Serotonin receptor type 1B constitutes a therapeutic target for MDS and CMML

Myelodysplastic syndromes (MDS) and chronic myelomonocytic leukemia (CMML) are chronic myeloid clonal neoplasms. To date, the only potentially curative therapy for these disorders remains allogeneic hematopoietic progenitor cell transplantation (HCT), although patient eligibility is limited due to high morbimortality associated with this procedure coupled with advanced age of most patients. Dopamine receptors (DRs) and serotonin receptors type 1 (HTR1s) were identified as cancer stem cell therapeutic targets in acute myeloid leukemia. Given their close pathophysiologic relationship, expression of HTR1s and DRs was interrogated in MDS and CMML. Both receptors were differentially expressed in patient samples compared to healthy donors. Treatment with HTR1B antagonists reduced cell viability. HTR1 antagonists showed a synergistic cytotoxic effect with currently approved hypomethylating agents in AML cells. Our results suggest that HTR1B constitutes a novel therapeutic target for MDS and CMML. Due to its druggability, the clinical development of new regimens based on this target is promising.

Due to the key role that HTR1s and DRs play in AML 23,24 and the close physiopathological relationship between AML and MDS, the expression of HTR1s and DRs was screened by flow cytometry in BM samples from MDS patients. Similar to AML, MDS patient samples highly expressed HTR1A (Fig. 1A) and HTR1B (Fig. 1B) on the cell surface. The expression of HTR1A and HTR1B on MDS BM cells was 3.4-and 3.7-fold higher compared to healthy donor samples (HD), respectively. Interestingly, expression levels of HTR1A and HTR1B were comparable in MDS and AML patients.
Similarly, DRD3 and DRD5 were highly expressed in MDS and AML patient samples as compared to HD blood cells (Fig. 1C,D). The expression of both DRs was equivalent, displaying a 3.5-fold increase with respect to HD cells. Although all samples expressed DRs on the surface, only DRs type 3 (DRD3) and 5 (DRD5) were significantly higher expressed than HD blood cells ( Figure S1). Moreover, DRD3 and DRD5 expression have been comprehensively studied in the context of AML 23 . Consequently, DRD3 and DRD5 were chosen for further validation as biomarkers for MDS cells.
MDS is driven by a complex combination of genetic and epigenetic changes that result in a wide heterogeneity in both clinical phenotype and disease outcome. According to the World Health Organization (WHO) classification 4 , MDS is a clonal disease characterized by morphological dysplasia, ineffective hematopoiesis leading to cytopenias and risk of transformation to AML 4 . MDS is subclassified into clinically relevant groups mainly based on morphological and cytogenetic criteria 4 . Thus, we next interrogated the expression of HTR1A/B and DRD3/ D5 within the most frequent MDS subtypes ( Figure S2). As shown in Fig. 2A-D, the expression of HTR1A, HTR1B, DRD3 and DRD5 was significantly higher in all MDS subgroups as compared to HD samples, with the exception of MDS-RS (MDS with ring sideroblasts) with multilineage dysplasia (MDS-RS-MLD), which showed only an increase in DRD3 expression. Indeed, the frequency of positive cells for HTR1s and DRs was above 50% in MDS-RS with single lineage dysplasia (MDS-RS-SLD); MDS with isolated del(5q) (MDS-5q); MDS with multilineage dysplasia (MDS-MLD); and MDS with excess blasts type 1 (MDS-EB-1) and type 2 (MDS-EB-2) ( Fig. 2A-D) (Table 1). Notably, MDS-RS-MLD samples expressed HTR1s and DRs similarly to HD samples. Due to the cell subpopulation heterogeneity within MDS samples, the expression of HTR1A/B and DRD3/5 was also interrogated in the CD34-positive cell population, which corresponds to the most primitive fraction. Expression levels of each receptor were similar in the bulk population and CD34 + fraction ( Figure S3). Indeed, CD34-positive HD cells lacked the expression of HTR1A/B ( Figure S3) and DRD3/5 23 .
The natural history of MDS varies considerably among individuals, which correlates with the mosaic of subtypes of MDS 3 . In order to better discriminate prognostic risk for assessing clinical outcomes in MDS, IPSS-R was developed based on five disease factors (percentage of blasts, cytogenetics, hemoglobin concentration, absolute neutrophil and platelet count) that categorized MDS patients into five risk categories: very low, low, intermediate, high, and very high 31 . Therefore, we further analyzed if the expression of HTR1A, HTR1B, DRD3 and DRD5 correlated with the diverse IPSS-R prognostic categories. As shown in Fig. 2E-H, the expression of HTR1A, HTR1B, DRD3 and DRD5 was independent of prognostic-risk, since no significant correlation was observed between the surface expression of these receptors and any of the IPSS-R prognostic groups. Inhibition of HTR1B in MDS cells induces cytotoxicity. HTR1s have been recently described as therapeutic targets in AML 24 . Indeed, inhibition of HTR1A and HTR1B leads to a reduction in cell viability coupled with the induction of the differentiation program of AML cells both ex vivo and in vivo 24 . Since MDS samples differentially expressed HTR1A and HTR1B similarly to AML (Figs 1 and 2), their sensitivity to HTR1A/B antagonists was evaluated ex vivo. MDS samples corresponding to MDS-RS-MLD (violet), MDS-RS-SLD (purple), MDS-5Q (red), MDS-MLD (orange) and MDS-EB1/2 (blue) were treated for 72 h with apomorphine (a promiscuous dual HTR1/2 antagonist and a DR partial agonist), methiothepin (a broad-spectrum HTR antagonist), SB-224289 (a highly selective HTR1B antagonist), and NAN-190 (a highly selective HTR1A antagonist), and subsequent cell viability and the differentiation stage were measured (Fig. 3B). A wide range of cell response to the different agents assayed was observed, in accordance with the intrinsic heterogeneity within MDS disease behavior among patients. An overall statistically significant but clinically irrelevant reduction in cell viability was observed upon apomorphine and SB-224289 treatment (Fig. 3B); cellular responses to methiothepin were heterogeneous, probably due to a lower affinity. On the contrary, NAN-190, the selective HTR1A antagonist, spared MDS samples in terms of cell survival (Fig. 3B). When the analysis was focused on MDS-EB-1/2 samples (blue), a higher sensitivity to HTR1 antagonists was observed ( Similarly to HTR1 antagonism 24 , DR antagonism in AML produces a reduction in cell viability coupled with the initiation of the differentiation program 23 . We therefore interrogated the cellular effect of the treatment with DR antagonists such as SCH-23390 (SC90) (DRD1 and DRD5 antagonists), UH-232 (DRD2 and DRD3 antagonist), chlorpromazine (CPZ) (pan-DR antagonist), and thioridazine (Thio) (pan-DR antagonist) in MDS.  Surprisingly, and in contrast to AML 23 , none of the DR antagonists was cytotoxic for MDS samples (Fig. 3F), even though DRs were overexpressed on these patient cells. Only thioridazine at a 50 μM concentration produced a therapeutic significant induction of cell death ( Figure S4C); however, at this concentration, thioridazine has been proven toxic to normal cells 23 . In contrast to HTR1B antagonists, DR antagonists showed neither cytotoxicity nor induction of differentiation in any MDS subgroup ( Figure S4B,C).
Recently, HMAs such as azaciditine and decitabine have shown a clinical benefit for high-risk MDS patients, becoming a major advance in the treatment of these patients. However, 50% of MDS patients are non-responders and the majority of responders relapse within 2 years 32 . As both agents act on proliferative cells (S-phase) and treatment with HTR and DR antagonists induces cell cycle entry 23,24 , a potential synergistic cytotoxic effect of the combination of HTR1 antagonists and HMAs was investigated. To date, only 2 MDS cell lines have been generated: MDS92 33 and its derivative MDS-L 34,35 . MDS-L successfully reproduces the disease in xenograft models 34 . However, the expression of DRs and HTR1s was absent in these cell lines, in contrast to primary MDS patient samples ( Figure S5). Therefore, the pharmacologic interaction between HTRs antagonists and HMAs was analyzed in AML cells that express DRs and HTRs similarly to MDS (Fig. 1), which express DRs and HTR1s similarly to MDS primary samples. The presence of apomorphine synergistically interacted with azaciditine inducing cytotoxicity (Fig. 4A,B). Indeed, 100 nM azaciditine in the presence of 5 μM apomorphine induced equivalent cell death level as 1 μM azaciditine (Fig. 4A). Similarly, decitabine treatment also showed synergistic anti-leukemic effect in combination with apomorphine ( Fig. 4C,D). Moreover, synergism between drugs could be demonstrated based on both combination index (CI) 36 (Fig. 4B,D) and Excess Over Bliss additivism (EOBA) 37 ( Figure S6A,B). Indeed, equivalent results were obtained with methiothepin in combination with HMAs ( Figure S6C,D). Next, interaction between different DR antagonists with the HMAs was evaluated similarly to HTR1 antagonists. In concordance to the cytotoxicity data observed with the use of DR antagonists alone, the anti-neoplastic effect of azaciditine (Fig. 4E,F) and decitabine (Fig. 4G,H) was not potentiated in the presence of the DR antagonists thioridazine (Thio) and SCH-224289. Taken together, HTR1 constitutes a potential therapeutic target in combination with currently approved HMAs in MDS.

CMML cells are sensitive to both HTR1B and DRs' antagonists.
CMML is a clonal hematopoietic stem cell disorder, characterized by overlapping features between MDS and myeloproliferative neoplasms, with an inherent tendency to transform to AML 38 . Based on these characteristics, the expression of HTR1 and DR was also analyzed in CMML patient samples. Similar to MDS, CMML samples differentially expressed HTR1A (Fig. 5A), HTR1B (Fig. 5B), DRD3 (Fig. 5C), and DRD5 subtypes (Fig. 5D). Next, sensitivity to HTR1 antagonists' treatment was assayed in similar conditions as described for MDS samples. In concordance with the expression profile, apomorphine, methiothepin, and SB-224289 reduced cell viability at 10 μM (Fig. 5E). In contrast to MDS, CMML samples were also sensitive to treatment with DR antagonists (such as CPZ, UH-232, SCH-23390 and Thio) (Fig. 5F). Interestingly, the inhibition of HTR1B resulted in the reduction of the clonogenic capacity of CMML samples (Fig. 5G-H). These results suggest that HTR1B and DRs may also act as a therapeutic target for CMML analogously to AML.

Simultaneous inhibition of HTR1 and DR enhances the cytotoxic effect.
Oligomerization is a general characteristic of GPCRs, and homo-and heterodimerizations are found in DRs and HTR1s 39 . Actually, DRs/ HTR1s complexes have been described 40 . The HTR and DR expression patterns observed in MDS samples were comparable (Figs 1 and 2). Indeed, HTR1A, HTR1B, DRD3 and DRD5 were expressed in each sample at the same level in both MDS and CMML (Fig. 6A). The cytotoxic effect observed upon treatment with HTR1 antagonists and DR antagonists was potentiated between 20-50% in co-treatment (Fig. 6B) in AML cell lines; thus, HTR1 and DR antagonists presented a synergistic anti-leukemic effect. Therefore, HTR1s and DRs seem to cooperate in the survival and/or proliferation of leukemic cells.

Discussion
The overexpression of HTR1 and DRs in MDS and CMML, described in this study, was also recently described in AML patients 23,24 . Regardless of disease subtype, these biomarkers arose as potential therapeutic targets in both disorders. Indeed, HTR1B antagonists induced cytotoxicity and differentiation of MDS and CMML cells; whereas DR antagonists showed an antiproliferative effect in CMML, in contrast to MDS cells. Interestingly, HTR1 antagonists displayed a synergistic effect with HMAs, which currently constitute the essential treatment for high-risk MDS patients; although due to technically difficulties, the synergist effect was studied in AML cells.
The in vitro activity observed with these agents warrant their translation to clinical trials, given the incurability of these diseases with currently approved therapies and, therefore, the urgent need of new therapeutic strategies for this challenging group of patients. HTR1B and DR3/5 receptors were consistently upregulated in a wide series of MDS patients diagnosed with different MDS subtypes, similar to previous description in AML patients 23,24 , and this finding was also observed among CMML specimens. Moreover, given the profound effect of HTR1B inhibition on cell viability and differentiation, the distinctive higher expression of these receptors in comparison to normal hematopoietic cells supports the key role of HTR1B signalling in the pathogenesis of these myeloid diseases. The lack of clinical correlation with HTR1B and/or DR expression across patient samples imply that these signaling pathways might not be involved in mechanisms governing disease progression and/or AML transformation.
The antineoplastic activity observed with different types of HTR antagonists identified HTR1B as the key target for further exploring for clinical purposes, either using a wide-spectrum agent as apomorphine or a more selective HTR1B antagonist. Of note, reduction of cell viability parallel induction of differentiation, supporting the rationale of differentiating therapy in MDS patients 41 .
Although AML, MDS and CMML are myeloid neoplasms that share common features such as an impaired terminal differentiation of mature cells, the intimate pathophysiology responsible and disease phenotype is remarkably diverse. AML is characterized by a marked block in differentiation coupled with an increased proliferation kinetics of leukemic progenitors that results in the accumulation of blasts in BM and peripheral blood. On the contrary, in MDS, the lack of terminal normal mature cells is mainly attributed to a failure of differentiation. CMML is a highly diverse entity, which combines dysplastic features resembling MDS with proliferative traits, especially evident in some patients with hyperleukocytosis. These discrepancies are probably due to differences in cell cycle status, cell survival and cell proliferation rate 42 . Accordingly, the degree of cell viability reduction of AML and CMML cells upon treatment with HTR1 and DR antagonists was similar. In contrast, MDS cells were more sensitive to HTR1 antagonists, as compared to DR antagonists, even though the expression level of these biomarkers was equivalent across samples. These results suggest that DRs may be implicated mainly in the block of differentiation observed in CMML and AML; whereas HTR1 may play a critical role in survival of myeloid-transformed blasts. Alternately, formation of HTR1/DR oligomers may constitute a fine-tuning signaling regulation that mediates leukemogenesis.
Due to the characteristic intraclonal and interpatient heterogeneity of myeloid neoplasms, monotherapy regimens are unlikely to be capable of a durable, clinically meaningful disease control. Therefore, combination therapies that may target the different neoplastic subpopulation are highly desirable. DR and HTR1 antagonists not only act synergically as anti-leukemic agents, but also positively interact with currently approved HMAs (azaciditine and decitabine), the current cornerstone in MDS treatment, fully justify the design of clinical trials exploring combination of these diverse drugs. Interestingly, the concentrations used are within the safety range, as previously published 23,24 . Moreover, dopamine and serotonin blockage are widely used among patients with  Parkinson's disease without remarkable side effects 43 , and this clinical experience could facilitate their repositioning as new agents for myeloid neoplasms 44 . Intriguingly, patients with Parkinson's disease have been reported to present a lower incidence of hematological cancers 45,46 ; although this epidemiological observation and putative casualty with generalized, long-standing use of DR and HTR antagonists should be prospectively addressed.
In conclusion, the differential expression of HTR1B and DRs in MDS and CMML cells, together with the observed effect on cell viability and differentiation induction upon HTR1B and DR inhibition in patient specimens identify these monoamine receptors as potential therapeutic targets in myeloid neoplasms. Moreover, the well proven clinical safety of DR and HTR antagonists and the synergic potential in combination with HMAs justify their rapid translation to clinical experimentation.

Methods
Primary samples. Primary MDS and CMML samples were obtained from patients diagnosed with MDS and CMML at Hospital Clínic of Barcelona (Spain) and Hospital Germans Trias i Pujol (Badalona, Spain). Alternatively, MDS and CMML samples were obtained from the Sample collection located at the Hematology Laboratory of IJC (ref#C0000886). MDS and CMML diagnosis and classification was based on standard WHO criteria 4,47 . Main MDS/CMML patient's characteristics are summarized in Tables 2 and S1. Samples were obtained from bone marrow and mononuclear cells (MNCs) were isolated by Ficoll density gradient centrifugation (GE). All patients provided written informed consent in accordance with the Declaration of Helsinki, and the study was approved by the Ethics Committee of Hospital Clínic of Barcelona and Hospital Germans Trias i Pujol; thus, all methods were performed in accordance with relevant guidelines and regulations. Healthy donor samples were obtained from the Banc de Sang i Teixits of Barcelona (Spain).

Drugs and antibodies.
All drugs were resuspended in H 2 O or DMSO (Sigma-Aldrich) according to manufacturer's instructions and were stored at −80 °C at 10 mM (Table S2). Antibody information is found in Table S3. Surface phenotype. Ficolled-primary samples were stained for DRD1, DRD2, DRD3, DRD4, DRD5, HTR1A or HTR1B on the surface, using the antibodies listed in Table S3, simultaneously with APC-conjugated anti-human CD45 (clone HI30) (BD Bioscience) and the live-dead discriminator dye 7-AAD (eBioscience). Cells were acquired in a FACSCanto II cytometer (BD). In patient samples, the analysis was performed within the live  blast gate (CD45 dim − SSC low/int ) 48,49 , whereas HD samples were gated based on CD45 pos based on the absent of a blast population, using FlowJo software (TriStar).
Cytotoxicity assay. 4 × 10 4 cells per mL were cultured in 96-well plates in complete medium and all drugs were added at the indicated concentration. 72 h after, cell viability was measured by 7-AAD (eBioscience) exclusion and Hoechst33342 (Sigma-Aldrich) positivity staining by flow cytometry, and cell count was obtained by volume in a FACSCanto II cytometer (BD). Statistical analysis was calculated in GraphPad (Prism software). FlowJo software (TriStar) was used for flow cytometry analysis.