The FMS like Tyrosine Kinase 3 (FLT3) Is Overexpressed in a Subgroup of Multiple Myeloma Patients with Inferior Prognosis

Therapy resistance remains a major challenge in the management of multiple myeloma (MM). We evaluated the expression of FLT3 tyrosine kinase receptor (FLT3, CD135) in myeloma cells as a possible clonal driver. FLT3 expression was analyzed in bone marrow biopsies of patients with monoclonal gammopathy of undetermined significance or smoldering myeloma (MGUS, SMM), newly diagnosed MM (NDMM), and relapsed/refractory MM (RRMM) by immunohistochemistry (IHC). FLT3 gene expression was analyzed by RNA sequencing (RNAseq) and real-time PCR (rt-PCR). Anti-myeloma activity of FLT3 inhibitors (midostaurin, gilteritinib) was tested in vitro on MM cell lines and primary MM cells by 3H-tymidine incorporation assays or flow cytometry. Semi-quantitative expression analysis applying a staining score (FLT3 expression IHC-score, FES, range 1–6) revealed that a high FES (>3) was associated with a significantly shorter progression-free survival (PFS) in NDMM and RRMM patients (p = 0.04). RNAseq and real-time PCR confirmed the expression of FLT3 in CD138-purified MM samples. The functional relevance of FLT3 expression was corroborated by demonstrating the in vitro anti-myeloma activity of FLT3 inhibitors on FLT3-positive MM cell lines and primary MM cells. FLT3 inhibitors might offer a new targeted therapy approach in a subgroup of MM patients displaying aberrant FLT3 signaling.


Introduction
The therapy spectrum for multiple myeloma (MM) has considerably broadened over the past decade, which has led to an improved overall survival (OS), particularly in patients up to 80 years of age [1][2][3]. In the case of unsuccessful first-line therapy, there is a wide range of alternative drugs available, and the former "watch and wait" approach has been replaced by early treatment initiation [4]. The shift from a disease defined by symptoms to a disease defined by biomarkers enables diagnosis and treatment of MM prior to the occurrence of end-organ damage [4][5][6]. Nevertheless, therapy resistance due to genetic alterations, multifocal tumor growth, and subclonal structure remains a big challenge in the treatment of MM [7]. According to the Revised International Staging System (RISS), patients with translocation t (4;14), translocation t (14;16), and deletion del(17p), amongst others, are considered to be high-risk patients with reduced OS and progression-free survival (PFS) [8]. In order to provide patients with personalized therapy regimens, it is an unmet medical need to find additional biomarkers for improved patient stratification and prediction of therapy response as well as new druggable targets for therapeutic intervention [9,10].
In addition to the genetic alterations in MM cells, the respective bone marrow microenvironment supports tumor progression and development of therapy resistances [11,12]. Assessing the prognostic value of several angiogenic factors, our group found that FMS-like tyrosine kinase 3 ligand (FLT3L) levels in peripheral blood and bone marrow plasma correlate directly with the stage of disease [13]. Kokonozaki et al. detected relations between FLT3L serum levels and those of multiple other MM proliferation markers (B-cell activating factor, c-reactive protein, lactate dehydrogenase (LDH), interleukin-6, and bone marrow ki-67 proliferative index) as well as known soluble angiogenic factors, such as vascular endothelial growth factor, tumor necrosis factor-α, endoglin, and hematopoietic growth factor [14,15]. FLT3L activates FMS-like tyrosine kinase 3 (FLT3), a tyrosine kinase receptor type 3, which is strongly expressed in common lymphoid progenitors. The phosphorylation cascades initiated by the FLT3L/FLT3-complex in association with other cytokines, such as interleukin-7 and interleukin-11, have been proven to play an important role in the differentiation of B-cell progenitor cells [16]. Murine experiments showed that an inadequate expression of FLT3L results in deficiencies in hematopoietic progenitors, natural killer cells, and dendritic cells, whilst mice with a null-mutation of the FLT3 gene displayed deficiencies in primitive B-lymphoid progenitors [17,18].
In order to properly understand the role of the FLT3L/FLT3-complex in the course of MM, we analyzed FLT3 expression in bone marrow biopsies of patients with plasma cell dyscrasias (Table 1) by immunohistochemistry and real-time PCR and tested effects of FLT3 inhibitors on MM cell lines and on primary myeloma cells.  Table 1 represents the patient characteristics at the time of sample collection. For most patients of the newly diagnosed multiple myeloma (NDMM) group, the bone marrow samples that were used in this study were the ones that led to diagnosis. Thus, these samples were obtained before the first treatment cycle was started and the respective patients are listed amongst "therapy not started". The patients in the monoclonal gammopathy of undetermined significance/smoldering myeloma (MGUS/SMM) group did not receive therapy due to their premalignant condition. The category "therapy at the time of sample" solely refers to ongoing therapy at the time of sample collection. Three of the relapsed/refractory multiple myeloma (RRMM) patients had previously received treatment but had not yet been started on a new therapy regimen at the time of sample collection.

FLT3 Protein Is Expressed in Bone Marrow Samples of MGUS, NDMM, and RRMM Patients
FLT3 tyrosine kinase receptor protein (CD135) was analyzed in an acute myeloid leukemia (AML) patient with FLT3 ITD mutation displaying clusters of positive AML cells within adipocytes and stromal tissue ( Figure 1A). FLT3 is also expressed in bone marrow biopsies of all groups of MGUS/MM as detected by IHC ( Figure 1B-C). Interestingly, plasma cells from MGUS patients and SMM also showed a clear and strong FLT3 staining around the cell nucleus and in the cytoplasm. AML and MM cells were further examined for subcellular localization of FLT3 protein ( Figure 2). FLT3 displayed a vesicular staining pattern around nuclei of leucocytes in the tonsil, used as the positive control, as well as in AML and MM cells, suggesting localization within the endoplasmic reticulum (ER) and in vesicles trafficking to the cell membrane.

High FLT3 Protein (CD135) Expression in Myeloma Cells Correlates with Shorter Progression-Free Survival
FLT3 expression was evaluated by the use of a semi-quantitative immunochemistry (IHC) expression score, FES (range 1-6), taking into account the percentage of FLT3-positive cells and the intensity of FLT3 staining. There was no significant difference between the three patient groups in regard to FLT3 expression levels according to the FES ( Figure 3A). Due to the small sample size, the data from the MGUS/SMM patients (n = 3) were excluded from further statistical analyses. As we wanted to focus our investigations on a survival analysis, we also excluded seven patients for whom we were not able to collect substantial follow-up data. The remaining NDMM and RRMM patients (n = 32) were grouped for FES > 3 (high, n = 15) and FES ≤ 3 (low, n = 17). These patient characteristics are listed in Table 2. Comparing the survival of the two groups in a Kaplan-Meier analysis, we found a significantly shorter PFS in the patients with a FES > 3 in direct comparison to those with a FES ≤ 3 (p = 0.04; Figure 3B). However, we found no significant difference in OS between the two groups (p = 0.296, data not shown).
we were not able to collect substantial follow-up data. The remaining NDMM and RRMM patients (n = 32) were grouped for FES > 3 (high, n = 15) and FES ≤ 3 (low, n = 17). These patient characteristics are listed in Table 2. Comparing the survival of the two groups in a Kaplan-Meier analysis, we found a significantly shorter PFS in the patients with a FES > 3 in direct comparison to those with a FES ≤ 3 (p = 0.04; Figure 3B). However, we found no significant difference in OS between the two groups (p = 0.296, data not shown).
.     Table 2 represents the patient characteristics at the time of sample collection. For most patients of the NDMM group, the bone marrow samples that were used in this study were the ones that led to diagnosis. Thus, these samples were obtained before the first treatment cycle was started and the respective patients are listed amongst "therapy not started". The category "therapy at the time of sample" solely refers to ongoing therapy at the time of sample collection. Three of the RRMM patients had previously received treatment but had not yet been started on a new therapy regimen at the time of sample collection.

High FLT3 Expression Scores in Bone Marrow Biopsies Correlate with Increased ß 2 -Microglobulin Levels
As mentioned above, the NDMM and RRMM patients who were used in our survival analyses were grouped for FES ≤ 3 (n = 17) and FES > 3 (n = 15). Patient characteristics are listed in Table 2. The median age was 69 years (range 41-85 years) in the FES ≤ 3 group and 71 years (range 49-88 years) in the FES > 3 group.
As already mentioned, patients with a FES > 3 had a shorter PFS than patients with a FES ≤ 3 (p = 0.04; Table 3). Additionally, the median serum ß 2 -microglobulin-level was significantly higher in Cancers 2020, 12, 2341 7 of 16 patients with a FES > 3 than in patients with a FES ≤ 3 (p = 0.005; Table 3). Increased levels of serum ß 2 -microglobulin in patients with MM are known to be associated with a poor prognosis. Moreover, patients with a FES > 3 trended towards a higher ISS stage (p = 0.07; Table 3). There were no significant differences between the groups regarding sex, stage of disease, type of heavy chain or light chain, cytogenetic risk, decreased platelet count, decreased hemoglobin, serum LDH, serum creatinine, serum calcium, bone lesions, therapy lines, or therapy regimen. Table 3. Patients with a FES > 3 had a shorter progression-free survival (PFS), a higher ß 2 -microglobulin level, and revealed a trend towards a higher ISS stage in comparison to patients with a FES ≤ 3.

FLT3 Tyrosine Kinase Receptor Gene Expression Is Elevated in a Subgroup of MM Patients
After our antibody-based FLT3 protein analysis in bone marrow biopsies, we wanted to confirm FLT3 gene expression. Therefore, we isolated CD138 (syndecan-1, SDC1)+ cells from bone marrow aspirates of refractory MM patients (n = 4) and performed RNA sequencing (RNAseq). Data were compared to RNAseq data from blood mononuclear cells obtained from healthy donors (PBMNCs, n = 4).
Regarding the tyrosine receptor gene expression pattern of MM cells, we found expression of FLT3 in all samples analyzed. When we normalized samples to transcript number/million (TPM), FLT3 gene expression levels were higher than FLT1, but lower than for the Src family of tyrosine kinases BLK (B lymphocyte kinase), LYN, and HCK. Moreover, FLT3 gene expression was lower than FGFR3 (fibroblast growth factor receptor 3), BTK (Bruton's tyrosine kinase), JAK1-3 (Janus kinase 1-3), and the highly expressed serine-threonine kinase AKT1 (protein kinase B) ( Figure 4A). Compared to PBMNCs from healthy donors, MM cells from RRMM patients displayed higher expression levels of FLT3, FGFR3, and AKT1. Purity of CD138 +-MM cells was proven by checking SDC1 (CD138) gene expression and a set of established myeloma-marker/differentiation factors, such as MZB1 (marginal zone B and B1 cell-specific protein), XBP1 (x-box binding protein 1), IRF4 (interferon regulatory factor 4, MUM-1), SLAM7 (activating transcription factor 4), and TNF receptor superfamily member 17 (BCMA) in direct comparison to PBMNCs of healthy donors ( Figure 4B).   Based on our RNAseq findings on the small group of RRMM, we performed a more detailed FLT3 gene expression analysis on CD138 + -purified MM cells from bone marrow aspirates of a larger collective of NDMM and RRMM patients (n = 24). We quantified FLT3 gene expression levels with a hybridization-probe based multiplex real-time PCR and identified a small subgroup of patients with a higher FLT3 gene expression in NDMM and in RRMM ( Figure 5A). In six patients, the number of FLT3 copies per 50 ng of RNA exceeded 2000, whilst in most other patients it was below 1000. We found that this group, when compared to the rest of the patient collective, displayed tendencies towards a higher bone marrow plasma cell percentage. Future studies employing larger patient collectives are, however, necessary to confirm this trend. In addition, FLT3 gene expression levels were determined by real-time PCR in myeloma cell lines and AML cell line ( Figure 5B). As expected, AML cells exhibited high levels of FLT3 mRNA, followed by the MM cell lines KMS-12-BM and KMS-12-PE. The MM cell line NCI-H929 had very low amounts of FLT3 transcripts.

FLT3 Inhibitors Block Proliferation of MM Cell Lines and Induce Apoptosis of Primary MM Cells
Next, we investigated whether the number of transcripts correlated with the cell line response to FLT3 inhibitors. Human myeloma cell lines (KMS12-BM, KMS12-PE, NCI-H929) and AML cell line MOLM13 were treated with the FLT3 inhibitors midostaurin and gilteritinib, an AKTserine/threonine inhibitor (API-2) in increasing concentrations ranging from 10 to 1000 nM, and the standard MM proteasome-inhibitor (bortezomib, 5-20 nM). Proliferation was measured by 3 H-thymidine incorporation.
An inhibition of proliferation was induced by midostaurin and gilteritinib at concentrations ranging from 100 to 1000 nM in AML cells (MOLM-13, Figure 6A). All MM cell lines analyzed were inhibited by midostaurin at 1000 nM, independently of FLT3 expression levels ( Figure 6B-D). In

FLT3 Inhibitors Block Proliferation of MM Cell Lines and Induce Apoptosis of Primary MM Cells
Next, we investigated whether the number of transcripts correlated with the cell line response to FLT3 inhibitors. Human myeloma cell lines (KMS12-BM, KMS12-PE, NCI-H929) and AML cell line MOLM13 were treated with the FLT3 inhibitors midostaurin and gilteritinib, an AKTserine/threonine inhibitor (API-2) in increasing concentrations ranging from 10 to 1000 nM, Cancers 2020, 12, 2341 9 of 16 and the standard MM proteasome-inhibitor (bortezomib, 5-20 nM). Proliferation was measured by 3H-thymidine incorporation.
An inhibition of proliferation was induced by midostaurin and gilteritinib at concentrations ranging from 100 to 1000 nM in AML cells (MOLM-13, Figure 6A). All MM cell lines analyzed were inhibited by midostaurin at 1000 nM, independently of FLT3 expression levels ( Figure 6B-D). In contrast, the more specific FLT3 inhibitor gilteritinib inhibited cell proliferation only in MM cells with higher FLT3 gene transcription (KMS12-PE and KMS12-BM), not in NCI-H929 displaying low amounts of FLT3 transcripts. The AKT inhibitor (API-2) was less effective in inhibiting cell proliferation than midostaurin. Cancers 2020, 12, x 10 of 17 Since primary myeloma cells might not proliferate extensively, we chose to assess the impact of the inhibitors on the viability of plasma cells. Therefore, we prepared bone marrow mononuclear cells (by density gradient centrifugation) and treated the malignant plasma cells together with their environmental cells (mononuclear cells) for 24 h with different concentrations of FLT3 inhibitors. The cell fraction containing plasma cells were gated according to CD38 high /CD45 low/neg staining. From these cells, we assessed apoptotic cells, i.e., cells that were stained by the apoptosis marker efluoro 780 and/or AnnexinV (for gating strategy see Figure 7A). Moreover, we determined FLT3 gene expression in CD138 + -purified MM cells from the respective patients ( Figure 7B).
FLT3 inhibitors induced apoptosis in a subgroup of patients at low concentrations ( Figure 7B). Gilteritinib was acting only on patients MM#257 and MM#258 at 100 and 1000 nM. Patients MM#261 and MM#262 neither showed a response to standard MM therapy (bortezomib) nor to FLT3 inhibitors. Since primary myeloma cells might not proliferate extensively, we chose to assess the impact of the inhibitors on the viability of plasma cells. Therefore, we prepared bone marrow mononuclear cells (by density gradient centrifugation) and treated the malignant plasma cells together with their environmental cells (mononuclear cells) for 24 h with different concentrations of FLT3 inhibitors. The cell fraction containing plasma cells were gated according to CD38 high /CD45 low/neg staining. From these cells, we assessed apoptotic cells, i.e., cells that were stained by the apoptosis marker efluoro 780 and/or AnnexinV (for gating strategy see Figure 7A). Moreover, we determined FLT3 gene expression in CD138 + -purified MM cells from the respective patients ( Figure 7B).  FLT3 inhibitors induced apoptosis in a subgroup of patients at low concentrations ( Figure 7B). Gilteritinib was acting only on patients MM#257 and MM#258 at 100 and 1000 nM. Patients MM#261 and MM#262 neither showed a response to standard MM therapy (bortezomib) nor to FLT3 inhibitors.

Discussion
During the past decade, the use of additional biomarkers has become more and more prevalent in the management of MM. As of 2014, the IMWG diagnostic criteria (myeloma defining events) include three biomarkers (clonal bone marrow plasma cell percentage ≥ 60%, FLC ratio ≥ 100, ≥1 focal lesion in MRI studies) in addition to the traditional CRAB criteria (hypercalcemia, renal damage, anemia, bone lesions) [5]. The R-ISS divides myeloma patients into the stages I-III depending on four different biomarkers (serum ß2-microglobulin, serum albumin, chromosomal abnormalities, and serum LDH) [8]. Due to heterogeneous genetic alterations, comorbidities, and other individual characteristics, the response to therapy can be very heterogeneous amongst MM patients. Hitherto, there is no single prognostic biomarker available that reflects directly on disease promoting dynamics in the bone marrow microenvironment.
In a former study, we established that plasma levels of FLT3L in MM patients correlate with the stage of disease [13]. In our current work, we demonstrate that FLT3, the receptor tyrosine kinase that FLT3L binds to, is expressed in MM cells. To the best of our knowledge, this study is the first to confirm the presence of FLT3 in CD138 + plasma cells in patients with MGUS/SMM, NDMM, and RRMM by the use of IHC, rt-PCR, and next generation sequencing (NGS). We developed a semi-quantitative IHC FLT3 expression score (FES) that takes into account the extent and intensity of FLT3 staining (range 1-6, score ≤ 3 and >3). Comparing the two patient groups, we found that patients with a FES > 3 had a significantly shorter PFS than those with a FES ≤ 3 (p = 0.04). This pertained to patients that were newly diagnosed at the time of sample collection as well as to relapsed/refractory patients.
In recent years, risk stratification, therapy planning, and biomarker research for MM evolved around cytogenetic alterations. According to the R-ISS, patients with translocation t(4;14), translocation t(14;16), and deletion del(17p), amongst others, are considered to be high-risk patients who face a worse OS and PFS [8]. Recently, Laganà et al. created a computational platform for personalized treatment recommendations for RRMM. The authors generated patient profiles based on DNA mutations and RNAseq data. In their mRNA expression findings, they listed the FLT3 gene amongst "overexpressed actionable deregulated genes" such as MAPK, FGFR3, and PI3K/AKT [19]. The FLT3 cytoplasmic domain has been shown to associate with the p85 subunit of phosphoinositol-3-kinase (PI3K), monomeric RAS, and phospholipase C-γ, and results in the phosphorylation and hence activation of these proteins [20] and consequently also further downstream pathways of PI3K/AKT and mitogen-activated protein kinase (MAPK) pathways [21,22]. Thus, targeting the PI3K/AKT or BRAF/MEK pathway might also contribute to blocking FLT3 downstream signaling in MM cells. In the clinical setting, data on the treatment of MM with BRAF-MEK inhibitors are emerging, so far confined to the specific setting of extramedullary disease [23]. It will be interesting to investigate if combining FLT3 inhibitors with BRAF/MEK inhibition reveals synergistic therapeutic effects.
FLT3 inhibitors are already in use in the treatment of AML. In approximately 30% of AML patients, the FLT3 receptor is mutated [24]. An internal tandem duplication (ITD), which leads to autonomous FLT3 activity, can be found in approximately 20% of AML patients [22,25]. FLT3 mutations are "actionable" mutations, which can be targeted with FLT3 inhibitors, such as the first generation FLT3 inhibitor midostaurin. Second generation FLT3 inhibitors, such as gilteritinib, are currently in clinical trials [26][27][28]. Although FLT3 is not mutated in MM, elevated serum levels of FLT3L in RRMM patients [13] might result in increased activation of the downstream phosphorylation cascades, which could potentially be prevented by FLT3 inhibitors. According to this idea, we demonstrate in our study that FLT3 inhibitors (i.e., midostaurin and gilteritinib) exert promising anti-myeloma activity in vitro. The most effective induction of apoptosis in primary MM cells was achieved by high dosages of midostaurin in four out of six patient samples. Interestingly, the two patient samples not affected by midostaurin showed also no response to the standard myeloma drug bortezomib. A limitation of our study was the short culture conditions for MM cells ex vivo from bone marrow aspirates (24 h) that required higher concentrations of FLT3 inhibitors. Basically, MM patients can be treated over a longer time window with lower concentrations of FLT3 inhibitors to achieve growth inhibition or apoptosis of MM cells.
Further limitations of our study are the retrospective character and the small number of MM patients for extensive statistical analysis. Clearly, a further prospective study with a larger patient population is needed to substantiate our findings and carve out the full potential of FLT3 and its ligand in the context of prognosis and targeted therapy of MM.

Patients and Cell Lines
According to the IMWG criteria, patients with MGUS (n = 2), SMM (n = 1), NDMM (n = 26), and RRMM (n = 13) were included in the study population (Table 1) [29]. The study was designed and conducted in accordance with national and international guidelines and with the ethical standards of the Declaration of Helsinki. Investigations have been approved by the authors' institutional review board (1163/2017 and 1220/2018 Innsbruck).

Immunohistochemistry of Bone Marrow Biopsies
IHC for the detection of FLT3 was performed on an automated platform (Benchmark ULTRA , Ventana Medical Systems, Tucson, AZ, USA) using a polyclonal antibody against FLT3 (CD135, dilution 1:100, SAB 2900166, Sigma-Aldrich, Darmstadt, Germany). Specificity of FLT3 signals in the ER was proven by competitive blocking of signals with a 25-fold molar excess of human recombinant FLT3 protein (Cat: 10445-H08H, Sino-Biological, Beijing, China) in the first antibody incubation step. Three cases of FLT3-mutated AML and human tonsils served as positive controls for FLT3, respectively. In brief, sections were cut at 1.5 µm from formalin-fixed, paraffin-embedded, and decalcified tissue blocks of bone marrow trephine biopsies. After de-paraffinization, slides were heat-pre-treated with cell-conditioning reagent 1 (CC1, Ventana Medical Systems, Tucson, AZ, USA) for antigen retrieval, and primary antibodies against FLT3 were incubated for 32 min at 37 • C. For visualization, the Ultra View DAB Detection Kit (Ventana Medical Systems, Tucson, AZ, USA) was used in accordance with the manufacturer's recommendation. Finally, slides were washed in distilled water, counterstained with Hematoxylin (12 min) and Bluing Reagent (4 min), dehydrated in descending order of alcohols, cleared in xylene, and cover-slipped with Tissue-Tek mounting medium (Sakura Finetek, Tokjo, Japan).

Scoring System for FLT3 Staining
All routinely prepared slides were re-evaluated by two observers (AB, NS) in regard to diagnosis and plasma cell content. Visual assessment of FLT3 (CD135) expression was then performed by applying a semi-quantitative FLT3 expression score (FES), taking into account the intensity of FLT3 staining of plasma cells (weak staining = 1, strong staining = 2) and the extent of FLT3-positive stained plasma cells/total plasma cells (0-10% FLT3-positive plasma cells = 1, >10-60% = 2, >60% = 3). The final score was calculated by multiplying the extent of stained cells (1-3) with the staining intensity (1-2) and ranged from 1-6. Only cytoplasmic staining was considered positive for FLT3.

Conclusions
A high FLT3 expression score (FES > 3) is associated with a shorter PFS in patients with NDMM and RRMM. FLT3 inhibitors showed activities in a subgroup of MM primary cells and MM cell lines. Thus, FLT3 inhibitors might offer a new targeted therapy approach for a subgroup of patients with ectopic FLT3 signaling.