Background: Even though PD-1/PD-L1 is an identified key "don't find me" signal to active adaptive immune system for cancer treatment, the overall response rate (ORR) for all cancer patients is still limited. Other effective therapeutic modalities to bridge the innate and adaptive immunity to improve ORR are urgently needed. Recently, CD47/SIRPα interaction is confirmed as a critical "don't eat me" signal to active innate immunity. However, the red blood cell (RBC) toxicity is the big concern for the development of CD47-based anti-cancer therapeutics. Methods: Here, we report the development of a CD47/PD-L1 bispecific antibody 6MW3211 to block both PD-1/PD-L1 and CD47/SIRPα signals, and studied the effects of 6MW3211 on anti-tumor immune functions in vitro and in vivo. The pharmacokinetic and toxicity profiles of 6MW3211 were evaluated in GLP non-human primate (NHP) studies. Results:The dual immune checkpoint inhibitory signaling blocker 6MW3211 shows high binding affinity to PD-L1 and low binding affinity to CD47. This inequivalent binding affinity design makes 6MW3211 preferentially bound to PD-L1 on tumor cells followed by disrupting the interaction of CD47/SIRPα. Complex structure determination and flow cytometry assay demonstrated that 6MW3211 has no binding to either human or rhesus monkey RBCs. 6MW3211 effectively blocked both PD-1/DP-L1 and CD47/SIRPα signaling and promoted macrophage phagocytosis of tumor cells. Potent therapeutic efficacies of 6MW3211 in three different mouse models were further observed. Moreover, 6MW3211 was demonstrated to have a fairly good safety profile in a GLP NHP study. In addition, multiplex fluorescent immunohistochemistry (mIHC) staining shows that PD-L1 and CD47 co-express on several different types of human tumor tissues. Conclusions: These results support the development of 6MW3211 for the treatment of PD-L1 and CD47 double positive cancers.

Alkaline pectate lyases have biotechnological applications in plant fiber processing, such as ramie degumming. Previously, we characterized an alkaline pectate lyase from Bacillus clausii S10, named BacPelA, which showed potential for enzymatic ramie degumming because of its high cleavage activity toward methylated pectins in alkaline conditions. However, BacPelA displayed poor thermo-alkaline stability. Here, we report the 1.78 Å resolution crystal structure of BacPelA in apo form. The enzyme has the characteristic right-handed β-helix fold of members of the polysaccharide lyase 1 family and shows overall structural similarity to them, but it displays some differences in the details of the secondary structure and Ca2+-binding site. On the basis of the structure, 10 sites located in flexible regions and showing high B-factor and positive ΔTm values were selected for mutation, aiming to improve the thermo-alkaline stability of the enzyme. Following site-directed saturation mutagenesis and screening, mutants A238C, R150G, and R216H showed an increase in the T5015 value at pH 10.0 of 3.0 °C, 6.5 °C, and 7.0 °C, respectively, compared with the wild-type enzyme, interestingly accompanied by a 24.5%, 46.6%, and 61.9% increase in activity. The combined mutant R150G/R216H/A238C showed an 8.5 °C increase in the T5015 value at pH 10.0, and an 86.1% increase in the specific activity at 60 °C, with approximately doubled catalytic efficiency, compared with the wild-type enzyme. Moreover, this mutant retained 86.2% activity after incubation in ramie degumming conditions (4 h, 60 °C, pH 10.0), compared with only 3.4% for wild-type BacPelA. The combined mutant increased the weight loss of ramie fibers in degumming by 30.2% compared with wild-type BacPelA. This work provides a thermo-alkaline stable, highly active pectate lyase with great potential for application in the textile industry, and also illustrates an effective strategy for rational design and improvement of pectate lyases.