{\it RAS\/} is required for the expression of interleukin-1$\beta$ in two diverse leukemic cell lines

Under normal physiological conditions, cells of the hematopoietic system produce Interleukin-1$\beta$(IL-1$\beta)$ only when a stimulus is present. Leukemic cells, however, can constitutively produce this cytokine without an exogenous source of activation. In addition, IL-1$\beta$ can operate as an autocrine and/or paracrine growth factor for leukemic blasts. In order to study the cellular basis for this aberrant production, we analyzed two leukemic cell lines (B1 and W1) which express high levels of IL-1$\beta$ and use IL-1$\beta$ as an autocrine growth factor. Initial studies demonstrated: (1) lack of rearrangement and/or amplification in the IL-1$\beta$ gene and its promoter; and (2) intact responsiveness to regulators such as cycloheximide and dexamethasone, implying that the molecular defect was upstream. Analysis of the Ras inducible transcription factors by gel shift assay demonstrated constitutive transcription factor binding in the IL-1$\beta$ promoter. Furthermore, RAS mutations were found at codon 12 in the K-RAS and N-RAS genes in the B1 and W1 cells, respectively. To deduce the effects of activated Ras on IL-1$\beta$ expression, two classes of farnesyltransferase inhibitors and an adenoviral vector expressing antisense targeted to K-RAS were utilized. The farnesyltransferase inhibitors perillyl alcohol and B581 were able to reduce IL-1$\beta$ levels by 80% and 50% in the B1 cells, respectively. In W1 cells, IL-1$\beta$ was reduced by 60% with 1mM perillyl alcohol. Antisense RNA targeted to K-RAS confirmed the results demonstrating a 50% reduction in IL-1$\beta$ expression in the B1 cells. In addition, decreased binding at the crucial NF-IL6/CREB binding site correlated with decreased IL-1$\beta$ production and cellular proliferation implying that this site was a downstream effector of Ras signaling. Our data suggest that mutated RAS genes may be responsible for autocrine IL-1$\beta$ production in some leukemias by stimulating signal transduction pathways that activate the IL-1$\beta$ promoter. ^

Analysis of the mechanisms that maintain coordinate production of $\alpha$- and $\beta$-tubulin in mammalian cells

Alpha and beta tubulin are essential proteins in all eukaryotic cells. To study how cells maintain coordinate levels of these two interacting proteins, we have used PCR to add a 9 amino acid epitope from influenza hemagglutinin protein onto the carboxyl terminus of $\alpha$1 and $\beta$1-tubulin. The chimeric tubulin genes (HA$\alpha$1 and HA$\beta$1) were transfected into CHO cells and cell lines that stably express each gene were selected. Cells transfected with HA-tubulin do not exhibit any gross changes in growth or morphology. Immunofluorescence analysis demonstrated that HA-tubulins incorporate into both cytoplasmic and spindle microtubules. A quantitative biochemical assay was used to show that HA-tubulins incorporate into microtubules to a normal extent and do not alter the steady state distribution of endogenous tubulin between monomer and polymer pools. Two-dimensional gel analysis of pulse-labeled cells indicated that when HA$\beta$1-tubulin is expressed at high levels, it slightly represses the synthesis of the endogenous $\beta$-tubulin but produces a small increase in the synthesis of $\alpha$-tubulin. Analysis of cells labeled to steady state showed that HA$\beta$1-tubulin accumulates to a similar level as the wild-type gene product, but together these polypeptides produce only a small increase in total tubulin content consistent with the increased synthesis of $\alpha$-tubulin. It thus appears that HA$\beta$1-tubulin successfully competes with endogenous $\beta$-tubulin for heterodimer formation and that free $\beta$-tubulin subunits (endogenous and HA$\beta$1) are selectively degraded to maintain coordinate amounts of $\alpha$- and $\beta$-tubulin. In addition, the increased synthesis of $\alpha$-tubulin suggested the existence of a mechanism to ensure coordinate synthesis of $\alpha$- and $\beta$-tubulin subunits. To analyze whether reciprocal changes in endogenous tubulin synthesis occur when $\alpha$-tubulin is overexpressed, stably transfected CHO cell lines were isolated in which HA$\alpha$1-tubulin represents 50% of the total $\alpha$-tubulin, and its relative abundance can be further increased to 85-90% by treatment with sodium butyrate. In contrast with results obtained using HA$\beta$1-tubulin, transfection of HA$\alpha$1-tubulin decreased the synthesis of endogenous $\alpha$-tubulin to 60% of normal with little or no change in $\beta$-tubulin synthesis. When the transfected cells were treated with sodium butyrate to further increase HA$\beta$1-tubulin production, a larger decrease in the synthesis of endogenous $\alpha$-tubulin (to 30% of normal) was observed. The repression on the synthesis of endogenous $\alpha$-tubulin polypeptide was found to be directly proportional to the expression of HA$\alpha$1-tubulin indicating the existence of an autoregulatory loop, where $\alpha$-tubulin inhibits its own synthesis. To determine whether overproduction of HA$\alpha$1-tubulin affected the transcription, message stability or translation of endogenous $\alpha$-tubulin, the steady state levels of $\alpha$-tubulin mRNA were analyzed by ribonuclease protection assays. The results showed that the steady state level of $\alpha$-tubulin mRNA is not affected by the overexpression of HA$\alpha$1-tubulin, indicating that the repression is translational. The results are compatible with a model in which $\beta$-tubulin synthesis is largely unperturbed by overexpression of other tubulin subunits, and excess $\beta$-tubulin subunits are rapidly degraded to maintain coordinate $\alpha$- and $\beta$-tubulin levels at steady state. In contrast, free $\alpha$-tubulin represses its own synthesis at the translational level, suggesting that its level of production may be controlled by the amount of $\beta$-tubulin available for heterodimer formation. ^

Accessing novel developmental mechanisms in the mouse by gene trapping

Genetic analysis is a powerful method for analyzing the function of specific genes in development. I sought to identify novel genes in the mouse using a genetic analysis relying on the expression pattern and phenotype of mutated genes. To this end, I have conducted a gene trap screen using the vector $\rm SA\beta geo,$ a promoterless DNA construct that encodes a fusion protein with lacZ and neomycin resistance activities. Productive integration and expression of the $\beta$geo protein in embryonic stem (ES) cells requires integration into an active transcription unit. The endogenous regulatory elements direct reporter gene expression which reflects the expression of the endogenous gene. Of eight mouse lines generated from gene trap ES cell clones, four showed differential regulation of $\beta$geo activity during embryogenesis. These four were analyzed in more detail.^ Three of the lines RNA 1, RNA2 and RNA 3 had similar expression patterns, within subsets of cells in sites of embryonic hematopoiesis. Cloning of the trapped genes revealed that all three integrations had occurred within 45S rRNA precursor transcription units. These results imply that there exists in these cells some mechanism responsible for the efficient production of the $\beta$geo protein from an RNA polymerase I transcript that is not present in most of the cells in the embryo.^ The fourth line, GT-2, showed widespread, dynamic expression. Many of the sites of expression were important classic embryonic induction systems. Cloning of the sequences fused to the $5\sp\prime$ end of the $\beta$geo sequence revealed that the trapped gene contained significant sequence homology with a previously identified human sequence HumORF5. An open reading frame of this sequence is homologous to a group of eukaryotic proteins that are members of the RNA helicase superfamily I.^ Analysis of the gene trap lines suggests that potentially novel developmental mechanisms have been uncovered. In the case of RNA 1, 2 and 3, the differential production of ribosomal RNAs may be required for differentiation or function of the $\beta$geo positive hematopoietic cells. In the GT-2 line, a previously unsuspected temporal and spatial regulation of a putative RNA helicase implies a role for this activity during specific aspects of mouse development. ^

Regulation of angiogenesis by interferon-beta

The purpose of these studies was to investigate the role of interferon-beta (IFN-$\beta$) in angiogenesis. IFN-$\alpha/\beta$ have been implicated in inhibiting a number of steps in the angiogenic pathway. We examined the balance of angiogenesis-regulating molecules in several systems including human infantile hemangiomas, UV-B irradiated mice, and dorsal incisional wound healing in mice. In each system, epidermal hyperplasia and cutaneous angiogenesis were directly related to the expression of positive angiogenic factors (bFGF and VEGF) and inversely related to the expression of endogenous IFN-$\beta.$ The re-expression of IFN-$\beta$ correlated with tumor regression and/or resolution of wound healing. In contrast to control mice, UV-B-induced cutaneous angiogenesis and hyperplasia persisted in IFN-$\alpha/\beta$ receptor knock-out mice. In normal mice, endogenous IFN-$\beta$ was expressed by all differentiated epithelial cells exposed to environmental stimuli. The expression of endogenous IFN-$\beta$ was necessary but insufficient for complete differentiation of epidermal keratinocytes.^ The tumor organ microenvironment can regulate angiogenesis. Human bladder carcinoma cells growing in the bladder wall of nude mice express high levels of bFGF, VEGF, and MMP-9, have higher vascular densities, and produce metastases to lymph nodes and lungs, whereas the same cells growing subcutaneously express less bFGF, VEGF, and MMP-9, have lower vascular densities, and do not metastasize. IFN-$\alpha/\beta$ was found to inhibit bFGF and MMP-9 expression both in vitro and in vivo in human bladder carcinoma cells. Systemic therapy with human IFN-$\alpha$ of human bladder cancer cells growing orthotopically in nude mice, resulted in decreased vascularity, tumorigenicity, and metastasis as compared to saline treated mice. Human bladder cancer cells resistant to the antiproliferative effects of IFN were transfected with the human IFN-$\beta$ gene. Hu-IFN-$\beta$ transfected cells expressed significantly less bFGF protein and gelatinase activity than parental or control-transfected cells and did not grow at ectopic or orthotopic sites. Collectively the data provide direct evidence that IFN-$\alpha/\beta$ can inhibit angiogenesis via down-regulation of angiogenesis-stimulating cytokines. ^

beta-catenin interacts with and inhibits NF-kappaB in colon cancer

The β-catenin pathway plays an important role in the progression of colon cancer as well as many other cancer types. Almost all colorectal tumors show an upregulation of β-catenin activity either through mutations in the β-catenin regulator APC or through mutations in β-catenin itself. Upregulation of β-catenin leads to the transcription of many target genes involved in tumorigenesis. NF-κB is a transcription factor which activates many target genes, including both anti-apoptotic and pro-apoptotic molecules. Recently, it has been shown that GSK-3β, a negative regulator of β-catenin, is involved in the activation of NF-κB. However, the mechanism of this regulation of NF-κB by GSK-3β is unclear. As GSK-3β inhibits β-catenin we hypothesized that β-catenin may be responsible for the regulation of NF-κB by GSK-3β; i.e. β-catenin may inhibit NF-κB activity. In this study we show that β-catenin physically interacts with NF-κB leading to the inhibition of NF-κB transcriptional and DNA-binding activities. We also show that in colon cancer cells with high β-catenin expression there is a suppressed NF-κB activity and depletion of β-catenin increases NF-κB activity. Similarly, in colon cancer cells that have a low level of β-catenin NF-κB activity is high and introduction of β-catenin reduces NF-κB activity. Importantly, we show that this suppression of NF-κB by β-catenin leads to a reduction of NF-κB target gene Fas expression. Also Fas-mediated apoptosis is reduced in β-catenin overexpressing cells, which can be reversed upon depletion of β-catenin. Introduction of the NF-κB subunit p65 can restore Fas expression indicating that the effect of β-catenin on Fas is through NF-κB. Furthermore, β-catenin expression was found to inversely correlate with Fas expression in human colon and breast primary tumor tissues. As Fas downregulation is important for tumors to evade immune surveillance, β-catenin inhibition of NF-κB and Fas downregulation likely plays and important role for colon cancer progression. Additionally, we found that phosphoinositide 3-kinase plays a role in the regulation of β-catenin inhibition of NF-κB through the disruption of the β-catenin/NF-κB complex. This study provides a link between two important signal transduction pathways as well as another mechanism of β-catenin oncogenesis. ^

The role of beta-arrestin 2 in lysophosphatidic acid-induced NF-kappaB activation

Lysophosphatidic acid (LPA) is a bioactive phospholipid and binds to its receptors, a family of G protein-coupled receptors (GPCR), which initiates multiple signaling cascades and leads to activation of several transcription factors, including NF-κB. NF-κB critically regulates numerous gene expressions, and is persistently active in many diseases. In our previous studies, we have demonstrated that LPA-induced NF-κB activation is dependent on a novel scaffold protein, CARMA3. However, how CARMA3 is recruited to receptor remains unknown. β-Arrestins are a family of proteins involved in desensitization of GPCR signaling. Additionally, β-arrestins function as signaling adaptor proteins, and mediate multiple signaling pathways. Therefore, we have hypothesized that β-arrestins may link CARMA3 to LPA receptors, and facilitate LPA-induced NF-κB activation. ^ Using β-arrestin-deficient MEFs, we found that β-arrestin 2, but not β-arrestin 1, was required for LPA-induced NF-κB activation. Also, we showed that the expression of NF-κB-dependent cytokines, such as interlukin-6, was impaired in β-arrestin 2-deficient MEFs. Mechanistically, we demonstrated the inducible association of endogenous β-arrestin 2 and CARMA3, and we found the CARD domain of CARMA3 interacted with 60-320 residues of β-arrestin 2. To understand why β-arrestin 2, but not β-arrestin 1, mediated NF-κB activation, we generated β-arrestin mutants. However, some mutants degraded quickly, and the rest did not rescue NF-κB activation in β-arrestin-deficient MEFs, though they had similar binding affinities with CARMA3. Therefore, it indicates that slight changes in residues may determine the different functions of β-arrestins. Moreover, we found β-arrestin 2 deficiency impaired LPA-induced IKK kinase activity, while it did not affect LPA-induced IKKα/β phosphorylation. ^ In summary, our results provide the genetic evidence that β-arrestin 2 serves as a positive regulator in NF-κB signaling pathway by connecting CARMA3 to LPA receptors. Additionally, we demonstrate that β-arrestin 2 is required for IKKα/β activation, but not for the inducible phosphorylation of IKKα/β. Because the signaling pathways around the membrane-proximal region of LPA receptors and GPCRs are quite conserved, our results also suggest a possible link between other GPCRs and CARMA3-mediated NF-κB activation. To fully define the role of β-arrestins in LPA-induced NF-κB signaling pathways will help to identify new drug targets for clinical therapeutics.^

Single nucleotide polymorphisms (SNPs) associated with TGF-beta pathway and their significance in systemic sclerosis - A multilevel analysis

Systemic sclerosis (SSc) or Scleroderma is a complex disease and its etiopathogenesis remains unelucidated. Fibrosis in multiple organs is a key feature of SSc and studies have shown that transforming growth factor-β (TGF-β) pathway has a crucial role in fibrotic responses. For a complex disease such as SSc, expression quantitative trait loci (eQTL) analysis is a powerful tool for identifying genetic variations that affect expression of genes involved in this disease. In this study, a multilevel model is described to perform a multivariate eQTL for identifying genetic variation (SNPs) specifically associated with the expression of three members of TGF-β pathway, CTGF, SPARC and COL3A1. The uniqueness of this model is that all three genes were included in one model, rather than one gene being examined at a time. A protein might contribute to multiple pathways and this approach allows the identification of important genetic variations linked to multiple genes belonging to the same pathway. In this study, 29 SNPs were identified and 16 of them located in known genes. Exploring the roles of these genes in TGF-β regulation will help elucidate the etiology of SSc, which will in turn help to better manage this complex disease. ^

RNA-SEQUENCING APPLICATIONS: GENE EXPRESSION QUANTIFICATION AND METHYLATOR PHENOTYPE IDENTIFICATION

My dissertation focuses on two aspects of RNA sequencing technology. The first is the methodology for modeling the overdispersion inherent in RNA-seq data for differential expression analysis. This aspect is addressed in three sections. The second aspect is the application of RNA-seq data to identify the CpG island methylator phenotype (CIMP) by integrating datasets of mRNA expression level and DNA methylation status. Section 1: The cost of DNA sequencing has reduced dramatically in the past decade. Consequently, genomic research increasingly depends on sequencing technology. However it remains elusive how the sequencing capacity influences the accuracy of mRNA expression measurement. We observe that accuracy improves along with the increasing sequencing depth. To model the overdispersion, we use the beta-binomial distribution with a new parameter indicating the dependency between overdispersion and sequencing depth. Our modified beta-binomial model performs better than the binomial or the pure beta-binomial model with a lower false discovery rate. Section 2: Although a number of methods have been proposed in order to accurately analyze differential RNA expression on the gene level, modeling on the base pair level is required. Here, we find that the overdispersion rate decreases as the sequencing depth increases on the base pair level. Also, we propose four models and compare them with each other. As expected, our beta binomial model with a dynamic overdispersion rate is shown to be superior. Section 3: We investigate biases in RNA-seq by exploring the measurement of the external control, spike-in RNA. This study is based on two datasets with spike-in controls obtained from a recent study. We observe an undiscovered bias in the measurement of the spike-in transcripts that arises from the influence of the sample transcripts in RNA-seq. Also, we find that this influence is related to the local sequence of the random hexamer that is used in priming. We suggest a model of the inequality between samples and to correct this type of bias. Section 4: The expression of a gene can be turned off when its promoter is highly methylated. Several studies have reported that a clear threshold effect exists in gene silencing that is mediated by DNA methylation. It is reasonable to assume the thresholds are specific for each gene. It is also intriguing to investigate genes that are largely controlled by DNA methylation. These genes are called “L-shaped” genes. We develop a method to determine the DNA methylation threshold and identify a new CIMP of BRCA. In conclusion, we provide a detailed understanding of the relationship between the overdispersion rate and sequencing depth. And we reveal a new bias in RNA-seq and provide a detailed understanding of the relationship between this new bias and the local sequence. Also we develop a powerful method to dichotomize methylation status and consequently we identify a new CIMP of breast cancer with a distinct classification of molecular characteristics and clinical features.

Regulation of myogenesis and suppression of {\it mck\/} 5$\sp\prime$ enhancer elements by TGF$\beta$ and RAS

The regulation of muscle differentiation, like cell differentiation in general, is only now beginning to be understood. Here are described several key features to myogenesis: a beginning, some intermediary events, and an endpoint. Muscle differentiation proceeds spontaneously when myoblasts are cultured in serum-poor medium. Transforming growth factor type $\beta$ (TGF$\beta$), a component of fetal serum, was found to potently suppress muscle differentiation. Prolonged blockade of differentiation required replenishing TGF$\beta$. When TGF$\beta$ was removed, cells rapidly differentiated. Both TGF$\beta$ and RAS, which also blocks myogenesis, suppress the genes for a series of muscle-specific proteins. Regions that regulate transcription of one such gene, muscle creatine kinase (mck), were located by linking progressively smaller parts of the mck 5$\sp\prime$ region to the marker gene cat and testing the constructs for regulated expression of cat in myoblasts and muscle cells. The mck promoter is not muscle-specific but requires activation. Two enhancers were found: a weak, developmentally regulated enhancer within the first intron, and a strong, compact, and tightly developmentally regulated enhancer about 1.2 Kb upstream of the transcription start site. Activity of this enhancer is eliminated by activated ras. Suppression of activated N-RAS restores potency to the upstream enhancer. Further deletion shows the mck 5$\sp\prime$ enhancer to contain an enhancer core with low but significant muscle-specific activity, and at least one peripheral element that augments core activity. The core and this peripheral element were comprised almost entirely of factor-binding motifs. The peripheral element was inactive as a single copy, but was constitutively active in multiple copies. Regions flanking the peripheral element augmented its activity and conferred partial muscle-specificity. The enhancer core is also modulated by its 5$\sp\prime$ flanking region in a complex manner. Site-specific mutants covering most of the enhancer core and interesting flanking sequences have been made; all mutants tested diminish the activity of the 5$\sp\prime$ enhancer. Alteration of the site to which MyoD1 is reported to bind completely inactivates the enhancer. A theoretical analysis of cooperativity is presented, through which the binding of a constitutively expressed nuclear factor is shown to have weak positive cooperativity. In summary, TGF$\beta$, RAS, and enhancer-binding factors are found to be initial, intermediary, and final regulators, respectively, of muscle differentiation. ^

Mechanism of molecular regulation of nuclear -encoded mitochondrial gene expression by electrical stimulation in the cultured neonatal rat cardiac myocytes

To answer the question whether increased energy demand resulting from myocyte hypertrophy and enhanced $\beta$-myosin heavy chain mRNA, contractile protein synthesis and assembly leads to mitochondrial proliferation and differentiation, we set up an electrical stimulation model of cultured neonatal rat cardiac myocytes. We describe, as a result of increased contractile activity, increased mitochondrial profiles, cytochrome oxidase mRNA, and activity, as well as a switch in mitochondrial carnitine palmitoyltransferase-I (CPT-I) from the liver to muscle isoform. We investigate physiological pathways that lead to accumulation of gene transcripts for nuclear encoded mitochondrial proteins in the heart. Cardiomyocytes were stimulated for varying times up to 72 hr in serum-free culture. The mRNA contents for genes associated with transcriptional activation (c-fos, c-jun, junB, nuclear respiratory factor 1 (Nrf-1)), mitochondrial proliferation (cytochrome c (Cyt c), cytochrome oxidase), and mitochondrial differentiation (carnitine palmitonyltransferase I (CPT-I) isoforms) were measured. The results establish a temporal pattern of mRNA induction beginning with c-fos (0.25-3 hr) and followed by c-jun (0.5-3 hr), junB (0.5-6 hr), NRF-1 (1-12 hr), Cyt c (12-72 hr), cytochrome c oxidase (12-72 hr). Induction of the latter was accompanied by a marked decrease in the liver-specific CPT-I mRNA. Electrical stimulation increased c-fos, $\beta$-myosin heavy chain, and Cyt c promoter activities. These increases coincided with a rise in their respective endogenous gene transcripts. NRF-1, cAMP response element (CRE), and Sp-1 site mutations within the Cyt c promoter reduced luciferase expression in both stimulated and nonstimulated myocytes. Mutations in the Nrf-1 and CRE sites inhibited the induction by electrical stimulation or by transfection of c-jun into non-paced cardiac myocytes whereas mutation of the Sp-1 site maintained or increased the fold induction. This is consistent with the appearance of NRF-1 and fos/jun mRNAs prior to that of Cyt c. Overexpression of c-jun by transfection also activates the Nrf-1 and Cyt c mRNA sequentially. Electrical stimulation of cardiac myocytes activates the c-Jun-N-terminal kinase so that the fold-activation of the cyt c promoter is increased by pacing when either c-jun or c-fos/c-jun are cotransfected. We have identified physical association of Nrf-1 protein with the Nrf-1 enhancer element and of c-Jun with the CRE binding sites on the Cyt c promoter. This is the first demonstration that induction of Nrf-1 and c-Jun by pacing of cardiac myocytes directly mediates Cyt c gene expression and mitochondrial proliferation in response to hypertrophic stimuli in the heart.^ Subsequent to gene activation pathways that lead to mitochondrial proliferation, we observed an isoform switch in CPT-I from the liver to muscle mRNA. We have found that the half-life for the muscle CPT-I is not affected by electrical stimulation, but electrical decrease the T1/2 in the liver CPT-I by greater than 50%. This suggests that the liver CPT-I switch to muscle isoform is due to (1) a decrease in T1/2 of liver CPT-I and (2) activation of muscle CPT-Itranscripts by electrical stimulation. (Abstract shortened by UMI.) ^

The role of Ski/Sno oncoproteins as negative regulators of the TGF-beta/SMAD signaling pathway in B -cell non -Hodgkin's lymphoma (NHL-B)

Non-Hodgkin's lymphomas are common tumors of the human immune system, primarily of B cell lineage (NHL-B). Negative growth regulation in the B cell lineage is mediated primarily through the TGF-β/SMAD signaling pathway that regulates a variety of tumor suppressor genes. Ski was originally identified as a transforming oncoprotein, whereas SnoN is an isoform of the Sno protein that shares a large region of homology with Ski. In this study, we show that Ski/SnoN are endogenously over-expressed both in patients' lymphoma cells and NHL-B cell lines. Exogenous TGF-β1 treatment induces down-regulation of Ski and SnoN oncoprotein expression in an NHL-B cell line, implying that Ski and SnoN modulate the TGF-β signaling pathway and are involved in cell growth regulation. Furthermore, we have developed an NHL-B cell line (DB) that has a null mutation in TGF-β receptor type II. In this mutant cell line, Ski/SnoN proteins are not down-regulated in response to TGF-β1 treatment, suggesting that downregulation of Ski and SnoN proteins in NHL-B require an intact functional TGF-β signaling pathway Resting normal B cells do not express Ski until activated by antigens and exogenous cytokines, whereas a low level of SnoN is also present in peripheral blood Go B cells. In contrast, autonomously growing NHL-B cells over-express Ski and SnoN, implying that Ski and SnoN are important cell cycle regulators. To further investigate a possible link between reduction of the Ski protein level and growth inhibition, Ski antisense oligodeoxynucleotides were transfected into NHL-B cells. The Ski protein level was found to decrease to less than 40%, resulting in restoring the effect of TGF-β and leading to cell growth inhibition and G1 cell cycle arrest. Co-immunoprecipitation experiments demonstrated that Ski associates with Smad4 in the nucleus, strongly suggesting that over-expression of the nuclear protein Ski and/or SnoN negatively regulates the TGF-β pathway, possibly by modulating Smad-mediated tumor suppressor gene expression. Together, in NHL-B, the TGF-β/SMAD growth inhibitory pathway is usually intact, but over-expression of the Ski and/or SnoN, which binds to Smad4, abrogates the negative regulatory effects of TGF-β/SMAD in lymphoma cell growth and potentiates the growth potential of neoplastic B cells. ^