyond. 784 Nat. Rev. Clin. Oncol. 17, 418–434 (2020). 785 4. Dumontet, C., Reichert, J. M., Senter, P. D., Lambert, J. M. & Beck, A. Antibody -786 drug conjugates come of age in oncology. Nat. Rev. Drug Discov. 22, 641–661 787 (2023). 788 5. Wells, J. A. & Kumru, K. Extracellular targeted protein degradation: an emerging 789 modality for drug discovery. Nat. Rev. Drug Discov. 23, 126–140 (2024). 790 6. Fu, Z., Li, S., Han, S., Shi, C. & Zhang, Y. Antibody drug conjugate: the “biological 791 missile” for targeted cancer therapy. Signal Transduct. Target. Ther. 7, 93 (2022). 792 7. Austin, C. D. et al. Endocytosis and sorting of ErbB2 and the site of action of 793 cancer therapeutics trastuzumab and geldanamycin. Mol. Biol. Cell 15, 5268–5282 794 (2004). 795 8. Pegram, M. D. et al. First-in-Human, Phase 1 Dose -Escalation Study of 796 Biparatopic Anti -HER2 Antibody -Drug Conjugate MEDI4276 in Patients with 797 HER2-positive Advanced Breast or Gastric Cancer. Mol. Cancer Ther. 20, 1442–798 1453 (2021). 799 9. Li, J. Y. et al. A Biparatopic HER2 -Targeting Antibody-Drug Conjugate Induces 800 Tumor Regression in Primary Models Refractory to or Ineligible for HER2-Targeted 801 Therapy. Cancer Cell 29, 117–129 (2016). 802 10. de Goeij, B. E. C. G. et al. Efficient Payload Delivery by a Bispecific Antibody-Drug 803 Conjugate Targeting HER2 and CD63. Mol. Cancer Ther. 15, 2688–2697 (2016). 804 11. DeVay, R. M. et al. Improved lysosomal trafficking can modulate the potency of 805 antibody drug conjugates. Bioconjug. Chem. 28, 1102–1114 (2017). 806 12. Békés, M., Langley, D. R. & Crews, C. M. PROTAC targeted protein degraders: 807 the past is prologue. Nat. Rev. Drug Discov. 21, 181–200 (2022). 808 13. Igawa, T., Haraya, K. & Hattori, K. Sweeping antibody as a novel therapeutic 809 antibody modality capable of eliminating soluble antigens from circulation. Immunol. 810 Rev. 270, 132–151 (2016). 811 14. Ahn, G. et al. LYTACs that engage the asialoglycoprotein receptor for targeted 812 protein degradation. Nat. Chem. Biol. 17, 937–946 (2021). 813 15. Banik, S. M. et al. Lysosome-targeting chimaeras for degradation of extracellular 814 proteins. Nature 584, 291–297 (2020). 815 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted June 9, 2025. ; https://doi.org/10.1101/2025.06.06.658366doi: bioRxiv preprint 27 16. Cotton, A. D., Nguyen, D. P., Gramespacher, J. A., Seiple, I. B. & Wells, J. A. 816 Development of Antibody-Based PROTACs for the Degradation of the Cell-Surface 817 Immune Checkpoint Protein PD-L1. J. Am. Chem. Soc. 143, 593–598 (2021). 818 17. Gramespacher, J. A., Cotton, A. D., Burroughs, P. W. W., Seiple, I. B. & Wells, J. 819 A. Roadmap for Optimizing and Broadening Antibody -Based PROTACs for 820 Degradation of Cell Surface Proteins. ACS Chem. Biol. 17, 1259–1268 (2022). 821 18. Marei, H. et al. Antibody targeting of E3 ubiquitin ligases for receptor degradation. 822 Nature 610, 182–189 (2022). 823 19. Pance, K. et al. Modular cytokine receptor -targeting chimeras for targeted 824 degradation of cell surface and extracellular proteins. Nat. Biotechnol. 41, 273–281 825 (2023). 826 20. Zheng, J. et al. Bifunctional Compounds as Molecular Degraders for Integrin -827 Facilitated Targeted Protein Degradation. J. Am. Chem. Soc. 144, 21831–21836 828 (2022). 829 21. Zhang, D. et al. Transferrin receptor targeting chimeras for membrane protein 830 degradation. Nature 638, 787–795 (2025). 831 22. Rennick, J. J., Johnston, A. P. R. & Parton, R. G. Key principles and methods for 832 studying the endocytosis of biological and nanoparticle therapeutics. Nat. 833 Nanotechnol. 16, 266–276 (2021). 834 23. Mettlen, M., Chen, P.-H., Srinivasan, S., Danuser, G. & Schmid, S. L. Regulation 835 of Clathrin-Mediated Endocytosis. Annu. Rev. Biochem. 87, 871–896 (2018). 836 24. Rudenko, G. et al. Structure of the LDL receptor extracellular domain at 837 endosomal pH. Science 298, 2353–2358 (2002). 838 25. Scully, T., Kase, N., Gallagher, E. J. & LeRoith, D. Regulation of low -density 839 lipoprotein receptor expression in triple negative breast cancer by EGFR -MAPK 840 signaling. Sci. Rep. 11, 17927 (2021). 841 26. Yuan, Q. et al. Low-density lipoprotein receptor promotes crosstalk between cell 842 stemness and tumor immune microenvironment in breast cancer: a large data -843 based multi-omics study. J. Transl. Med. 21, 871 (2023). 844 27. Bonacina, F. et al. The low-density lipoprotein receptor-mTORC1 axis coordinates 845 CD8+ T cell activation. J. Cell Biol. 221, (2022). 846 28. Brown, M. S. & Goldstein, J. L. A receptor -mediated pathway for cholesterol 847 homeostasis. Science 232, 34–47 (1986). 848 29. Goldstein, J. L. & Brown, M. S. The LDL receptor. Arterioscler. Thromb. Vasc. Biol. 849 29, 431–438 (2009). 850 30. Schaefer, K. et al. Direct Identification of Proteolytic Cleavages on Living Cells 851 Using a Glycan-Tethered Peptide Ligase. ACS Cent. Sci. 8, 1447–1456 (2022). 852 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted June 9, 2025. ; https://doi.org/10.1101/2025.06.06.658366doi: bioRxiv preprint 28 31. Banerjee, S. et al. Proteolysis of the low density lipoprotein receptor by bone 853 morphogenetic protein-1 regulates cellular cholesterol uptake. Sci. Rep. 9, 11416 854 (2019). 855 32. Herling, T. W. et al. Nonspecificity fingerprints for clinical -stage antibodies in 856 solution. Proc Natl Acad Sci USA 120, e2306700120 (2023). 857 33. Goldstein, J. L., Basu, S. K., Brunschede, G. Y. & Brown, M. S. Release of low 858 density lipoprotein from its cell surface receptor by sulfated glycosaminoglycans. 859 Cell 7, 85–95 (1976). 860 34. Spano, J. P. et al. Impact of EGFR expression on colorectal cancer patient 861 prognosis and survival. Ann. Oncol. 16, 102–108 (2005). 862 35. Arienti, C., Pignatta, S. & Tesei, A. Epidermal growth factor receptor family and its 863 role in gastric cancer. Front. Oncol. 9, 1308 (2019). 864 36. Von Kreudenstein, T. S. et al. Improving biophysical properties of a bispecific 865 antibody scaffold to aid developability: quality by molecular design. MAbs 5, 646–866 654 (2013). 867 37. Schlothauer, T. et al. Novel human IgG1 and IgG4 Fc-engineered antibodies with 868 completely abolished immune effector functions. Protein Eng. Des. Sel. 29, 457–869 466 (2016). 870 38. Li, Z. et al. Ribosome stalling during selenoprotein translation exposes a 871 ferroptosis vulnerability. Nat. Chem. Biol. 18, 751–761 (2022). 872 39. Wollscheid, B. et al. Mass-spectrometric identification and relative quantification 873 of N-linked cell surface glycoproteins. Nat. Biotechnol. 27, 378–386 (2009). 874 40. Berditchevski, F. & Odintsova, E. ErbB receptors and tetraspanins: Casting the 875 net wider. Int. J. Biochem. Cell Biol. 77, 68–71 (2016). 876 41. Sugiyama, M. G. et al. Confinement of unliganded EGFR by tetraspanin 877 nanodomains gates EGFR ligand binding and signaling. Nat. Commun. 14, 2681 878 (2023). 879 42. Kwon, H. J., Lagace, T. A., McNutt, M. C., Horton, J. D. & Deisenhofer, J. 880 Molecular basis for LDL receptor recognition by PCSK9. Proc Natl Acad Sci USA 881 105, 1820–1825 (2008). 882 43. Maxwell, K. N., Fisher, E. A. & Breslow, J. L. Overexpression of PCSK9 883 accelerates the degradation of the LDLR in a post -endoplasmic reticulum 884 compartment. Proc Natl Acad Sci USA 102, 2069–2074 (2005). 885 44. Freeman, G. J. et al. Engagement of the PD -1 immunoinhibitory receptor by a 886 novel B7 family member leads to negative regulation of lymphocyte activation. J. 887 Exp. Med. 192, 1027–1034 (2000). 888 45. Zhang, F. et al. Structural basis of the therapeutic anti -PD-L1 antibody 889 atezolizumab. Oncotarget 8, 90215–90224 (2017). 890 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted June 9, 2025. ; https://doi.org/10.1101/2025.06.06.658366doi: bioRxiv preprint 29 46. Mauvezin, C. & Neufeld, T. P. Bafilomycin A1 disrupts autophagic flux by inhibiting 891 both V -ATPase-dependent acidification and Ca -P60A/SERCA-dependent 892 autophagosome-lysosome fusion. Autophagy 11, 1437–1438 (2015). 893 47. Kisselev, A. F. & Goldberg, A. L. Proteasome inhibitors: from research tools to 894 drug candidates. Chem. Biol. 8, 739–758 (2001). 895 48. Oh, D.-Y. & Bang, Y.-J. HER2-targeted therapies - a role beyond breast cancer. 896 Nat. Rev. Clin. Oncol. 17, 33–48 (2020). 897 49. Chatterjee, S., Behnam Azad, B. & Nimmagadda, S. The intricate role of CXCR4 898 in cancer. Adv. Cancer Res. 124, 31–82 (2014). 899 50. Claes, K. et al. Modular Integrated Secretory System Engineering in Pichia 900 pastoris To Enhance G-Protein Coupled Receptor Expression. ACS Synth. Biol. 5, 901 1070–1075 (2016). 902 51. Scherl-Mostageer, M. et al. Identification of a novel gene, CDCP1, overexpressed 903 in human colorectal cancer. Oncogene 20, 4402–4408 (2001). 904 52. Martinko, A. J. et al. Targeting RAS-driven human cancer cells with antibodies to 905 upregulated and essential cell-surface proteins. eLife 7, e31098 (2018). 906 53. Moroz, A. et al. Theranostic targeting of CUB domain containing protein 1 (CDCP1) 907 in pancreatic cancer. Clin. Cancer Res. 26, 3608–3615 (2020). 908 54. Lim, S. A. et al. Targeting a proteolytic neoepitope on CUB domain containing 909 protein 1 (CDCP1) for RAS-driven cancers. J. Clin. Invest. 132, e154604 (2022). 910 55. Holly, R. W. et al. Monitoring lysosomal catabolism: a sensitive probe for 911 assessing targeted lysosomal degradation of extracellular proteins. BioRxiv (2024) 912 doi:10.1101/2024.10.18.619124. 913 56. Doronina, S. O. et al. Development of potent monoclonal antibody auristatin 914 conjugates for cancer therapy. Nat. Biotechnol. 21, 778–784 (2003). 915 57. Rao, M., Murali, S., Amores, D., Yu, F. & Tsourkas, A. Exploring the Sensitivity of 916 Antibody-Drug Conjugate Efficacy to the Selection of Payload, Antibody, and Cell 917 line. Bioconjug. Chem. 35, 115–124 (2024). 918 58. Ba, Q. et al. Dihydroartemisinin exerts its anticancer activity through depleting 919 cellular iron via transferrin receptor-1. PLoS ONE 7, e42703 (2012). 920 59. Hungria, V. & Mateos, M. -V. Belantamab mafodotin, bortezomib, and 921 dexamethasone for multiple myeloma. reply. N. Engl. J. Med. 391, 1364 –1365 922 (2024). 923 60. Baines, A. C. et al. FDA Approval Summary: Belantamab Mafodotin for Patients 924 with Relapsed or Refractory Multiple Myeloma. Clin. Cancer Res. 28, 4629–4633 925 (2022). 926 61. Jain, N. et al. Loncastuximab tesirine, an anti -CD19 antibody-drug conjugate, in 927 relapsed/refractory B-cell acute lymphoblastic leukemia. Blood Adv. 4, 449–457 928 (2020). 929 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted June 9, 2025. ; https://doi.org/10.1101/2025.06.06.658366doi: bioRxiv preprint 30 62. Bardia, A. et al. Sacituzumab Govitecan in Metastatic Triple -Negative Breast 930 Cancer. N. Engl. J. Med. 384, 1529–1541 (2021). 931 63. Cortés, J. et al. Trastuzumab Deruxtecan versus Trastuzumab Emtansine for 932 Breast Cancer. N. Engl. J. Med. 386, 1143–1154 (2022). 933 64. Lonial, S. et al. Belantamab mafodotin for relapsed or refractory multiple myeloma 934 (DREAMM-2): a two-arm, randomised, open-label, phase 2 study. Lancet Oncol. 935 21, 207–221 (2020). 936 65. Gerber, H.-P. et al. Potent antitumor activity of the anti -CD19 auristatin antibody 937 drug conjugate hBU12 -vcMMAE against rituximab -sensitive and -resistant 938 lymphomas. Blood 113, 4352–4361 (2009). 939 66. Hommelgaard, A. M., Lerdrup, M. & van Deurs, B. Association with membrane 940 protrusions makes ErbB2 an internalization -resistant receptor. Mol. Biol. Cell 15, 941 1557–1567 (2004). 942 67. Cuello, M. et al. Down-regulation of the erbB-2 receptor by trastuzumab (herceptin) 943 enhances tumor necrosis factor -related apoptosis -inducing ligand -mediated 944 apoptosis in breast and ovarian cancer cell lines that overexpress erbB-2. Cancer 945 Res. 61, 4892–4900 (2001). 946 68. Strop, P. et al. RN927C, a Site-Specific Trop-2 Antibody-Drug Conjugate (ADC) 947 with Enhanced Stability, Is Highly Efficacious in Preclinical Solid Tumor Models. 948 Mol. Cancer Ther. 15, 2698–2708 (2016). 949 69. Brown, M. S., Anderson, R. G. & Goldstein, J. L. Recycling receptors: the round -950 trip itinerary of migrant membrane proteins. Cell 32, 663–667 (1983). 951 70. Xu, S. Internalization, Trafficking, Intracellular Processing and Actions of 952 Antibody-Drug Conjugates. Pharm. Res. 32, 3577–3583 (2015). 953 71. Andreev, J. et al. Bispecific Antibodies and Antibody -Drug Conjugates (ADCs) 954 Bridging HER2 and Prolactin Receptor Improve Efficacy of HER2 ADCs. Mol. 955 Cancer Ther. 16, 681–693 (2017). 956 72. Li, S. et al. Structural basis for inhibition of the epidermal growth factor receptor 957 by cetuximab. Cancer Cell 7, 301–311 (2005). 958 73. Hattori, T. & Koide, S. Next -generation antibodies for post -translational 959 modifications. Curr. Opin. Struct. Biol. 51, 141–148 (2018). 960 74. Walsh, G. & Jefferis, R. Post -translational modifications in the context of 961 therapeutic proteins. Nat. Biotechnol. 24, 1241–1252 (2006). 962 75. Xie, N. et al. Neoantigens: promising targets for cancer therapy. Signal Transduct. 963 Target. Ther. 8, 9 (2023). 964 76. Mayor, S., Sabharanjak, S. & Maxfield, F. R. Cholesterol -dependent retention of 965 GPI-anchored proteins in endosomes. EMBO J. 17, 4626–4638 (1998). 966 77. Hornsby, M. et al. A High Through -put Platform for Recombinant Antibodies to 967 Folded Proteins. Mol. Cell. Proteomics 14, 2833–2847 (2015). 968 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted June 9, 2025. ; https://doi.org/10.1101/2025.06.06.658366doi: bioRxiv preprint 31 78. Fisher, T. S. et al. Effects of pH and low density lipoprotein (LDL) on PCSK9 -969 dependent LDL receptor regulation. J. Biol. Chem. 282, 20502–20512 (2007). 970 79. Bausch-Fluck, D. et al. The in silico human surfaceome. Proc Natl Acad Sci USA 971 115, E10988–E10997 (2018). 972 80. Rogers, T. F. et al. Isolation of potent SARS -CoV-2 neutralizing antibodies and 973 protection from disease in a small animal model. Science 369, 956–963 (2020). 974 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted June 9, 2025. ; https://doi.org/10.1101/2025.06.06.658366doi: bioRxiv preprint