Targeting human CALR‐mutated MPN progenitors with a neoepitope‐directed monoclonal antibody

Both type 1 and type 2 CALR mutations in myelofibrosis result in a loss of the endoplasmic reticulum retention signal KDEL producing aberrant localization of mutant protein (Klampfl et al, 2013; Nangalia et al, 2013; How et al, 2019) and the acquisition of a 31 amino-acid-long sequence towards the C-terminus (Fig 1A). We reasoned that this neoepitope would be an ideal target for precision medicine approaches and set out to develop a panel of rat monoclonal antibodies to a chemically synthesized peptide corresponding to the C-terminal mutant CALR neoepitope sequence (Fig 1A). Of three hybridoma screens, several antibodies showed positive binding to the peptide and to the CALR-mutant protein in immunoblotting with clone 4D7 showing the strongest reactivity (Fig EV1A). By using peptides covering different regions of the immunizing neopeptide, we mapped the binding of 4D7 to the 11 amino acids of the N-terminus with no binding of 4D7 observed to a scrambled peptide or the C-terminal portion of the immunogen (Fig 1B). Clone 4D7 mAb showed a dissociation constant (Kd) of 1.53 nM to the full-length peptide as measured by 125I-Scatchard analysis (Figs 1C and EV1B). To investigate the specific binding of 4D7 on CALR mutant-bearing cells, we generated cytokine-independent TF-1 TpoR CALRdel61 and TF-1 TpoR CALRdel52 (Fig EV1C). Using these cells in a horizontal co-culturing system, we found no evidence of a paracrine effect by the mutant CALR protein (Fig EV1D). Specific binding of phycoerythrin (PE)-conjugated 4D7 to the cell surface of mutant CALR-expressing cells, but not TF-1 TpoR CALRWT-expressing cells, was observed in comparison to unstained or isotype-PE control (Figs 1D and EV1E).

Monoclonal antibody 4D7 blocks TPO-independent signalling

We then tested cytokine-independent TF-1 TpoR CALRdel61 and TF-1 TpoR CALRdel52 cells for their ability to proliferate after treatment with increasing concentrations of 4D7 antibody or isotype control. Firstly, we noted no inhibition by 4D7 of TF-1 cells expressing TpoR alone (with endogenous wild-type CALR; Fig 1E). In contrast, we observed blockade of TPO-independent cell growth over 5 days using 2, 10 and 20 µg/ml 4D7 in cytokine-independent cells (Fig 1F and G) with evidence of concentration-dependent inhibition observed after 48 h. No inhibition was observed with IgG isotype control antibody at 20 µg/ml. No inhibition by 4D7 was observed in TF-1 cells with an overexpression of wild-type CALR (Fig EV2A). Additionally, TF-1 cells harbouring either the CALRdel52 or CALRdel61 mutations but lacking TpoR were unaffected when treated with 4D7 (Fig EV2B and C). Interestingly, MARIMO cells also did not show inhibition with 4D7, consistent with CALR no longer acting as a driver mutation in this NRAS mutant cell line (Han et al, 2018) (Fig EV2D). In addition, no inhibition was observed in factor-independent JAK2V617F SET2 cells (Fig EV2E) or factor-independent TF-1 expressing the PTPN11E76K mutation (Fig EV2F) indicating a mutation-specific effect.

Details are in the caption following the image

Figure EV2. Biological specificity of 4D7

Cytokine-dependent TF-1 TpoR cells with an overexpression of WT CALR cultured in the absence of TPO, 10 ng/ml hTPO and 10 or 20 µg/ml 4D7 or 20 µg/ml control IgG antibody for 5 days and the number of trypan blue-negative cells were counted every 24 h (n = 3 biological replicates with three technical replicates). Cytokine-dependent TF-1 CALRdel52 cells lacking TpoR were cultured in the presence of 2 ng/ml GM-CSF and 10 or 20 µg/ml 4D7 or 20 µg/ml control IgG antibody (n = 3 biological replicates with three technical replicates). Cytokine-dependent TF-1 CALRdel61 cells lacking TpoR were cultured in the presence of 2 ng/ml GM-CSF and 10 or 20 µg/ml 4D7 or 20 µg/ml control IgG antibody (n = 3 biological replicates with three technical replicates). MARIMO cells from which the CALRdel61 mutation was originally amplified were cultured in the presence of 2, 10 or 20 µg/ml 4D7 or 20 µg/ml control IgG antibody (n = 3 biological replicates with three technical replicates). Cytokine-independent SET2 cells which harbour the pathogenic JAK2V617F mutation were cultured in the presence of 2, 10 or 20 µg/ml 4D7 or 20 µg/ml control IgG antibody (n = 3 biological replicates with three technical replicates). Cytokine-independent TF-1 PTPN11E76K cells were cultured in the presence of 10 or 20 µg/ml 4D7 or 20 µg/ml control IgG antibody (n = 3 biological replicates with three technical replicates).

Data information: For all panels, bars represent standard error of the mean.

To understand the mechanism of action of 4D7, we performed signalling experiments in TF-1 cells. Firstly, we observed no inhibition of signalling in cytokine-dependent TF-1 cells expressing TpoR alone, which showed the expected down-regulation of phospho-STAT1,3,5 and phospho-ERK only after TPO withdrawal (Figs 2A and EV3A). In contrast, we observed blockade of constitutive factor-independent phospho-STAT1/3/5 and phospho-ERK after incubation with 10 or 20 µg/ml 4D7, but not with IgG control in both CALRdel61 and CALRdel52 cells (Figs 2B and C, EV3B and C). Importantly, this blockade was also observed in peripheral blood mononuclear cells (PBMNCs) from a primary CALRdel52 PMF (Fig 2D) but not in JAK2V617F PMF (Fig EV3D). Treatment with 4D7 resulted in a significant reduction in phospho-signalling, whereas ruxolitinib resulted in complete suppression. This was expected as ruxolitinib is known to inhibit all JAK-dependent signalling, lacking mutation specificity (Tvorogov et al, 2018). A significant increase in the sub-G0 fraction was observed by 4D7 compared to IgG control (P = 0.001; Fig 2E) in the absence of a major effect on G2-M on cell cycle analysis (Fig EV3E) consistent with induction of an apoptotic response. This was confirmed by caspase 3 cleavage that occurred within 24 h and peaked at 48 h after 4D7 treatment (Fig 2F).

Details are in the caption following the image

Figure 2. 4D7 disrupts interaction with TpoR and downstream signalling

Cell extracts blotted for phospho-STAT5, total STAT5, phospho-ERK, total ERK and actin from TF-1 TpoR cells after incubation with 10 or 20 µg/ml 4D7 or IgG for 4 or 8 h, in presence of TPO. The last line indicates TPO withdrawal. Similar experiment using TPO-independent TF-1 TpoR CALRdel61 cells. Similar experiment using TPO-independent TF-1 TpoR CALRdel52 cells at 8 h. Similar experiment using PBMNCs from CALRdel52 PMF primary cells at 8 h. Additionally, cells were treated with 280 nM of ruxolitinib as a positive control. Fraction of apoptotic sub-G0 population of TF-1 TpoR CALRdel61 cells after 48 h of 4D7 or IgG treatment (n = 3 biological replicates, bars represent standard deviation for three replicates, normalized to IgG, with a Student's unpaired t-test used to determine statistical significance). Western blot showing caspase 3 cleavage occurring within 24 h of TPO withdrawal in TPO-dependent TF-1 TpoR cells. An increase in cleaved caspase 3 is observed after 48 h of treatment with 20 µg/ml 4D7 in TF-1 TpoR CALRdel61 and TF-1 TpoR CALRdel52 cells. Western blot of TpoR immunoprecipitation under non-reducing conditions showing associated CALR 50 kDa monomers and 100 kDa dimers (red arrowheads) present only in TF-1 TpoR CALRdel61 disrupted by 8-h treatment with 20 µg/ml 4D7 but not PBS or 20 µg/ml IgG. CALR monomers and dimers are detectable by polyclonal anti-wild-type CALR or anti-mutant CALR monoclonal antibodies. Red arrowheads, detected mutant CALR protein; brown arrowheads, detected wild-type CALR protein; asterisk, non-specific bands.

Source data are available online for this figure.

Details are in the caption following the image

Figure EV3. Monoclonal antibody blocks STAT1, 3 signalling and TpoR phosphorylation

Cell extracts blotted for phospho-STAT1, total STAT1, phospho-STAT3, total STAT3 and actin from TF-1 TpoR cells after incubation with 10 or 20 µg/ml 4D7 or IgG for 4 or 8 h as indicated. Similar experiment using TPO-independent TF-1 TpoR CALRdel61 cells at 4 and 8 h. Similar experiment shown using TPO-independent TF-1 TpoR CALRdel52 cells, 8 h. Similar experiment using PBMNCs from JAK2V617F PMF primary cells at 8 h. Additionally, cells were treated with 280 nM of ruxolitinib as a positive control. Flow cytometry analysis for cell cycle distribution of TF-1 TpoR CALRdel61 cells exposed to 20 µg/ml IgG or 4D7 for 48 h. Cells were harvested and fixed and stained with propidium iodide and their DNA contents were analysed. Results from one representative experiment shown. Percentages of cells in Sub G0, G1, S and G2/M cycle indicated. Western blot showing specific co-immunoprecipitation of mutant CALRdel61 cells with TpoR anti-FLAG antibody under reducing conditions detected by WT and mutant-specific CALR antibodies. Red arrowheads, detected mutant CALR protein; brown arrowheads, detected wild-type CALR protein. Western blot showing decreased TpoR phosphorylation in a TpoR immunoprecipitated after 8 h of 20 µg/ml 4D7 treatment compared to PBS or IgG control in TF-1 TpoR CALRdel61.

Source data are available online for this figure.

4D7 disrupts the mutCALR/TpoR complex

To investigate the mechanism of action of 4D7, we examined how it affects mutant CALR interactions with TpoR. Analysis of TF-1 TpoR CALRdel61 cells revealed that CALR mutants constitutively existed in complex with the TpoR (Fig EV3F) consistent with previous findings (Araki et al, 2019). Importantly, pre-treatment with 4D7 diminished the presence of CALR monomers (slightly < 50 kDa) and dimers (~ 100 kDa) bound to TpoR (Fig 2G), supporting the notion that 4D7 interferes with the mutCALR/TpoR complexes at the cell surface by disrupting the constitutive association of mutCALR to TpoR. Similar results were obtained with TF-1 TpoR CALRdel52 cells (Fig EV4A). 4D7 also prevented TpoR activation in the presence of CALR-mutant protein as demonstrated by its inhibition of TpoR tyrosine phosphorylation (Fig EV3G) and downstream signalling (Fig 2B–D), consistent with inhibition of receptor activation.

Details are in the caption following the image

Figure EV4. 4D7 blocks CALR interaction with TpoR in CALRdel52 cells

Western blot of TpoR immunoprecipitation under non-reducing conditions showing associated CALR 50 kDa monomers and 100 kDa dimers (red arrowheads) present only in TF-1 TpoR CALRdel52 disrupted by 8-h treatment with 20 µg/ml 4D7 but not PBS or 20 µg/ml IgG. CALR monomers and dimers are detectable by polyclonal anti-wild-type CALR or anti-mutant CALR monoclonal antibodies. Red arrowheads, detected mutant CALR protein; brown arrowheads, detected wild-type CALR protein; asterisk, non-specific bands. Peripheral blood mononuclear cells from PMF samples were thawed and stained for CD34, CD14, CD19 and CD3 prior to FACS sorting. Each population was collected and purity was verified prior to proceeding with any further analysis. PCR amplification of CALR exon 9 was carried out to confirm mutational status of CD34+ cells which were utilized in megakaryocyte differentiation assays and colony forming assays in the presence of 4D7 or IgG. Representative flow cytometry plots for determination of CD41+/61+ populations from liquid culture assay from one PMF patient. Beads shown in the upper left panel with high SSC-A. Live cell population shown in hexagon gate (top panel). CD41/61+ population gates shown in lower panel with % CD41/61+ cells indicated for unstained, IgG- and 4D7-treated cells over 12 days. The number of CD41/61+ cells-to-bead ratio used to enumerate effect of 4D7 on megakaryopoiesis.

Source data are available online for this figure.

Monoclonal antibody 4D7 inhibits TPO-independent megakaryocyte formation

Primary myelofibrosis is characterized by abnormal proliferation and morphology of clonal megakaryocytes. To examine whether 4D7 suppressed CALR-mutant megakaryocytes, we tested its activity on purified primary CD34+ cells obtained from patients with CALR-mutant myelofibrosis using two orthogonal assays: (i) TPO-independent megakaryocyte differentiation in liquid culture, and (ii) TPO-independent megakaryocyte colony formation on a collagen-based medium. Five of eight patients had the most common type 1 (52 base pair deletion) CALR mutation which was heterozygous in the bulk of the CD34+ fraction (Fig 3A). The sorting strategy for myelofibrosis stem cells is shown in Fig EV4B and clinical details for all patients are listed in Table EV1. Strikingly, four of five mutant CALR MF patient samples that displayed robust TPO-independent growth of CD41+CD61+ megakaryocyte progenitors showed inhibition by 4D7 of at least 50% (Figs 3B and EV4C). Interestingly, the patient sample exhibiting less inhibition harboured a 34 base pair deletion. No significant inhibition was observed when JAK2V617F was present in all three myelofibrosis samples (black bars; Fig 3B). An additional two patient samples obtained from the University of Graz also showed evidence of inhibition compared to IgG control, including a type 2 CALR patient (five base pair insertion; Fig 3C). Overall, a mean decrease of 55% was observed across CALR-mutant samples after 4D7 treatment (P < 0.0001; Fig 3D). Similarly, we saw a dramatic reduction in the absolute numbers of primary TPO-independent megakaryocyte colonies cultured on collagen (colony-forming unit-mega) treated with 4D7 in multiple patient samples (decrease of 62%, P < 0.0001; Fig 3E–G). Strikingly, the size of 4D7-treated megakaryocyte colonies was substantially smaller as enumerated in a patient with a type 2 mutation (Fig 3H).

Details are in the caption following the image

Figure 3. 4D7 monoclonal antibody specifically inhibits primary megakaryocyte differentiation of mutated CALR myelofibrosis samples

PCR amplification of CALR exon 9 from patients confirming heterozygous del52 mutation in sorted CD34+ cells obtained from CALR-mutated myelofibrosis samples. Graph showing the decreased fold change of CD41+CD61+ megakaryocytes cultured in 4D7 compared to IgG. FACS-sorted CD34+ from myelofibrosis patients with CALR or JAK2 mutation was cultured over 12 days without TPO in the presence of SCF, IL-6 and IL-9. Black columns show JAK2V617F mutation-positive samples (n = 4 technical replicates of samples from different patients). Number of CD41+ megakaryocytes derived from isolated CD34+ progenitors from myelofibrosis patients. Number of CD41/CD61+ cells counted on day 12 using trypan blue exclusion (n = 1 technical replicate). Summary of fold change reduction of CD41/CD61+ megakaryocytes by 4D7 in all tested CALR-mutated patient samples compared to CALR wild type, normalized to IgG (n = 11 samples from different patients, with four technical replicates per sample). Number of CD41+ megakaryocyte colonies from patient samples after 4D7 treatment. CD34+ from patients with myelofibrosis with CALRdel52 or CALRins5 was plated on collagen-based matrix in presence of 20 µg/ml 4D7 or IgG control (n = 3 biological replicates). Representative micrographs showing CD41+ colonies in pink and CD41- colonies in blue after treatment with 4D7 or IgG at 100× or 40× magnification. Scale bar indicates 100 µm. Summary of fold change reduction in CD41+ megakaryocytes in CALR-mutated samples cultured MegaCult treated with 4D7 compared to IgG (n = 4 biological replicates). Number of megakaryocyte colony forming units grouped according to colony cell number after 4D7 treatment in a mutant CALRins5 sample (n = 3 biological replicates).

Data information: Bars represent standard error of means in (D) and (G). Unpaired Student's t-test used to determine statistical significance in (B, D, E, G and H). Bars represent standard deviations in (B, E and H). *P = 0.05–0.01, **P = 0.01–0.001, ***P = 0.001–0.0001, ****P < 0.0001, n.s, not significant.

Monoclonal antibody 4D7 inhibits the growth of ruxolitinib-persistent cells

Ongoing follow-up of COMFORT-1 and COMFORT-2 studies (Harrison et al, 2012; Verstovsek et al, 2012) suggests patients with CALR mutations are less responsive to the JAK inhibitor ruxolitinib than patients with JAK2 mutations, but cytopenia limits the use of higher treatment doses (Ross et al, 2021). We therefore tested whether an immunotherapeutic approach would have efficacy in cells with ruxolitinib “persistence/resistance” (Meyer et al, 2015). We first showed that ruxolitinib had inhibitory effects on megakaryocyte formation from healthy cord blood with > 30% inhibition after 12 days at 50 or 100 nM (P = 0.008 and 0.07, respectively; Fig 4A). In contrast, 4D7 had no effect on normal haematopoiesis. At the highest concentration, 4D7 did not inhibit physiological megakaryopoiesis in liquid culture (Fig 4B), haematopoietic granulocyte-macrophage colony formation and erythroid colony formation (Fig 4C), or collagen-plated megakaryocyte colony numbers (CFU-Mega; Fig 4D and E) compared to IgG control.

Details are in the caption following the image

Figure 4. 4D7 has no effect on normal progenitor cells and shows activity against ruxolitinib-resistant cells without haematological toxicity

Graph showing effect of ruxolitinib on megakaryocyte differentiation at 50 and 100 nM in healthy CD34+ cells (n = 4 different cords collected with three technical replicates). Effect of 4D7 on TPO-dependent megakaryocyte differentiation of healthy CD34+ cells cultured with 20 µg/ml 4D7 or IgG control (n = 2 biological replicates, with two technical replicates). Total numbers of haematopoietic colonies plated in MethoCult from healthy cord blood after treatment with 4D7 or IgG. CFU-GM, colony forming unit–granulocyte macrophage; BFU-E, blast forming unit–erythroid; CFU-GEMM, colony forming unit–granulocyte, erythroid, monocyte and megakaryocyte (n = 3 biological replicates). CD41+ megakaryocyte colonies from healthy CD34+ cells observed on collagen matrix after 4D7 treatment. Colonies were cultured in TPO, SCF, IL-9 and IL-3. Scale bar indicates 100 µm. Number of megakaryocyte colony forming units observed in MegaCult assay after treatment of healthy cord blood CD34+ cells with either IgG or 4D7 (n = 3 biological replicates). Western blot showing signalling in ruxolitinib-resistant TF-1 TpoR CALRdel61 compared to ruxolitinib-sensitive TF-1 TpoR CALRdel61 cells after treatment with 100 nM ruxolitinib for 16 h or 30 min and blotted for phospho-STAT5, total STAT5, phospho-ERK, total ERK and actin as indicated. Ruxolitinib-sensitive cells were non-viable after 16 h of treatment. Comparison of cell growth after DMSO, 20 µg/ml 4D7, IgG or 100 nM ruxolitinib treatment over 4 days of ruxolitinib-resistant TF-1 TpoR CALRdel61. Cells counted using trypan blue exclusion (n = 3 biological replicates with three technical replicates). Number of colonies observed from cells plated in MethoCult following 72 h of treatment with DMSO, 20 µg/ml 4D7, IgG or 100 nM ruxolitinib performed in ruxolitinib-resistant TF-1 TpoR CALRdel61 cells (n = 3 biological replicates with three technical replicates).

Data information: Error bars on (A, B, C and E) indicate standard error of mean. Error bars on (G and H) represent standard deviation. All P-values are unpaired Student's unpaired t-test.

Source data are available online for this figure.

To generate ruxolitinib-resistant cells, we cultured TF-1 TpoR CALRdel61 cells with increasing concentrations of ruxolitinib over 4 weeks, up to 100 nM. Ruxolitinib-resistant cells showed enhanced phospho-STAT5 and phospho-ERK phosphorylation in the presence of ruxolitinib but still showed some inhibition after withdrawal and re-stimulation (Fig 4F). Importantly, 4D7 had strong inhibitory activity on cells that were resistant to ruxolitinib, in both liquid culture at 96 h (Fig 4G) and in colony formation (Fig 4H). Together, these results suggest that an immunotherapeutic approach with 4D7 may have clinical utility in CALR-mutant patients who develop resistance/persistence of myelofibrosis during ruxolitinib treatment.

A recent report suggests TpoR co-expressed with mutant CALR may have a distinct conformation that is less responsive to physiological TPO compared to normal TpoR (Basso-Valentina et al, 2021). We therefore performed rescue experiments with 10 ng/ml TPO in the presence of 4D7. We found in the presence of 4D7, TPO could only partially rescue the proliferation defect in both TF-1 TpoR CALRdel61 cell line and patient-derived megakaryocyte colonies (Fig EV5A–C), suggesting endogenous TPO may not be sufficient to overcome 4D7 effect in vivo.

Details are in the caption following the image

Figure EV5. Effect of 4D7 antibody on TPO receptor biology

Proliferation curves of factor-independent TF-1 TpoR CALRdel61 cells cultured with 10 or 20 µg/ml 4D7 or 20 µg/ml of control IgG antibody in presence or absence of 10 ng/ml TPO (n = 3 biological replicates with three technical replicates). Number of CD41+ megakaryocyte colonies CALRdel52 patient after 4D7 treatment in the presence of TPO. Samples were seeded in a collagen-based matrix in presence of 20 µg/ml 4D7 or IgG control with 50 ng/ml TPO (n = 2 patient samples with two technical replicates). Representative micrographs showing CD41+ colonies in pink and CD41- colonies in blue after treatment with IgG or 4D7 in absence or presence of 50 ng/ml TPO. Scale bar indicates 100 µm.

Data information. Error bars represent standard error of the mean in (A) and standard deviation (B). Unpaired Student's t-test used to determine statistical significance in (B). *P = 0.05–0.01, **P = 0.01–0.001, ***P = 0.001–0.0001.

The 4D7 monoclonal antibody blocks CALR-dependent proliferation and prolongs survival in xenograft models

Myeloproliferative neoplasms are fundamentally disorders of unregulated proliferation. To test whether 4D7 could block mutant CALR-dependent proliferation in vivo, we developed two distinct cell-line xenograft models. The first was a bone marrow engraftment model, which measures mutant CALR-dependent proliferation of TF-1 TpoR CALRdel61 cells in the bone marrow microenvironment. The second was a chloroma model created by subcutaneous injection of TF-1 TpoR CALRdel61 cells into the flanks of NSG mice, which mimics extramedullary haematopoiesis. In both models, fibrosis is not an evaluable disease outcome. In the bone marrow engraftment model (Fig 5A), 4D7 treatment (12.5 mg/kg twice weekly via intraperitoneal injection starting day 7) showed an excellent pharmacokinetic profile, achieving a serum concentration of more than 100 µg/ml 48 h post-injection (Fig 5B), lowered peripheral blood engraftment of human CD33+ cells at 3 weeks (0.04 vs. 19.8% CD33+ 4D7 vs. IgG, P = 0.01; Fig 5C) and significantly prolonged survival (log-rank hazard ratio 0.24, P = 0.003; Fig 5D). In the chloroma model, 4D7 treatment significantly slowed tumour growth at 21 days post-engraftment (353 vs. 3,317 mm3 mean tumour volume, 4D7 vs. IgG, P = 0.04; Fig 5E) and prolonged survival (hazard ratio 0.19, P = 0.02; Fig 5F). Strikingly, mutant CALR cells induced to be resistant to 100 nM ruxolitinib also showed a survival advantage after treatment with 4D7 (12.5 mg/kg twice weekly beginning at day 7) in the bone marrow engraftment model (hazard ratio 0.26, P = 0.0008; Fig 5G). Together, these results suggest an immunotherapeutic approach with antibody 4D7 may have clinical utility in CALR-driven MPNs as well as in CALR-mutant patients who develop resistance/persistence to ruxolitinib.

Details are in the caption following the image

Figure 5. 4D7 blocks mutant CALR-dependent proliferation in vivo and prolongs survival

Illustration showing bone marrow NSG engraftment model with TPO-independent TF-1 TpoR CALRdel61 cells treated with 4D7 or IgG control twice weekly, starting 7 days after engraftment via intraperitoneal injection and final measurements taken from euthanized mice. Pharmacokinetic measurements of serum levels of 4D7 in mice after intraperitoneal injection at 0, 1, 24, 48, 72 and 110 h since administration. Percentage of TF-1 TpoR CALRdel61 human CD33+ cells measured in peripheral blood at 3 weeks post-tail vein engraftment (n = 5 mice per treatment). Kaplan–Meier survival curve of bone marrow engraftment model of TF-1 TpoR CALRdel61 cells showing improved survival of mice treated with 4D7 compared with IgG control commencing 1 week after tail vein injection (n = 5 mice per treatment). Mean tumour volume at 21 days after subcutaneous injection of TF-1 TpoR CALRdel61 cells following treatment with 4D7 or IgG. Cells were pre-treated with 4D7 or IgG control for 1 h prior to injection and treatment continued twice weekly at 12.5 mg/kg until euthanasia (n = 3 mice per treatment). Kaplan–Meier survival curve of TF-1 TpoR CALRdel61 chloroma mice treated with 4D7 or IgG until humane killing due to tumour diameter > 30 mm or ulceration (n = 3 mice per treatment). Kaplan–Meier survival curve of NSG mice engrafted with ruxolitinib-resistant TF-1 TpoR CALRdel61 treated with 12.5 mg/kg 4D7 or IgG twice weekly (n = 5 and 6 mice for IgG and 4D7, respectively).

Data information: For (C and E), a Student's unpaired t-test was used to determine statistical significance. For all survival curves, the log-rank Mantel–Cox test P-value is shown.

Primary myelofibrosis is an insidious and poorly understood disorder that encompasses features of both cancer and chronic inflammation. It is a clonal neoplasm driven by a handful of somatic mutations that activate cell signalling, presumably residing in the long-term stem cell compartment (Wernig et al, 2006; Nangalia et al, 2013; Reinisch et al, 2016), which also has a constellation of cytokine-mediated symptoms that are disproportionately severe. After the discovery of activating JAK2 mutations in 50–60% of patients with PMF (Baxter et al, 2005; Kralovics et al, 2005), clinicians were hopeful that myelofibrosis would respond to tyrosine kinase inhibitor therapy in a similar fashion to chronic myeloid leukaemia. However, maturing data from the COMFORT studies and other trials have shown ruxolitinib also does not cause widespread regression of fibrosis or eradicate disease clones in a mutation-specific manner (Verstovsek et al, 2012; Cervantes et al, 2013; Harrison et al, 2016). Exceptional cases of complete haematological or molecular remission have been reported (Masarova et al, 2016), but are uncommon. Typically, there is a gradual loss of response over time, with only 27% of patients remaining on first-line ruxolitinib treatment after 5 years in the COMFORT studies, underscoring a need for alternative approaches.

Somatic mutations in exon 9 of the endoplasmic chaperone protein CALR are found in 70–80% of patients with JAK2-negative PMF, accounting for ~ 30% of PMF cases overall, and also present in a similar proportion of essential thrombocythemia (Klampfl et al, 2013; Nangalia et al, 2013). In a subsequent analysis of patients from the COMFORT-2 study, the response rate to ruxolitinib was 20% (confidence interval 5.7–43.7%) in patients with a CALR mutation vs. 34% (24.8–44.1%) in CALR-negative patients, most of whom had a JAK2 V617F mutation (Guglielmelli et al, 2016), suggesting that CALR-mutant myelofibrosis may be less responsive to JAK inhibitor therapy. Although more than 50 CALR mutations have been described, the majority can be classified as type 1 like or type 2 like based on predicted protein structures (Eder-Azanza et al, 2014). Our data suggest that both type 1 and type 2 CALR mutations could be amenable to an immunotherapeutic approach.

Here, we show for the first time that primary cells from patients with CALR mutations can be targeted with a mutation selective immunotherapy. Our data confirm the observations from other studies using human cells that mutant CALR protein is secreted and accessible on the cell surface and can drive growth factor independence, provided TpoR is co-expressed. Data from independent laboratories have shown that mutant CALR protein requires TpoR for signalling and factor-independent cell growth, and that the normal lectin domain of CALR is essential to bind glycosylated sites on TpoR (Elf et al, 2016, 2018; Pecquet et al, 2019). The relative contribution of autocrine vs. paracrine vs. endosomal signalling of mutant CALR protein has not been fully examined, with most studies to date performed in murine IL-3-dependent Ba/F3 cells. In our human TF-1 model of mutant CALR-induced proliferation, no evidence of a paracrine effect could be demonstrated in a horizontal co-culturing assay (Fig EV1D). Remarkably, despite a demonstrated lack of autocrine/paracrine signalling, mutant CALR protein appears to be exquisitely sensitive to 4D7 binding in a dose-dependent manner with concomitant inhibition of cell proliferation. Our results indicate that the CALR-TpoR complex can be disrupted via a monoclonal antibody, and specifically we identified evidence for mutCALR bound constitutively to TpoR that could be inhibited by 4D7, thus blocking TpoR phosphorylation as well as downstream STAT and ERK signalling in primary samples. Future studies should determine the 3D structure of 4D7 bound to CALR and help test whether TPO analogues can overcome the effects that we describe (Basso-Valentina et al, 2021).

Recently, it was reported that mutant CALR protein is present at high levels in the plasma of myelofibrosis patients (compared to normal individuals) (Sollazzo et al, 2016; Pecquet et al, 2018). The cell surface expression of mutant CALR is an ideal target for immunological therapies. It is present on the cell surface (Pecquet et al, 2018; Balligand et al, 2020; Masubuchi et al, 2020) and contains sequences not normally present in healthy mammalian cells or conserved across mammalian species. CALR mutations appear to be an early event in MPN ontogeny (perhaps only preceded by TET2 in some cases). This results in it being present in all of the cells of the clone, unlike JAK2V617F which can be either an initiating or a secondary lesion, and found in only around half of patients with AML arising from an antecedent JAK2V617F MPN (Ross et al, 2021).

The development of a therapeutic antibody directed against mutant CALR protein has enormous clinical potential. The mouse model utilized within this study shows safety and efficacy for a rapidly proliferating tumour that is mutant CALR dependent, however, is not a relevant disease model for MPN. Future studies should investigate efficacy in relevant heterozygous mouse models (Li et al, 2018; Balligand et al, 2020; Achyutuni et al, 2021), including evidence for fibrosis reversal. Additionally, the antibody that we have developed against the common neoepitope C-terminus of CALR can be used in both insertion and deletion CALR mutation-positive patients and appears to have minimal or no effects on normal cells ex vivo. The safety and tolerability of this antibody therapy could lead to its application in the earlier stages of PMF to prevent clinical progression and the acquisition of clonal heterogeneity. In parallel, an anti-mutant CALR antibody such as 4D7 could also be combined with other treatments with the aim of augmenting their effect, as has been successfully applied in B-cell lymphoma using anti-CD20 antibodies (Czuczman et al, 1999). Future studies should examine a similar approach in patients with essential thrombocythemia with CALR mutations and secondary acute myeloid leukaemia that has transformed from CALR-mutant myeloproliferative disorders.

Myelofibrosis can be associated with immune defects which are compounded by the effects of JAK inhibitors like ruxolitinib. There is, therefore, an argument in favour of immunotherapeutics that can be used in the less advanced stages of myeloproliferative neoplasms. Although immunodeficiency may limit the potential of antibody therapeutics that require intact complement-mediated cytotoxicity, this can be overcome by the addition of toxin conjugates or the use of chimeric antigen receptor T cells. Importantly, our data suggest that antibody-mediated inhibition of cell signalling may also contribute to suppression of CALR-mutant cell proliferation, independent of any immune-mediated effects. Physiologically, the SIRP alpha protein is normally present as a ligand for the “don't eat me” CD47 signal protecting cells from phagocytosis (Majeti et al, 2009; Majeti, 2011; Liu et al, 2020) but inhibition of macrophage phagocytosis does not appear to be involved in the disease progression of MPN (Daitoku et al, 2016). It will be intriguing to test anti-CALR monoclonal antibody efficacy in the setting of macrophage activation or CAR T-cell therapies.

留言 (0)

沒有登入
gif