Big part of $TRIL future valuation will be solid tumor data. $ALXO ALX148 in a ph1 trial demonstrated 40% ORR in combo w/ PDL1i for checkpoint inhibitor naive HNSCC tumors. Could $TRIL beat this? ALX148 was engineered to bind CD47 w/ high affinity. 148 avoids toxicity by (1/n)

using an inactive IgG. Thus the strength of 148 is its ability to be dosed high and generate good tumor coverage. However, its strong CD47 binding domain means it can still engage RBCs and create an antigen sink, even if it won't stimulate off-target phagocytosis. This (2/n)

potentially pulls a lot of available macrophages away from the tumor. This is likely very detrimental to the efficacy, as weak IgGs cannot convert tumor associated macrophages, only block new macrophages from being turned off by the tumor. However, if most macrophages are (3/n)

being pulled away to other CD47 expressing cells they will not be available to infiltrate tumor. $ALXO knows this, and on their website markets 148 as a drug that will combo well as it blocks CD47, but needs an additional signal for efficacy. They seem to pump that this is (4/n)

the best way to design a CD47 inhibitor, as any active effector domain will generate dose-limiting toxicities. Here they make the assumption that all other CD47 assets bind off-target CD47. If there is an active IgG that selectively binds cancer CD47, then through (5/n)

their own logic, it would be superior. $TRIL 621/622 contain a CD47 binding domain that selectively binds mobile CD47 (see attached). Interestingly, human RBC CD47 localizes in insoluble structures that likely lack mobility. Therefore, 621/622 avoid off-target (6/n)

RBC CD47 binding through an odd human-specific biophysical property of CD47 w/in RBCs. Because of this, 622 cannot only block cancer CD47 w/ good coverage but also avoid an antigen sink. This leads to higher efficacy in blood cancer, in which 622 had 50% ORR at 8m/k and (7/n)

ALX148 showed no single agent activity. So what does this mean for combo solid tumor trials? Probably a lot. ALX148 achieved 40% ORR w/ PDL1i. PDL1 axis also inhibits macrophages, and thus the combined I/O likely activates phagocytosis in some patients. However, these (8/n)

responders likely need both high CD47 and PDL1 expression along with other factors. Interestingly ALXO did not find correlation between CD47 expression and patient response, suggesting responding patients need multiple prerequisites, limiting the target patient pop. (9/n)

$TRIL 622 in theory would not need additional stimuli to generate phagocytosis, and thus could turn many more cold tumors hot regardless of other checkpoints. This could be critical in a PDLi combo trial. Finally, 622 can activate M1s and prevent macrophage conversion to (10/n)

M2 TAMs. IgG1 based mAbs like rituximab provide an IgG1, which can convert M2 TAMs. Thus in 622+Rtx combo, you could expect to prevent shut-off of new infiltrating M1 macrophages and convert M2 TAMs, thus generating phagocytic activity across macrophage species. ALX148 (11/n)

sequesters circulating macrophages at RBCs, and thus cannot promote phagocytosis from this population of cells. Therefore, I believe we will see superior data for $TRIL assets in combo with both IgG1 based mAbs and other checkpoint inhibitors. Just my model, we'll see.