Type 2 but Not Type 1 Calreticulin Mutants Exhibit Loss of Chaperone Function Due to C-Terminal Structural Interference, and Bind to Wild Type Calreticulin in a Dominant Negative Manner

Calreticulin (CALR) mutations are the second most common genetic driver of myeloproliferative neoplasms (MPNs). Most CALR mutations are classified as either type 1 (52 bp deletion; del52) or type 2 (5 bp insertion; ins5). Both lead to the generation of an identical mutant C-terminus, which is required to transform cells through aberrant binding to and activation of the thrombopoietin receptor MPL. Despite their identical C-termini and shared mechanism of transformation, patients with type 1 and type 2 CALR mutations display significant clinical and prognostic differences. The mechanisms underlying these divergent clinical phenotypes remain incompletely understood.

Though classified as gain-of-function mutations, we previously found that both CALRdel52 and CALRins5 mutants display differential losses-of-function; CALRdel52 loses its ability to bind calcium (ASH abstract #147859, 2021), while CALRins5 loses its function as a molecular chaperone. We demonstrated that loss of CALRins5 chaperone function leads to activation of the ATF6 arm of the unfolded protein response (ASH abstract #3588, 2021), and that this signaling culminates in ATF6-mediated up-regulation of BCL-xL, which renders CALRins5 cells dependent upon this pathway for survival However, these studies did not address the molecular mechanism by which CALRins5 loses chaperone function.

To dissect this, we performed structural modeling studies and found that the mutant C-terminal tail of the CALRins5 but not the CALRdel52 protein exhibits structural proximity to the globular N domain of the protein, which is critical for CALR chaperone function. We thus hypothesized that the observed loss of chaperone function may be due to structural interference by the CALRins5 C-terminus. To test this, we created a series of CALRins5 truncation variants in which we progressively trimmed back the mutant C-terminal tail. We found that such trimming restored chaperone function to the CALRins5 protein, suggesting that indeed the loss of chaperone function is due to interference from the mutant C-terminal tail.

Because CALR mutations are heterozygous in patients, we next asked why the wild type copy of CALR does not compensate for the loss of chaperone function in CALRins5-expressing cells. Though it has been shown that CALRdel52 does not bind to CALRwt, this has never been tested for CALRins5. We thus hypothesized that CALRins5 may bind to CALRwt in a dominant negative manner, leading to complete loss of CALR-mediated chaperone function in CALRins5-expressing cells. To test this, we performed co-immunoprecipitation assays and found that CALRins5 does indeed bind to CALRwt, while, consistent with previous findings, CALRdel52 and CALRwt do not interact. To determine whether this binding was in fact dominant negative, we performed an in vitro chaperone assay and found that CALRwt chaperone function is suppressed when in the presence of CALRins5. This suggests that CALRins5 exhibits a dominant negative effect over CALRwt chaperone function.

Finally, we sought to determine if this loss in chaperone ability was important for CALRins5 cell survival. To test this, we utilized the chemical chaperone tauroursodeoxycholic acid (TUDCA) to restore ER chaperone capacity. We found that CALRins5 but not CALRdel52 or CALRwt cell lines and patient samples displayed decreased viability upon TUDCA treatment, suggesting chaperone loss is necessary for survival of CALRins5 cells. Importantly, TUDCA has shown immense clinical benefit for diseases characterized by molecular chaperone dysfunction (e.g., neurodegenerative diseases). To test whether this benefit translates to MPNs, we utilized our mutant CALR-driven bone marrow transplantation model. We found that CALRins5 mice treated with TUDCA displayed significantly decreased platelet counts compared to those treated with vehicle alone, while TUDCA had no effect on CALRdel52-driven disease.

Taken together, we have dissected the mechanism by which CALRins5 proteins lose chaperone ability, as well as demonstrate that CALRins5 proteins exhibit a dominant negative effect over CALRwt chaperone function. Finally, we reveal TUDCA as a potential therapeutic strategy for CALRins5 patients that has immediate clinical tractability.