Preclinical human models and emerging therapeutics for advanced systemic mastocytosis

Abstract

Mastocytosis is a term used to denote a group of rare diseases characterized by an abnormal accumulation of neoplastic mast cells in various tissues and organs. In most patients with systemic mastocytosis, the neoplastic cells carry activating mutations in KIT. Progress in mastocytosis research has long been hindered by the lack of suitable in vitro models, such as permanent human mast cell lines. In fact, only a few human mast cell lines are available to date: HMC-1, LAD1/2, LUVA, ROSA and MCPV-1. The HMC-1 and LAD1/2 cell lines were derived from patients with mast cell leukemia. By contrast, the more recently established LUVA, ROSA and MCPV-1 cell lines were derived from CD34⁺ cells of non-mastocytosis donors. While some of these cell lines (LAD1/2, LUVA, ROSA^{KIT WT} and MCPV-1) do not harbor KIT mutations, HMC-1 and ROSA^{KIT D816V} cells exhibit activating KIT mutations found in mastocytosis and have thus been used to study disease pathogenesis. In addition, these cell lines are increasingly employed to validate new therapeutic targets and to screen for effects of new targeted drugs. Recently, the ROSA^{KIT D816V} subclone has been successfully used to generate a unique in vivo model of advanced mastocytosis by injection into immunocompromised mice. Such a model may allow in vivo validation of data obtained in vitro with targeted drugs directed against mastocytosis. In this review, we discuss the major characteristics of all available human mast cell lines, with particular emphasis on the use of HMC-1 and ROSA^{KIT D816V} cells in preclinical therapeutic research in mastocytosis.

Introduction

Mast cells (MC) are tissue-fixed cells found in all vascularized organs. These cells are involved in a number of physiological processes, such as adaptive and innate immune responses.1 Moreover, MC play a central role in many pathological conditions, including allergic reactions and mastocytosis.2 MC develop from bone marrow CD34/CD117 progenitor cells,3 which enter the circulation and migrate into tissues, where they mature into MC in response to their major growth factor, stem cell factor (SCF), the ligand of KIT, also known as CD117. KIT is a transmembrane receptor with intrinsic tyrosine kinase activity (Figure 1).4 Besides, mature tissue MC express the high affinity receptor for IgE (FcεRI) and can be activated through this receptor during allergic reactions.5

Figure 1.Normal structure of the KIT receptor and KIT mutations described in human mast cell leukemia-like cell lines and in patients with mastocytosis. In humans, KIT, located on chromosome 4q12, contains 21 exons transcribed/translated into a transmembrane receptor with tyrosine kinase activity (145 kDa and 976 amino acids). KIT is presented in its monomeric form, but dimerizes as a result of stem cell factor (SCF) ligation. The extracellular domain, in yellow, comprises five immunoglobulin (Ig)-like subunits where the ligand binding site (SCF for KIT) and the dimerization site are located. The cytoplasmic region contains a transmembrane domain (TMD) made by a single helix, in blue. The intracellular portion of KIT, in gray, contains an auto-inhibitory juxtamembrane domain (JMD) and a kinase domain (in dark gray) which is split into two parts: an ATP-binding domain (ABD) and a phosphotransferase domain (PTD) linked by a large kinase insert (KI) domain of ~ 60–100 residues. For the sake of clarity and readability, only the most frequent mutations found in patients and/or in human MC lines are represented. In red, mutants found in adult patients and in cell lines (KIT D816V found in >80% of adult patients with systemic mastocytosis and 30% of children with cutaneous mastocytosis, as well as in HMC-1.2 and ROSA^{KIT D816V} MCL-like cell lines, and KIT V560G found in the MCL-like cell lines HMC-1.1 and HMC-1.2, but only in a very few adult patients). In black, three of the KIT defects most frequently found in pediatric patients: KIT Del419 (NM_000222.2(KIT):c.1255_1257del, p.Asp419del), KIT ITD501-502 (NM_000222.2(KIT):c.1500_1505dup, p.Ser501_Ala502dup) and KIT ITD502-503 (NM_000222.2(KIT):c.1503_1508dup, p.Ala502_Tyr503dup) and in brown, the KIT K509I mutant found in several familial cases of the disease. For a complete overview of the various KIT mutations found in pediatric and adult mastocytosis patients, see Valent et al.⁷ Del, deletion; ITD, internal tandem duplication.

Mastocytosis designates a group of rare disorders characterized by a pathological accumulation of MC in one or more organs.6 Clinical presentations of mastocytosis range from skin-limited disease (cutaneous mastocytosis) occurring mainly in childhood and often regressing spontaneously, to systemic disease categories (systemic mastocytosis; SM), mostly seen in adults. SM variants usually involve the bone marrow and sometimes other internal organs, such as the spleen, liver, and/or gastrointestinal tract.

According to the World Health Organization (WHO), mastocytosis can be classified into three major categories: cutaneous mastocytosis, the most common variant, followed by SM, and MC sarcoma, a rare localized MC tumor (Online Supplementary Table S1).7 SM is subdivided into five distinct categories: indolent SM (ISM), smoldering SM (SSM), SM with an associated hematologic neoplasm (SM-AHN), aggressive SM (ASM) and MC leukemia (MCL) (Online Supplementary Table S1).7 While patients with ISM have a normal or near-normal life expectancy, patients with SM-AHN, ASM or MCL, collectively termed advanced SM, share a poor prognosis.8 The diagnosis of SM is based on WHO criteria and is established when one major criterion and one minor criterion or at least three minor criteria are present (Online Supplementary Table S2).9 Once the diagnosis of SM has been established, patients are further graded according to the presence of B-findings reflecting a high MC burden, and of C-findings reflecting organ damage related to MC infiltration (Online Supplementary Table S3).10 The pathophysiology of mastocytosis is complex and if acquired activating mutations in KIT (mostly KIT D816V: NM_000222.2(KIT):c.2447A>T, p.Asp816Val) seem to be major drivers of disease in ISM, the same cannot be said for advanced SM in which, in addition to KIT mutants, KIT-independent signaling pathways are activated and additional genetic defects are frequently found.

Given the complex pathophysiology of mastocytosis, in vitro models mimicking neoplastic MC found in SM patients could be useful for developing new therapeutic approaches. To date only a few human MC lines have been described, namely HMC-111 and its subclones (HMC-1.1 and HMC-1.2),12 LAD (subclones 1 through 5),13 LUVA,14 ROSA and its subclone ROSA,15 and MCPV-1.1 through MCPV-1.4.16 While LAD, LUVA and ROSA cells express KIT wild-type (WT), HMC-1.1, HMC-1.2 and ROSA cells harbor KIT activating mutations,1715 and MCPV-1 are RAS-mutated cells.16 Although these molecular aberrancies do not recapitulate all the characteristics of neoplastic MC found in advanced SM, these last four cell lines are currently the best available models for identifying molecular targets and defining the effects of several interventional (targeted) drugs currently used to treat advanced SM.

Pathophysiology of mastocytosis

The pathophysiology of mastocytosis is governed by the presence of KIT activating mutations in neoplastic MC.18 Indeed, various KIT activating mutations have been described, initially in patients with SM,19 then in children with cutaneous mastocytosis.20 In adult SM patients, KIT mutations affect primarily exon 17 encoding for the phosphotransferase domain, usually D816V (>80% of all patients) (Figure 1).21 Other less frequent mutations affect exons 2, 8 and 9 encoding for the extracellular domain or exons 13 and 14 encoding for kinase domain 1.21 By contrast, in children, KIT mutations are found in nearly 75% of biopsies of skin lesions, but the KIT D816V mutation is found in only 30% of all cases.20 Indeed, a significant percentage of children present with KIT mutants located in the extracellular domain (codons 8 and 9) (Figure 1).20

In KIT D816V SM patients, the development of neoplastic MC is principally governed by the PI3K/AKT and JAK/STAT5 signaling pathways activated downstream of KIT.2322 Indeed, AKT and STAT5 are constitutively acti vated in neoplastic MC in such patients and in KIT-mutated MCL-like cell lines, and inhibition of these pathways induces growth arrest in such cells.2422 Other intracellular pathways and molecules, such as the Feline sarcoma oncoprotein,25 or the mechanistic target of rapamycin (mTOR) complex,26 are also potential triggers of oncogenesis. In addition, the KIT mutant activates ERK independently of SRC, in contrast to KIT WT.27 Finally, LYN and BTK are found activated in neoplastic MC in a KIT-independent manner.28

Because KIT-activating mutations are found in most SM patients, several KIT-targeted tyrosine kinase inhibitors (KIT-TKI) have been developed. However, the nature of the mutation influences the sensitivity of the mutant to these TKI. For instance, the KIT D816V mutant is insensitive to imatinib.29 By contrast, patients presenting with KIT WT, or KIT mutant outside exon 17, may potentially respond to imatinib.30 While in ISM the KIT D816V mutant seems to be the unique molecular abnormality found, additional and recurrent somatic mutations of myeloid malignancy-related genes have been reported in advanced SM. The genes most frequently affected are TET2, SRSF2, ASXL1, RUNX1, JAK2, N/KRAS and CBL,3531 while EZH2, IDH2, ETV6, U2AF or SF3B1 are less often affected.36 All these mutations may be co-expressed with KIT D816V in the same cells or may be expressed in other myeloid cells but not in MC, especially in (A)SM-AHN with TET2, SRSF2 and ASXL1 mutants, in which acquisition of KIT D816V is often a late event conferring a mastocytosis phenotype on a pre-existing clonal condition.37 These defects, and particularly the SRSF2, ASXL1 and RUNX1 (S/A/R) mutations, have a negative impact on the disease prognosis.353431 Thus, targets other than KIT and drug combinations might be considered in order to develop more effective therapies for advanced SM. The potential of new targets and of new targeted drugs or drug combinations is currently investigated using the available human MCL-like MC lines, which will be described hereafter.

Major characteristics of available human mast cell leukemia-like mast cell lines

It is beyond the scope of this review to develop all the applications of the available human MC lines. We will only provide a detailed description of the cell lines that qualify as MCL-like, namely HMC-1, ROSA and MCPV-1. Indeed, these cell models have been and are still used in vitro and/or in vivo to evaluate the potential effects of drugs and drug combinations to treat mastocytosis. Table 1 shows a summary of the major characteristics of all available human MC lines, while Tables 2 and 3 provide detailed phenotypic information.

Table 1.Major characteristics of the available human mast cell lines.

Table 2.Expression of activation-induced antigens, cytokine receptors and molecular targets^† by the human mastocytosis-like mast cell lines.

Table 3.Expression of lineage-related markers and adhesion molecules^† by the human mastocytosis-like mast cell lines.

The origin, major characteristics and phenotype of HMC-1 cells are presented in Tables 1 through 4. HMC-1 cells are metachromatic cells containing histamine and tryptase.11 The original cell line presented with a complex karyotype (Table 1), which might have potentially played a role in the cells’ immortalization. However, despite this fact, HMC-1 cells remained sensitive to KIT inhibitors, in favor of a critical role of the KIT mutants in their maintenance. Indeed, in HMC-1 cells, KIT is constitutively phosphorylated on tyrosine residues in the absence of SCF.17 Sequencing of the coding region of KIT cDNA revealed that KIT in HMC-1 cells is composed of a normal WT allele and a mutant allele with two point mutations, KIT V560G (NM_000222.2(KIT):c.1679T>G, p.Val560Gly) and KIT D816V.38

Table 4.Expression by the human mastocytosis-like mast cell lines of antigens aberrantly expressed or overexpressed on malignant mast cells and/or their neoplastic progenitors in patients with systemic mastocytosis.

Seven years later, two HMC-1 subclones, namely HMC-1.1 and HMC-1.2, were described.12 Both subclones have a heterozygous KIT V560G mutation,12 but only HMC-1.2 cells display the KIT D816V mutation.12 In both subclones, KIT was found constitutively phosphorylated in the absence of SCF, although the presence of the KIT D816V mutation seemed to confer a slight growth advantage to HMC-1.2 cells over HMC-1.1 cells.12

HMC-1 cells have been extensively used to study KIT mutant-related and KIT mutant-independent signaling pathways and to evaluate anti-neoplastic effects of cytoreductive/targeted drugs developed to treat advanced SM.

The SCF-dependent ROSA cell line was established from a CD34 fraction of normal umbilical cord blood cells.15 CD34 cord blood cells were cultured in the presence of human SCF and, after an 8-week culture period, cells continued to proliferate, with virtually all cells being MC. The doubling time was relatively short (48–72 h) in the presence of SCF.15 ROSA cells are round cells with a relatively high nuclear-to-cytoplasm ratio and metachromatic cytoplasmic granules.15 ROSA cells stain strongly positive for tryptase and KIT, but express only little if any chymase.15

ROSA cells were found to express FcεRI, KIT (CD117), CD33, CD4, CD9, CD203c, and CD300a, consistent with a MC phenotype, while they did not express CD2 or CD25 (Tables 2 and 3).15 Moreover, similar to primary cord blood-derived MC, incubation of ROSA cells with interleukin-4 and IgE for 4–5 days enhanced surface expression of FcεRI. In addition, ROSA cells sensitized with interleukin-4 and IgE were fully activated by anti-IgE.15 However, over long periods of continuous culture, expression of FcεRI tends to fade on the cells, which become less sensitive to FcεRI cross-linking (unpublished observation).

ROSA cells have a normal KIT structure, but harbor a complex karyotype, with a derivative chromosome 1 [der(1)inv(1)(p31q21)del(1)(q24q32)]. In fact, the cell line consists of two subclones: one minor subclone carrying a complete trisomy 5, and the other predominant subclone carrying a partial trisomy 5 [+del(5)(q14q34)].15 In addition, molecular studies revealed that both ROSA subclones have a P53 deletion and a hot spot K700E mutation in SF3B1 (unpublished observation). We assume that these alterations contributed to the immortalization of ROSA cells and provide a premalignant (permissive) cellular background sufficient to trigger proliferation when a driver, such as KIT D816V, is introduced.

Indeed, when ROSA cells were further transfected with a lentivirus encoding for GFP + KIT D816V, the resulting subclone proliferated independently of SCF. This subclone, termed ROSA, has the same doubling time as ROSA cells cultured in SCF.15 Similar to their parental cells, ROSA cells have a rather mature morphology with numerous cytoplasmic granules.15

ROSA cells exhibit the same complex karyotype and the same SF3B1 K700E mutation as ROSA cells.15 Moreover, the phenotype of ROSA cells is similar to that of the parental cell line, including expression of the FcεRI and negativity for CD2 and CD25 (detailed in Tables 1–4), although, in contrast to ROSA cells, repeated attempts to activate ROSA cells by cross-linking FcεRI failed in our hands.15 Interestingly, KIT (CD117) is expressed at higher levels in ROSA cells than in ROSAcells.15 While KIT phosphorylation in ROSA cells needs the presence of SCF, KIT is constitutively phosphorylated in ROSA cells.15 In addition, STAT5 and AKT are constitutively phosphorylated in ROSA cells, as in primary neoplastic MC.392422 Interestingly, inhibition of AKT or STAT5 decreases ROSA cell proliferation.15 As expected, ROSA cells responded to imatinib, while ROSA cells were resistant to imatinib, but sensitive to dasatinib or midostaurin (PKC412),15 making these couple of cell lines a convenient tool for determining the relative selectivity of TKI towards the two forms of KIT (WT versus D816V).

Of note, ROSA cells were reported to engraft NOD/SCID IL-2Rγ (NSG) mice efficiently, giving rise to an ASM/MCL-like disease in vivo,15 described later in this manuscript. Thus, the ROSA cell line is a unique model of human KIT D816V ASM/MCL useful for in vitro and in vivo studies.

Finally, ROSA cells also appear well suited to investigating the transforming potential of KIT mutants found in other categories of mastocytosis. For example, starting from ROSA, we created ROSA subclones stably expressing the mutant KIT Del417-419insY (NM_000222.2 (KIT): c.1249_1255 delins T, p.Thr417_Asp419delinsTyr), or the mutant KIT K509I (NM_000222.2(KIT):c.1526A>T, p.Lys509Ile), both found in pediatric patients.20 In each case, the cells became SCF-independent, and KIT was found constitutively phosphorylated. In addition, both subclones remained, as expected, sensitive to the growth inhibitory effects of imatinib (unpublished observations). These additional data demonstrate that ROSA cells are reasonable tools for investigating the oncogenic potential of newly discovered KIT mutants as well as for screening for their sensitivity to TKI.

The human MCPV-1 subclones (MCPV-1.1 through -1.4) were generated from cord blood-derived CD34 progenitors by culturing these cells with SCF and interleukin-6 for 8 weeks and then stably transducing HRAS G12V, SV40 TAg and TERT.16 Single-cell clones were then isolated and cultured for more than 2 years to demonstrate immortalization. Light microscopy of Wright-Giemsa-stained MCPV-1.1 cells reveals large, immature cells with bi-, tri-, or multi-lobed (often cloverleaf-like-shaped) nuclei characteristic of MC precursors.16 MCPV-1 cells contain a basophilic cytoplasm, cytoplasmic protrusions and sparse granulation. Moreover, MCPV-1.1 cells exhibit an immunophenotype consistent with MC progenitors (Tables 2 and 3).16 MCPV-1 cells express tryptase but lack surface FcεRI.16 MCPV-1 cells grow independently of SCF and produce a MCL-like disease in NSG mice.

Human mast cell leukemia-like cell lines as models for in vitro testing of growth-inhibiting drugs

Treatment of ISM mainly aims at symptomatic relief of MC mediator symptoms.40 By contrast, treatment of advanced SM is challenging and relies principally on non-targeted and/or targeted cytoreductive therapy.41 In unusual cases (rare KIT-mutant forms or WT KIT) the disease may respond to imatinib or masitinib.434230 In a subgroup of patients with slowly progressing ASM, low-dose prednisolone and interferon-a may be efficacious.44 In addition, low-dose methylprednisolone and cyclosporine A may show some (usually minor) effects in ASM patients.45 Cladribine (2CdA) is often recommended as first-line therapy in patients with advanced SM with multi-organ involvement and slow progression.4746 A forthcoming new standard of therapy in advanced SM is midostaurin (PKC412).4948 This drug was approved for treatment of advanced SM by the American Food and Drug Administration and the European Medicines Agency in 2017. For ASM/MCL patients with rapid progression and those resistant to 2CdA or midostaurin, poly-chemotherapy is usually recommended, followed, when possible, by allogeneic hematopoietic stem cell transplantation.50 Almost all drug-based cytoreductive therapies have been validated preclinically in vitro using MCL-like MC lines. The most widely used cells for this purpose have been (and still remain) the two HMC-1 subclones. However, the newly emerging MCL-like human MC lines, ROSA and MCPV-1, have also been used repeatedly in such drug-testing studies. A summary of drug-testing approaches and of results obtained with these cell lines is provided in the following paragraphs.

Numerous antineoplastic drugs have been tested for their effects on HMC-1 cells.51 Among conventional anti-neoplastic drugs, doxorubicin and cytosine arabinoside were the most active agents.51 Other effective agents were vinblastine, etoposide and mitomycin.51 The potent effects of these chemotherapy-type drugs, otherwise used to treat acute myeloid leukemia, formed the basis to suggest treatment of patients with rapidly progressing ASM and MCL as well as patients with SM-acute myeloid leukemia with standard induction chemotherapy, often as preparation for allogeneic stem cell transplantation.

The effects of interferon(s) on the growth of HMC-1 cells have also been analyzed.52 HMC-1 cell numbers decreased in the presence of interferon-γ but were unaffected by interferon-α,52 contrasting with the activity of interferon-a in a subset of patients with advanced SM.44 This example highlights the fact that not all drug effects observed in vitro can be translated into clinical practice and that in each case, drugs and drug combinations need to be tested in additional disease models and finally in interventional clinical trials.

Studies of the in vitro anti-proliferative activity of 2CdA on HMC-1 cells were published after this drug was used in vivo to treat patients with advanced SM. Indeed, the first reports on the in vivo effects of 2CdA in patients were published between 2001 and 2004,5553 but it was not until 2006 that the in vitro effects of 2CdA on HMC-1 cells were described.56 While 2CdA alone produced growth-inhibitory effects on HMC-1 cells, the drug was also found to cooperate with midostaurin.56 The observation that midostaurin can induce apoptosis and growth inhibition in HMC-1 cells and that efficacy was identical in HMC-1.1 and HMC-1.2 cells prompted further investigations and led to the initiation of clinical trials.4948

Because most patients with SM harbor an activating point mutation in KIT (mostly KIT D816V) which is associated with disease pathology, considerable efforts have been made to identify drugs capable of inhibiting the kinase activity of the KIT mutant. The effects of imatinib, a drug targeting KIT WT, on cell lines harboring various KIT mutations, were investigated soon after the drug was found to block growth of leukemic cells in Philadelphia chromosome-positive chronic myeloid leukemia. In 2000, Ma et al. reported that imatinib inhibited KIT WT at low concentrations, without significant effects on the KIT D816V mutant.57 In 2003, these findings were confirmed using HMC-1.2 cells and patient-derived KIT D816V MC.5829 More recently, it was also confirmed that ROSA cells are insensitive to imatinib.15 Masitinib, another TKI active on KIT WT, although devoid of activity on KIT D816V in vitro,59 was administered in a randomized, double-blind, placebo-controlled, phase 3 study in a cohort of severely symptomatic ISM or SSM patients resistant to classical anti-mediator therapy.60 Interestingly, masitinib improved mediator-related symptoms in a subset of patients as compared to placebo-treated patients, regardless of KIT mutational status.60 This clinical activity was linked to the in vitro inhibitory effects of masitinib on two molecules involved in MC activation, namely LYN and FYN.59

Given the inefficacy of imatinib on the KIT D816V mutant, several other TKI have been evaluated in vitro (and for a few of them in vivo) for their potential activity in the SM context. Dasatinib is a multikinase inhibitor highly active on BCR-ABL1, KIT and PDGFRa.6261 The potential activity of this drug against KIT D816V was investigated in vitro in HMC-1 cells, SM patient-derived KIT D816V cells and ROSA cells.6315 In each instance, dasatinib exerted in vitro cytotoxic effects at relatively low half maximal inhibitory concentrations (IC₅₀), although the IC₅₀ for dasatinib was higher in KIT D816V cells than in KIT D816V cells.6315 However, when evaluated in vivo in clinical trials or in individual SM patients, dasatinib unexpectedly demonstrated only marginal activity.6664 While the in vivo effects of dasatinib have been disappointing, midostaurin, a potent multikinase inhibitor, has proven to be highly promising. Indeed, midostaurin decreased the proliferation of KIT D816V cell lines at pharmacological concentrations.67635615 In addition, the drug abrogated KIT phosphorylation in MCL-like cell lines harboring KIT D816V and induced their apoptosis.685615 Moreover, midostaurin suppressed the growth of primary human KIT D816V neoplastic MC.68 Finally, midostaurin was found to block IgE-dependent histamine release from MC and basophils.706967 Based on these data, clinical trials have been conducted to determine the efficacy of midostaurin in patients with advanced SM, with promising results.714948 An overview of the effects of midostaurin and of several other TKI on the growth of HMC-1 and ROSA cells is presented in Figure 2 and Online Supplementary Table S4.

Figure 2.Dose-dependent inhibition of the proliferation of wild-type and mutant KIT human mast cell lines by various tyrosine kinase inhibitors in vitro. Human mast cell lines expressing KIT D816V (HMC-1.2 or ROSA^{KIT D816V;} black lines) or lacking KIT D816V (HMC-1.1 or ROSA^{KIT WT}; gray lines) were incubated in control medium (Co) or in medium containing various concentrations of tyrosine kinase inhibitors (as indicated) at 37°C for 48 h. Thereafter, ³H-thymidine uptake was measured. Results are expressed as percent of ³H-thymidine uptake compared to the control and represent the mean ± standard deviation of three different experiments.

Several other TKI with different mechanisms of action were also found to exert antineoplastic effects in vitro on KIT D816V neoplastic MC, including HMC-1 cells. These drugs include 17-AAG (17-allylamino-17-demethoxygeldanamycin),72 EXEL-0862 (a TKI active against fibroblast growth factor receptors, vascular endothelial growth factor receptors, platelet-derived growth factor receptors, FLT3 and KIT),73 triptolide (a diterpenoid),74 ponatinib (a multi-kinase blocker),68 and bosutinib (a LYN/BTK-inhibiting TKI),28 which was administered to a patient with advanced SM, with no benefit.75 Nilotinib, which showed some effects in vitro on mutant KIT,76 was recently administered to 61 SM patients, with transient activity in some patients.77

More recently, several new KIT-TKI with inhibitory activity in vitro on several KIT mutants, including KIT D816V, have been developed. DCC-2618 (Deciphera Inc.) is a switch control type II KIT inhibitor, which arrests KIT in an inactive state, regardless of activating mutations, such as KIT D816V.78 In a recent study, it was found that DCC-2618 inhibits proliferation and survival of HMC-1.1, HMC-1.2 and ROSA cells at IC₅₀ <1 μM.79 BLU- 285 is a TKI developed by Blueprint Medicines. At low concentrations, BLU-285 selectively inhibited KIT D816V enzymatic activity (IC₅₀ = 0.27 nM).80 The cellular activity of BLU-285 on this mutant was also measured by autophosphorylation in HMC-1.2 cells with an IC₅₀ of 4.0 nM.80 Finally, BLU-285 inhibited in vitro the proliferation of KIT D816V HMC-1.2 cells with an IC₅₀ of 125 nM, while being less active on KIT D816V HMC-1.1 cells (IC₅₀ = 344 nM).81

Since KIT D816V is equally present in ISM and advanced SM patients, who have different life expectancies,82 the current assumption is that additional, KIT-independent pathways and pro-oncogenic hits and lesions are responsible for disease progression in advanced SM. Such pathways and pro-oncogenic molecules include LYN, BTK, STAT5, PI3-K, mTOR and members of the BCL-2 family.

For instance, LYN and BTK are phosphorylated in HMC-1.1 and HMC-1.2 subclones independently of KIT, and short interfering RNA against LYN and BTK decreased the survival of both subclones.28 In the same set of experiments, dasatinib blocked not only the kinase activity of KIT, but also LYN and BTK activation in neoplastic MC, while bosutinib inhibited LYN and BTK activation without decreasing KIT kinase activity.28

Another molecule, STAT5, seems critical for KIT D816V-driven proliferation in MCL-like MC lines as well as in neoplastic MC in SM patients.8324 Chaix et al. reported that the KIT D816V mutant can directly phosphorylate STAT5 in vitro.83 Interestingly, STAT5 is transcriptionally active in the HMC-1 cell line and in ROSA cells,2415 and drugs targeting STAT5 are effective in decreasing the growth rate of these cells.8584 Figure 3 shows representative curves of dose-dependent inhibition of the viability of MC lines by STAT5 inhibitors.

Figure 3.Dose-dependent inhibition of the proliferation of KIT D816V⁺ human mast cell lines by STAT5 inhibitors in vitro. ROSA^{KIT D816V} and HMC-1.2 cells were cultured in 96-well plates for 72 h in control medium containing 0.1% dimethylsulfoxide (DMSO) or with increasing concentrations (between 1.0 and 50.0 μM) of SF-1066, a very weak STA5 inhibitor (Ki >25 μM on STAT5), and of more specific and potent STAT5 inhibitors BP-1-102 and BP-1-108 (Ki >10 μM on STAT5).⁷⁴ Viability was calculated in each condition by the MTT method. Results are expressed as percent of control and represent the mean ± standard deviation of triplicate experiments. The half maximal inhibitory concentration (IC₅₀) at day 3 of each compound for each cell line was calculated using Prism GraphPad 4.0 software after plotting log concentration versus response. As expected, while the IC₅₀ for SF-1066 was >50 μM, IC₅₀ values for the more STAT5-specific compounds were lower: 11 μM for BP-1-102 on both KIT D816V⁺ neoplastic human MC lines and 34 and 22 μM for BP-1-108 for, respectively, HMC-1.2 and ROSA^{KIT D816V} cells. Although these values are still irrelevant at the pharmacological level, they open hopes that drug optimization might lead in a near future to the design of more potent small molecules inhibiting STAT5 activity.

Gabillot-Carre et al. reported constitutive activation of the mTOR signaling pathway in both HMC-1 subclones.86 However, the mTOR inhibitor rapamycin induced apoptosis only in HMC-1.2 cells.86 To support this unexpected selectivity, the authors demonstrated that rapamycin inhibited the phosphorylation of 4E-BP1, a downstream substrate of the mTOR pathway, only in HMC-1.2 cells.86 More recently, it was reported that the dual PI3-kinase/mTOR blocker NVP-BEZ235 has similar growth inhibitory effects in HMC-1.1 and HMC-1.2 cells.87 However, despite these encouraging data, no objective response was observed in a study in which everolimus, an oral mTOR inhibitor, was given at a dose of 10 mg daily to ten SM patients.88

Finally, aberrant accumulation of neoplastic MC in SM might result from deregulation of apoptosis pathways.89 Indeed, the anti-apoptotic molecules BCL-2, BCL-xL and MCL-1 are overexpressed in KIT D816V neoplastic MC in SM patients,9290 while the expression of the proapoptotic molecule BIM is suppressed in these cells.93 It has also been reported that MCL-1 is detectable in HMC-1.1 and HMC-1.2 cells.93 Moreover, exposure of these cells to MCL-1-specific antisense oligonucleotides or to MCL-1-specific short interfering RNA resulted in reduced cell survival and increased apoptosis.93 In further studies, evidence was provided that the pan-BCL-2 family blocker obatoclax inhibited the proliferation of HMC-1 cells, together with increased expression of PUMA, NOXA, and BIM mRNA, and apoptosis.94

Although drugs targeting KIT D816V have demonstrated activity on MC in vitro and in vivo, these agents do not cure patients with advanced SM.714948 Apart from several different mechanisms of resistance, neoplastic cells in these patients may exhibit a complex pattern of genetic alterations together with, or even often preceding the appearance of, the KIT mutant, as it is the case for TET2, SRSF2 and ASXL1 mutants,37 which could explain resistance to TKI. For this reason, attention has been focused recently on alternative targets which could help to overcome such resistance, namely surface antigens specifically or aberrantly expressed by neoplastic MC and epigenetic targets. Antibodies or drugs directed against these targets may also be able to overcome intrinsic neoplastic stem cell resistance, often associated with quiescence and altered drug influx or rapid drug efflux. Antibody-based drugs may exert antineoplastic effects independently of such mechanisms of resistance.

Several antigens are aberrantly expressed or overexpressed on neoplastic MC and on their progenitors in SM, including CD13, CD25, CD30, CD33, CD44, CD52, CD117 and CD123,9795 and might, therefore, be considered as potential therapeutic targets. Table 4 provides an overview of cell surface targets expressed on human MCL-like cell lines.

CD30 is aberrantly expressed by neoplastic MC in a subset of patients with SM, but not by normal/reactive MC.98 In a recent study, it was observed that the MCPV-1.1 subclone expressed high levels of CD30, while HMC-1.1 cells expressed low CD30 levels, and HMC-1.2 cells did not express CD30.99 The CD30-targeting antibody-conjugate brentuximab-vedotin inhibited the in vitro proliferation of neoplastic MC, with lower IC₅₀ values obtained for MCPV-1.1 cells (10 μg/mL) than for HMC-1.2 cells (>50 μg/mL).99 In addition, brentuximab-vedotin produced apoptosis in primary CD30 neoplastic MC.99 However, overall, the effects of brentuximab-vedotin on neoplastic MC are relatively weak. Correspondingly, no major clinical activity has been reported in clinical trials to date. In addition, neoplastic stem cells in advanced SM usually lack CD30 (personal information, PV).

In contrast to normal MC and MC from ISM patients, CD52 is abundantly expressed on neoplastic MC in most patients with advanced SM.16 Recently, it was reported that the CD52-targeting antibody alemtuzumab counteracts growth of MCPV-1.1 cells.16 Additionally, MCPV-1.1 cells were injected into NSG mice which were then treated with alemtuzumab or control vehicle. The alemtuzumab-treated mice had increased survival compared to controls, and reduced organ infiltration by neoplastic MC.16 Given that neoplastic (leukemic) stem cells identified in advanced SM may also express CD52,100 it can be hypothesized that a combination of a KIT-TKI and a monoclonal antibody against CD52 might help to achieve major antineoplastic effects in advanced SM.

Another potential surface target is CD33. In fact, CD33 is invariably expressed on neoplastic MC and their stem cells in patients with advanced SM.95 In addition, it has been described that the CD33-targeted antibody-construct gemtuzumab-ozogamicin is able to suppress growth and survival of neoplastic MC.101 In the light of the revival of gemtuzumab-ozogamicin, its clinical efficacy in patients with acute myeloid leukemia and its effects on neoplastic MC,102 it might be reasonable to propose clinical trials testing the effects of gemtuzumab-ozogamicin alone or in combination with other antineoplastic drugs or stem cell transplantation in advanced SM.

The epigenetic reader bromodomain-containing 4 protein (BRD4), a member of the BET family proteins, has recently been identified as a promising target in acute myeloid leukemia.104103 In addition, highly selective BET bromodomain inhibitors, including JQ1,105 I-BET151,107106 and I-BET762,108106 have demonstrated in vitro and in vivo activity against several hematopoietic malignancies. It has also been evaluated whether BRD4 might be a target in advanced SM. Indeed, BRD4 was found to be expressed in HMC-1.1, HMC-1.2, ROSA and ROSA cells as well as in primary neoplastic MC.109 Independently of the grade or variant of disease, neoplastic MC exhibit nuclear BRD4.109 However, in ASM and MCL, neoplastic MC also express substantial amounts of cytoplasmic BRD4.109 In line with this observation, HMC-1 and ROSA cells express cytoplasmic and nuclear BRD4 as well.109 The KIT-TKI midostaurin and dasatinib suppressed the expression of BRD4 in all MC lines.109 BRD4-specific short hairpin RNA and JQ1 decreased the proliferation of HMC-1 and ROSA cells.109 Based on these data, BRD4 is a promising target in advanced SM, although this needs to be confirmed in forthcoming clinical studies.

Human mast cell leukemia-like cell lines as tools to develop in vivo models

In vivo models have been developed in order to understand the pathophysiology of SM better. In addition to transgenic mouse models,111110 another approach is to create a SM-like disease in vivo by transplanting human neoplastic MC into immunodeficient mice.

The HMC-1 cell line engrafts immunodeficient mice after intravenous or subcutaneous injection, giving rise to subcutaneous tumors after 2 to 5 months.11211 The reason why intravenous injection does not give rise to a MCL-like disease is unknown, but limits the usefulness of HMC-1 cells to establish an in vivo model of advanced SM. More recently, an in vivo model of advanced SM was established using ROSA cells. Indeed, we engineered a ROSA subclone, termed ROSA, which naturally secretes Gaussia princeps luciferase (Gluc), used as a reporter.113 In this study, intravenous injection of NSG mice with ROSA cells led to an advanced SM phenotype, with neoplastic MC invading the bone marrow, spleen and liver, as testified by the quantification of engrafting cells by measuring Gluc reporter activity in peripheral blood and by an in vivo imaging system (IVIS).113 The detailed characteristics of this in vivo model are presented in Figure 4. All in all, this in vivo model of advanced SM is potentially the best available to date for in vivo testing of drugs previously identified as active in vitro on neoplastic MC.

Figure 4.Engraftment of ROSA^{KIT D816V-Gluc} cells in NSG mice. Increasing numbers of ROSA^{KIT D816V-Gluc} cells (1×10⁶, 5×10⁶ or 10×10⁶) were injected into the tail vein of NOD-SCID IL-2Rγ-/- (NSG) mice (3 groups; 6 mice per group) 24 h after irradiation at 2.5 Gy from a cesium-137 source.¹¹³ After 10 weeks, engraftment was assessed using (A) quantitative measurements of Gluc activity in plasma and (B) in vivo bioluminescence imaging (IVIS) on engrafted mice. Ten weeks after engraftment, mice were sacrificed and (C) bone marrow and (D) spleen sections from the three groups of mice were stained by immunohistochemistry with an anti-human tryptase antibody. Staining was visualized by a Histomouse Kit, showing human MC with a brownish stain, which massively invaded the bone marrow and to a lesser extent the spleen. (C) and (D) are from one representative mouse from the group injected with 10×10⁶ ROSA^{KIT D816V-Gluc} cells. Original magnification, ×100.

Summary and future perspectives

Despite decades of intensive research, only a few human MC lines have been established to date: HMC-1, LAD-2, LUVA, ROSA and MCPV-1. While none of these cell lines simultaneously expresses the KIT D816V mutant and a functional FcεRI, making them useless for testing MC-stabilizing drugs or drugs interfering with FcεRI-induced signaling in the context of KIT D816V SM, some of these cell lines may qualify as MCL-like since they harbor SM-related KIT variants and/or other oncogenic molecules relevant to SM. Among all MC lines, HMC-1 cells have been most frequently used, but other more recently established MC lines, such as ROSA and MCPV-1, are now available and are being used in various preclinical studies. For example, these cell lines have been used to analyze in vitro the growth-inhibitory effects of antineoplastic drugs, including various KIT-TKI, on neoplastic MC. However, because neoplastic MC in advanced SM are triggered by KIT-independent signaling pathways and additional genetic lesions that confer resistance against KIT-TKI, it might be interesting to establish in vitro models of multi-mutated neoplastic MC, starting from established human KIT mutant-positive MC lines in which additional lesions, such as the S/A/R combination of molecular lesions might be introduced. Such multi-mutated neoplastic MC lines should be useful to test combination therapies in vitro, and later in clinical trials in patients with advanced SM. With these approaches, new therapeutic concepts should be established in order to improve therapy in advanced SM.

Acknowledgments

The authors would like to thank Dr Patrick T. Gunning (Department of Medical Biophysics, University of Toronto, Ontario, Canada) for kindly providing the STAT-specific inhibitors presented in the manuscript. We also express our deep thanks to Dr Fawzia Louache (Inserm UMRS-1170, CNRS GDR 3697 Micronit, Institut Gustave Roussy, Université Paris-Sud, Université Paris-Saclay, Villejuif, France) for her invaluable help with the in vivo studies presented in the manuscript. K.H. is supported by a research grant from the German Research Council (DFG; HA 2393/6-1). DM is supported by the DIR, NIAID, and PV is supported by the Austrian Science Fund (FWF) grants F 4701-B20 and F 4704-B20.

Footnotes

• Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/103/11/1760

• Received May 4, 2018.
• Accepted June 27, 2018.