1. Introduction

Thalidomide was developed in the late 1950s for the treatment of morning sickness during pregnancy. However, several years after its introduction, more than 10,000 babies were born with severe birth defects and the drug was withdrawal from the markets. Nevertheless, after years of interrogation, in 2006, thalidomide was approved by the FDA for the treatment of multiple myeloma (MM) and in the year 2010 a thalidomide binding protein, called cereblon was discovered. Despite this, the molecular pathways modulated by the IMiDs and the mechanism of the embryopathy are still unresolved. Recently, we elaborated a new theory to explain the teratogenic effects of thalidomide, based on: 1. The stereotypical sequence of limb defect accumulation along the severity scale, which partially overlaps the spatiotemporal differentiation sequence of the developing bone elements. 2. The coexistence of well-formed bone segments side by side to absent, distorted or incompletely shaped nearby segments, that many times remained unseparated from adjacent elements (e.g., radius from ulna; humerus from ulna) and the whole structure kept growing in size while keeping the shape and proportions fixed (e.g., with normal elongation of fused humero-ulnar elements or continues growth of residual radio-ulnar segment with age). 3. The dose independence but timing dependence of the malformation extent and distribution ^[1]. All these characters are consistent with a premature differentiation casualty of prospective bone elements with fixation of their temporary shape and proportions, while sparing the already patterned segments. Moreover, induction of differentiation or exit from a proliferating state can be also the explanation for the antimyeloma activities of the IMiDs, either via direct activity of the drug on the tumor cells proper or alternatively by changing the fate or function of their supportive stromal cells, endothelial cells or monocytes/ osteoclasts. Maturation can be also the explanation for the activity of lenalidomide in the 5q- dysplasia, which is associated with maturation arrest. To examine this theory further and search for differentiation markers induced in response to IMiD treatment, we began analyzing gene expression profile (GEP) changes in bone marrow samples from untreated patients with multiple myeloma (MM) before and after thalidomide/lenalidomide treatment in vitro and also followed the clinical response of these patients to lenalidomide. However, because the teratogenic and the antimyeloma activities of thalidomide have been attributed to drug metabolites which are generated in vivo after microsomal processing of the parent drug in a species specific manner ^[2], we finally decided to rely principally on the published dataset from the Seattle/Little Rock's study, which analyzed the GEP in paired bone marrow samples taken before and 48-hours after administration of thalidomide, lenalidomide or dexamethasone to MM patients ^[3]. In that study, newly diagnosed patients received dexamethasone (n = 45) or thalidomide (n = 42); in the case of relapsed MM, microarray data were obtained prior to (n = 36) and after (n = 19) lenalidomide administration. The analysis by the authors, which was based on false discovery rate (FDR) indicated that dexamethasone and thalidomide induce both common and unique GEP changes in tumor cells, with many of them being related to oxidative stress and cytoskeletal dynamics. Some of the changes were also predictive for the outcome of newly diagnosed MM patients receiving tandem transplants. Thalidomide-altered genes also changed following lenalidomide exposure and predicted event-free and overall survival in relapsed patients receiving lenalidomide as a single agent. To get further insight into the GEP changes of interest to our hypothesis, we extracted the entire raw data (which was available for 42 patients who were treated with dexamethasone, 40 patients in the thalidomide arm and 15 patients in the lenalidomide arm) and examined the average changes in GEP on each treatment arm. This analytical strategy was associated however with two major pitfalls. First, certain genes like DDIT4, IRF4, TNFSF10, DKK1 and FOS changed extremely in both directions, being among the top upregulated genes is some of the cases and among the top downregulated genes in others so the average changes missed the individual changes, possibly resulting from spontaneous fluctuations unrelated to the treatment or alternatively from a rebound phenomenon between drug doses (Supplementary Table 1). Second, an unacceptable bias was noticed in the case of disproportional change in the expression of any gene relative to the average change. For example SPP1 and CCND2 were induced >135 folds and >117 folds, respectively following thalidomide treatment of case #237, resulting in average fold changes (FC) of 27.2 and 18.8, respectively as compared to 2.66 and 2.51, respectively after exclusion of this case. To overcome this bias, we sorted the fold changes for each case separately and ranked them from 1 (most upregulated gene) to 12,625 (most downregulated gene). On this scale, the average position of SPP1 and CCND2 relative to all other genes were 4075 and 410, respectively whereas the most upregulated gene after either thalidomide or lenalidomide treatment was PLSCR1, which encodes for phospholipid scramblase (Figure 1 and Supplementary Table 2). PLSCR1 was also among the top 100 upregulated genes in 13 of the 40 (32.5%) thalidomide treated cases and in 11 of the 15 (73.3%) lenalidomide treated cases (Supplementary Table 1). Likewise, PLSCR1 was induced x2.55 folds in the bone marrow sample from our lenalidomide responsive patient, after treatment in vitro with lenalidomide (0.5 µg/ml) but not in the sample which was treated with thalidomide (0.8 µg/ml) neither in the samples from our two lenalidomide refractory cases which were treated in vitro with lenalidomide (Supplementary Table 3). As already mentioned, some of the activities of thalidomide depend on in vivo modification of the parent drug ^[2] and thus the significance of the findings when using this compound in vitro is unclear. Phospholipid scramblase 1 was originally identified as a type II transmembrane protein that mediates the calcium-dependent bidirectional movement of membrane phospholipids. It was also identified as a substrate for several kinases, including c-Abl, c-Src, and protein kinase Cδ (PKCδ). In addition, PLSCR1 plays potential roles in hematopoiesis and leukemogenesis. PLSCR1^−/− bone marrow cells exhibit defective myeloid proliferation and differentiation in response to stimulation by selected growth factors. Moreover, PLSCR1 RNA levels correlate with significantly longer overall survival in acute myeloid leukemia (AML), and PLSCR1 gene expression has been identified as a new prognostic factor in AML. All-trans retinoic acid (ATRA), an effective differentiation-inducing agent of acute promyelocytic leukemia (APL) cells, elevates PLSCR1 expression in ATRA- sensitive APL cells NB4 and HL60, but not in maturation-resistant NB4-LR1 cells. ATRA- and phorbol 12-myristate 13-acetate (PMA)-induced monocytic differentiation is accompanied by increased PLSCR1 expression, whereas only a slight or no elevation of PLSCR1 expression was observed in U937 cells differentiated with dimethyl sulfoxide (DMSO), sodium butyrate, or vitamin D3. Cell differentiation with ATRA and PMA, but not with vitamin D3 or DMSO, results in phosphorylation of PKCδ, and the PKCδ -specific inhibitor rottlerin nearly eliminates the ATRA- and PMA-induced expression of PLSCR1, while ectopic expression of a constitutively active form of PKCδ directly increases PLSCR1 expression. Finally, decreasing PLSCR1 expression with small interfering RNA inhibits ATRA/PMA-induced differentiation ^[4]. More recently it was shown that Wogonoside, the main flavonoid component derived from the root of the Chinese herb medicine Scutellaria baicalensis, can effectively inhibit the proliferation of several cancer cell lines and induce granulocytic differentiation and upregulation of PLSCR1 expression in AML cells. Wogonoside also promoted PLSCR1 trafficking into the nucleus and facilitated its binding to the inositol 1,4,5-trisphosphate receptor 1 (IP3R1) promoter, thus increasing the expression of IP3R1. Finally, inhibition of PLSCR1 expression with small interfering RNA partially blocked Wogonoside-induced cell cycle arrest and differentiation and disturbed the Wogonoside-associated molecular events ^{[5, 6]}. Similarly, after treatment of U937 cells with the flavonoid III-10 the cells differentiated into monocyte-like cells in association with upregulation of the differentiation-related proteins PLSCR1 and promyelocytic leukemia protein (PML) ^[7]. Moreover, III-10 stimulated PLSCR1 and PML probably through activation of PKCδ. In ovarian carcinoma cells, both IFN-2α and Aresnic trioxide modulate PLSCR1 mRNA levels. In turn, PLSCR1 modulates aspects of the Aresnic trioxide cellular response. Finally, PLSCR1 is also expressed in maturing chondrocytes and the latter cells show scramblase activity ^[8].

The second most upregulated gene at 48-hours after thalidomide treatment in the Seattle/Little Rock dataset, based on the average fold change position described, was the glutamate transporter ARL6IP5. This gene was also among the top 100 upregulated genes in 10 of the 40 thalidomide treated cases and in 6 of the 15 lenalidomide treated cases. ARL6IP5 (JWA) has been recently shown to be important for differentiation induced by some chemicals, including ATRA. In HeLa cells, the inhibition of proliferation and the induction of apoptosis by ATRA are dependent on ARL6IP5 expression ^[9]. ARL6IP5 is also induced by the synthetic retinoid 12-O-tetradecanoylphorbol 13-acetate, N-4-hydroxy-phenyl-retinamide and by arsenic trioxide. In HeLa and MCF-7 cells, treatment with arsenic trioxide produced apoptosis in a dose-dependent manner in parallel to increase in ARL6IP5 expression ^[10]. In the human myeloid leukemia HL-60 cells, concomitant with the progressive cell differentiation, ARL6IP5 expression was up-regulated by ATRA in a dose- and time-dependent manner while inhibition of ARL6IP5 expression by RNA interference partially blocked ATRA-induced differentiation and growth inhibition of HL-60 cells. Pre-treatment with PMA, a PKC activator, decreased ATRA-mediated differentiation, together with downregulation of ARL6IP5 expression. In addition, Arsenic trioxide enhanced the cellular differentiation induced by low dose ATRA concurrent with the enhancement of ARL6IP5 expression, suggesting that up-regulation of ARL6IP5 expression is essential for ATRA-induced differentiation of HL-60 cells ^[11].

Figure 1. The most upregulated genes at 48-h after treatment

Download as

PowerPoint Slide

Larger image(png format)

Figures index

Veiw figure

View current figure in a new window

View next figure

Other reproducibly induced genes included RAB13, which is highly upregulated during differentiation of human peripheral blood monocytic cells into osteoclasts ^[12] and MEF2C, which is essential for endochondral bone development. MEF2C, controls bone development by activating the gene program for chondrocyte hypertrophy ^[13]. Genetic deletion of MEF2C or expression of a dominant-negative MEF2C mutant in endochondral cartilage impairs hypertrophy, cartilage angiogenesis, ossification, and longitudinal bone growth in mice. Conversely, a super activating form of MEF2C causes precocious chondrocyte hypertrophy, ossification of growth plates, and dwarfism. Endochondral bone formation is exquisitely sensitive to the balance between MEF2C and the corepressor histone deacetylase 4 (HDAC4), such that bone deficiency of MEF2C mutant mice can be rescued by an HDAC4 mutation, and ectopic ossification in HDAC4 null mice can be diminished by a heterozygous MEF2C mutation. Interestingly, PLSCR1 and ARL6IP5 were upregulated in the Seattle/Little Rock series by either thalidomide or lenalidomide but not dexamethasone, whereas MEF2C was induced by thalidomide only.

Importantly, when the analysis of the thalidomide arm was restricted to the subset of 13 cases whose PLSCR1 position was < 100 (the average PLSCR1 induction of this group was 3.81 folds as compared to 2.15 folds in the thalidomide arm as a whole), the majority of the top upregulated genes from the thalidomide arm remained so and continued to overlap the top upregulated genes from the lenalidomide arm (in which arm PLSCR1 was among the top 100 upregulated genes in 73.3% of cases). In contrast, when the analysis of the thalidomide arm was restricted to the 7 cases whose PLSCR1 position was >5000 (in the scale from 1 to 12,625) (the average fold change for PLSCR1 in this subset of patients was 0.86), the highly modulated gene list changed completely and now it had almost nothing in common with the induction signature characterized the lenalidomide and thalidomide arms including the subset of thalidomide treated patients whose PLSCR1 position was < 100. Therefore, it is obvious that PLSCR1 is part of a cluster of genes whose expression is strongly associated and context specific. The specificity of the PLSCR1 associated signature is also reflected by the absence of this signature in the dexamethasone arm (Figure 2). However, even when PLSCR1 induction was maximal, individual PLSCR1 associated genes were not always concordant and many times some of them did not change at all or even changed in the opposite direction. The explanation for this observation can be that many of the changes induced by the IMiDs are sequential and/or transient and therefore not necessarily exist simultaneously. The analysis of the 13 samples with maximal PLSCR1 induction separately exhibited also the association of PLSCR1 with additional relevant genes. For example, ID2/ID2B progressed from average position 30 in the thalidomide arm (and average fold change induction of 2.09) to position 2 (and average fold change induction of 3.58) in the subset of thalidomide treated patients whose PLSCR1 position was < 100. ID1 and ID2 are retinoic acid responsive genes ^[14] and ID2 controls chondrogenesis during maxillary morphogenesis ^[15]. Likewise, the position of IGF1, which is another retinoic acid responsive gene ^[16], progressed from 808 to 21, and that of SP100, which is part of the PML nuclear body ^[17] from position 91 to 25. Moreover, both SP100 and SP110 (another PML nuclear body transcript) were induced after lenalidomide treatment in vitro in our lenalidomide responsive case (x2.21 folds and 3.74 folds, respectively). The highly induced gene list also included FGL2, which encodes for prothrombinase ^[18]. The contribution of FGL2 to the hypercoagulable state associated with IMiD administration seeks further investigation.

Figure 2. PLSCR1 associated signature

Download as

PowerPoint Slide

Larger image(png format)

Figures index

Veiw figure

View current figure in a new window

View previous figure

CD27 was the most downregulated gene in the thalidomide arm (position 12,625, fold change 0.84) and nearly so in the lenalidomide arm (position 12,622, fold change 0.60). In contrast, in the dexamethasone arm CD27 was upregulated (position 82, average fold change induction = 1.50). Expression of CD27 was associated with increased clonogenic capacity and engraftment of human myeloma cells in immunodeficient nonobese diabetes/severe combined immunodeficient (NOD/SCID) mice during both primary and secondary transplantations ^[19]. However, the most downregulated gene in both the lenalidomide arm and the thalidomide subset of patients whose PLSCR1 position was < 100 was AMPD1. This gene encodes for adenylate deaminase, which is mutated in various muscular abnormalities. AMPD1 was also down regulated after treatment of OCI/AML2 cells with ATRA, with decrease in the signal intensity from 69.7 in the control cells to 44.6 after ATRA and to 31.9 after ATRA + Valproic acid treatment ^[20]. In the same dataset (GDS1215), ATRA induced the expression of PLSCR1, ARL6IP5 and ITGB7, similar to their induction in the lenalidomide arm and in the thalidomide subset of patients whose PLSCR1 position was < 100 in the Seattle/little Rock dataset. Finally, the most specific dexamethasone modulated gene in the Seattle/Little Rock series was the Prolyl hydroxylase P4HA, which was among the top 100 induced genes in 23 of 42 dexamethasone treated cases and its induction was limited to dexamethasone.

Although our original data could add only little to the authentic ("in vivo" quality) information existing in the large dataset from Seattle/Little Rock, the whole genome platform we used can still add to the whole picture. In principle, the in vitro GEP changes evolved in the bone marrow sample from our lenalidomide responsive patient at 24-hours after incubation of the cells with either thalidomide or lenalidomide could be divided into three main groups: 1. Non protein coding genes. 2. Interferon inducible genes. 3. Other genes. In contrast, most of the above changes were absent in the samples from the two lenalidomide resistant cases. The upregulated non protein coding genes included small nucleolar RNA (SNORD), small Cajal body-specific RNA (SCARNA), micro RNAs and histone modifiers (supplementary Table 3). These genes regulate the expression at the transcriptional, translational and posttranslational levels epigenetically. SNORD61 was the mostly induced gene after thalidomide treatment in vitro (x27.3 folds) and it was also induced (x2.97 folds) after lenalidomide treatment. Other upregulated genes included SNORD115-42 (induced 7.2 folds after thalidomide and 6.2 folds after lenalidomide treatment) and the histone modifier HIST2H2AC (3.78 folds induction following thalidomide treatment). SNORD115 and SNORD116 gene clusters believed to play key roles in the fine-tuning of serotonin receptor (5-HT2C) pre-mRNA processing and in the etiology of the Prader-Willi Syndrome, respectively ^[21]. Among the interferon inducible genes were IFIT1, CXCL10, IFI44L, XAF1, FCGR2B, LCE1B, S100A10, STAT1 and TNFSF10.

In summary, our analysis gives support to the hypothesis that both the teratogenic and the antitumoral effects of the IMiDs are mediated via differentiation induction, which involves molecular pathways regulated by ATRA, Arsenic trioxide and other differentiating agents, possibly with involvement of epigenetic changes. Maturation induction also plays role in the activity thalidomide in hereditary vascular malformations ^[22]. The identified PLSCR1 associated signature may help to predict the clinical response to the IMiDs and to discriminate between IMiD responsive versus IMiD refractory tumor cells. In this regard our prediction is that the 13 Seattle/Little Rock patients whose PLSCR1 position was >5000 were less responsive to thalidomide and if so it will be possible to personalize the treatment. Our analysis also gives the rational for combining the IMiDs with classical differentiating agents in both IMiD responsive tumors (e.g., 5q- dysplasia, MM) and ATRA/ Arsenic trioxide responsive tumors (e.g., APL). Finally, premature differentiation (and the PLSCR1 associated signature) may prove to be a common pathway for the teratogenic effects of thalidomide, retinoids, valproic acid and related agents.

2. Methods

The bone marrow samples we used were cultured and analyzed by microarray as described elsewhere ^{[23, 24]}. The study was approved by Asaf Harofhe Institute Review Board, Zerifin, Israel and patients provided written informed consent.

Acknowledgments

The authors thank Professor Dov Zipory for the training of the first author in the Weitzmann Institute of Science, Dr Izhar Hardan from Meir Medical Center for his kind technical support, Dr Relly Forer, Dr Inna Vulin and Dina Volodarsky from Dyn diagnostics Ltd for professional microarray and Bioinformatic service.

Author Contribution

YC designed the experiments, processed the samples, analyzed the data and wrote the article. All other authors assisted in patient recruitment.

Conflict of Interest Disclosure

All authors declare no conflict of interest

References

[1]	Cohen Y. The mechanism of action of novel antimyeloma agents and newer therapeutic strategies. In: Steve Holt Multiple Myeloma: Risk Factors, Diagnosis and Treatment. Nova Publishers Ltd: USA 2014 (in press).
	In article
[2]	Yaccoby S, Johnson CL, Mahaffey SC, Wezeman MJ, Barlogie B, et al (2002) Antimyeloma efficacy of thalidomide in the SCID-hu model. Blood 100: 4162-4168.
	In article		CrossRef PubMed
[3]	Burington B, Barlogie B, Zhan F, Crowley J, Shaughnessy JD Jr (2008) Tumor cell gene expression changes following short-term in vivo exposure to single agent chemotherapeutics are related to survival in multiple myeloma. Clin Cancer Res 14: 4821-4829.
	In article		CrossRef PubMed
[4]	Zhao KW, Li X, Zhao Q, Huang Y, Li D, et al (2004) Protein kinase Cdelta mediates retinoic acid and phorbol myristate acetate-induced phospholipid scramblase 1 gene expression: its role in leukemic cell differentiation. Blood 104: 3731-3738.
	In article		CrossRef PubMed
[5]	Chen Y, Hui H, Yang H, Zhao K, Qin Y, et al (2013) Wogonoside induces cell cycle arrest and differentiation by affecting expression and subcellular localization of PLSCR1 in AML cells. Blood 121: 3682-3691.
	In article		CrossRef PubMed
[6]	Zhang K, Guo QL, You QD, Yang Y, Zhang HW, et al (2008) Wogonin induces the granulocytic differentiation of human NB4 promyelocytic leukemia cells and up-regulates phospholipid scramblase 1 gene expression. Cancer Sci 99: 689-695.
	In article		CrossRef PubMed
[7]	Qin Y1, Li Z, Chen Y, Hui H, Sun Y, et al (2012) III-10, a newly synthesized flavonoid, induced differentiation of human U937 leukemia cells via PKCδ activation. Eur J Pharm Sci 45: 648-656.
	In article		CrossRef PubMed
[8]	Damek-Poprawa M, Golub E, Otis L, Harrison G, Phillips C, et al (2006) Chondrocytes utilize a cholesterol-dependent lipid translocator to externalize phosphatidylserine. Biochemistry 45: 3325-3336.
	In article		CrossRef PubMed
[9]	Mao WG, Liu ZL, Chen R, Li AP, Zhou JW (2006) JWA is required for the antiproliferative and pro-apoptotic effects of all-trans retinoic acid in Hela cells. Clin Exp Pharmacol Physiol 33: 816-824.
	In article		CrossRef PubMed
[10]	Zhou J, Ye J, Zhao X, Li A, Zhou J (2008) JWA is required for arsenic trioxide induced apoptosis in HeLa and MCF-7 cells via reactive oxygen species and mitochondria linked signal pathway. Toxicol Appl Pharmacol 230: 33-40.
	In article		CrossRef PubMed
[11]	Huang S, Shen Q, Mao WG, Li AP, Ye J, et al (2006) JWA, a novel signaling molecule, involved in all-trans retinoic acid induced differentiation of HL-60 cells. J Biomed Sci 13: 357-371.
	In article		CrossRef PubMed
[12]	Hirvonen MJ, Mulari MT, Büki KG, Vihko P, Härkönen PL, et al (2012) Rab13 is upregulated during osteoclast differentiation and associates with small vesicles revealing polarized distribution in resorbing cells. J Histochem Cytochem 60: 537-549.
	In article		CrossRef PubMed
[13]	Arnold MA, Kim Y, Czubryt MP, Phan D, McAnally J, et al (2007) MEF2C transcription factor controls chondrocyte hypertrophy and bone development. Dev. Cell 12: 377-389.
	In article		CrossRef PubMed
[14]	Nigten J, Breems-de Ridder MC, Erpelinck-Verschueren CA, Nikoloski G, van der Reijden BA, et al (2005) ID1 and ID2 are retinoic acid responsive genes and induce a G0/G1 accumulation in acute promyelocytic leukemia cells. Leukemia 19: 799-805.
	In article		CrossRef PubMed
[15]	Sakata-Goto T, Takahashi K, Kiso H, Huang B, Tsukamoto H, et al (2012) Id2 controls chondrogenesis acting downstream of BMP signaling during maxillary morphogenesis. Bone 50: 69-78.
	In article		CrossRef PubMed
[16]	Shim JH, Shin DW, Lee TR, Kang HH, Jin SH, et al (2012) The retinoic acid-induced up-regulation of insulin-like growth factor 1 and 2 is associated with prolidase-dependent collagen synthesis in UVA-irradiated human dermal equivalents. J. Dermatol. Sci 66: 51-59.
	In article		CrossRef PubMed
[17]	Weidtkamp-Peters S, Lenser T, Negorev D, Gerstner N, Hofmann TG, et al (2008) Dynamics of component exchange at PML nuclear bodies. J Cell Sci 121: 2731-2743.
	In article		CrossRef PubMed
[18]	Yuwaraj S, Ding J, Liu M, Marsden PA, Levy, G. A (2001) Genomic characterization, localization, and functional expression of FGL2, the human gene encoding fibroleukin: a novel human procoagulant. Genomics 71: 330-338.
	In article		CrossRef PubMed
[19]	Matsui W, Wang Q, Barber JP, Brennan S, Smith BD, et al (2008) Clonogenic multiple myeloma progenitors, stem cell properties, and drug resistance. Cancer Res 58: 190-197.
	In article		CrossRef PubMed
[20]	Trus MR, Yang L, Suarez Saiz F, Bordeleau L, Jurisica I, et al (2005) The histone deacetylase inhibitor valproic acid alters sensitivity towards all trans retinoic acid in acute myeloblastic leukemia cells. Leukemia 19: 1161-1168.
	In article		CrossRef PubMed
[21]	Bortolin-Cavaillé ML, Cavaillé J (2012) The SNORD115 (H/MBII-52) and SNORD116 (H/MBII-85) gene clusters at the imprinted Prader-Willi locus generate canonical box C/D snoRNAs. Nucleic Acids Res 40: 6800-6807.
	In article		CrossRef PubMed
[22]	Lebrin F, Srun S, Raymond K, Martin S, van den Brink S, et al (2010) Thalidomide stimulates vessel maturation and reduces epistaxis in individuals with hereditary hemorrhagic telangiectasia. Nat. Med 16: 420-428.
	In article		CrossRef PubMed
[23]	Cohen Y, Garach-Jehoshua O, Bar-Chaim A, Kornberg A (2012) Niche-modulated and niche-modulating genes in bone marrow cells. Blood Cancer J 2: e97.
	In article		CrossRef PubMed
[24]	Cohen Y, Gutwein O, Garach-Jehoshua O, Bar-Haim A, Kornberg A (2014) The proliferation arrest of primary tumor cells out-of-niche is associated with widespread downregulation of mitotic and transcriptional genes. Hematology 19: 286-292.
	In article		CrossRef PubMed

Supplementary Information

The top 100 upregulated (red) and 100 downregulated (green) genes (of overall 12,625 analyzed genes) in the Seattle/Little Rock dataset at 48-h following treatment of MM patients with dexamethasone (Dex), lenalidomide (Len) or thalidomide (Dex). In the red and green columns appearing the number of cases who exhibited the radical gene modulations (of overall 42 cases who received Dex, 15 who received Len and 40 who received Thal). Note that many of these genes, like STAT1 were both upregulated and downregulated extremely in different cases, possibly due to a rebound phenomenon between doses.

Fold changes (FC) of each case from the Seattle/Little Rock dataset were sorted and ranked in a decreasing order from 1 (most upregulated gene) to 12,625 (most downregulated gene). The position of any gene in this scale was used to calculate the average FC position in each treatment arm and the 500 most upregulated and 500 most downregulated genes in the thalidomide arm (column oo) are presented and compared to the average position of the same genes in the other treatment arms. Thal, thalidomide; Len, lenalidomide; Dex, dexamethasone.

Bone marrow samples from a patient who responded to lenalidomide clinically and two patients who did not respond were cultured for 24-h with or without thalidomide or lenalidomide. The list includes the upregulated genes for which the fold changes (FC) were > 2 and the downregulated genes for which the FC were < 0.5. Note the marked induction of non-protein coding genes using thalidomide and induction of interferon inducible genes using either compound.