Retinoid -induced regression of human head and neck squamous cell carcinomas is associated with the induction of a pro -inflammatory response

Retinoids are Vitamin A derivatives that are effective chemopreventative and chemotherapeutic agents for head and neck squamous cell carcinomas (HNSCC). Despite the wide application of retinoids in cancer treatment, the mechanism by which retinoids inhibit head and neck squamous cell carcinomas is not completely understood. While in vitro models show that drugs affect cell proliferation and differentiation, in vivo models, such as tumor xenografts in nude mice drugs affect more complex parameters such as extracellular matrix formation, angiogenesis and inflammation. Therefore, we studied the effects of retinoids on the growth of the 22B HNSCC tumors using a xenograft model. In this system, retinoids had no effect on tumor cell differentiation but caused invasion of the tumor by inflammatory cells. Retinoid induced inflammation lead to tumor cell death and tumor regression. Therefore, we hypothesized that retinoids stimulated the 22B HNSCC xenografts to produce a pro-inflammatory signal such as chemokines that in turn activated host inflammatory responses. ^ We used real time quantitative RT-PCR to measure cytokine and chemokine expression in retinoid treated tumors. Treatment of tumors with an RAR-specific retinoid, LGD1550, had no effect on the expression of TNFα, IL-1α, GROα, IP-10, Rantes, MCP-1 and MIP-1α but induced IL-8 mRNA 5-fold. We further characterized the retinoid effect on IL-8 expression on the 22B HNSCC and 1483 HNSCC cells in vitro. Retinoids increased IL-8 expression and enhanced TNFα-dependent IL-8 induction. In addition, retinoids increased the basal and TNFα-dependent expression of MCP-1 but decreased the basal and TNFα dependent expression of IP-10. The effect of retinoids on IL-8 and MCP-1 expression was very rapid with increased levels of mRNA detected within 1–2 hours. This effect did not require new protein synthesis and did not result from mRNA stabilization. Both RAR and RXR ligands increased IL-8 expression whereas only RAR ligands activated MCP-1 expression. ^ We identified a functional retinoid response element in the IL-8 promoter that was located adjacent to the C/EBP-NFkB response element. TNFα treatment of the 22B cells caused rapid, transient and selective acetylation of regions of the IL-8 promoter associated with the NFkB response element. Co-treatment of the cells with retinoids plus TNF increased the acetylation of chromatin in this region without altering the kinetics of acetylation. These results demonstrate that ligand activated retinoid receptors can cooperate with NFkB in histone acetylation and chromatin remodeling. We believe that in certain HNSCC tumors this cooperation and the resulting enhancement of IL-8 expression can induce an inflammatory response that leads to tumor regression. We believe that the induction of inflammation in susceptible tumors, possibly coupled with cytotoxic interventions may be an important component in the use of retinoids to treat human squamous cancers. ^

Identification and characterization of cellular genes differentially expressed in a subset of tumors with non -functional WT1

Wilms tumor (WT) or nephroblastoma is a genetically heterogeneous pediatric renal tumor that accounts for 6–7% of all childhood cancers in the U.S. WT1, located at 11p13, is the sole WT gene cloned to date. Additional genomic regions containing genes that play a role in the development of Wilms tumor include 11p15, 7p, 16q, 1p, 17q and 19q. This heterogeneity has made it extremely difficult to develop an understanding of the pathways involved in the development of WT, even in the 5–20% of tumors that show mutations at the WT1 locus. My research addresses this gap in our current comprehension of the development of WT. ^ I have used two complementary approaches to extend the current understanding of molecular changes involved in the development of WT. In order to minimize complexities due to genetic heterogeneity, I confined my analysis to the WT1 pathway by assessing those genetically defined tumors that carry WT1 mutations. WT1 encodes a zinc finger transcription factor, and in vitro studies have identified many genes that are potentially regulated in vivo by WT1. However, there is very little in vivo data that suggests that they are transcriptionally regulated endogenously by WT1. In one approach I assessed the role of WT1 in the in vivo regulation of PDGFA and IGF2, two genes that are strong contenders for endogenous regulation by WT1. Using primary tissue samples, I found no correlation between the level of RNA expression of WT1 with either PDGFA or IGF2, suggesting that WT1 does not play a critical role in their expression in either normal kidney or WT. ^ In a parallel strategy, using differential display analysis I compared global gene expression in a subset of tumors with known homozygous inactivating WT1 mutations (WT1-tumors) to the gene expression in a panel of appropriate control tissues (fetal kidney, normal kidney, rhabdoid tumor and pediatric renal cell carcinoma). Transcripts that are aberrantly expressed in this subset of Wilms tumors are candidates for endogenous transcriptional regulation by WT1 as well as for potentially functioning in the development of WT. By this approach I identified several differentially expressed transcripts. I further characterized two of these transcripts, identifying a candidate WT gene in the process. I then performed a detailed analysis of this WT candidate gene, which maps to 7p. Future studies will shed more light on the role of these differentially expressed genes in WT. ^

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. ^

*Ras proteins and human tumor cell radiosensitivity

Ras proteins (H-, N-, K4A-, and K4B) are associated with cellular resistance to ionizing radiation (IR) and, consequently, may provide a potential target for radiosensitization strategies in cancer treatment. Several approaches have been used to compromise Ras activity and enhance IR-induced cell killing; however, these techniques either target proteins in addition to Ras or only target one member of the Ras family. In this study, I have used an adenovirus (AV1Y28) that expresses a single-chain antibody fragment directed against Ras proteins to investigate the mechanism(s) responsible for Ras-mediated radiation resistance. AV1Y28 enhanced the radiosensitivity of a number of human tumor cell lines without affecting the radiosensitivity of normal human fibroblasts. Whereas AV1Y28-mediated sensitization was independent of ras gene mutational status, it was dependent on active Ras proteins suggesting that AV1Y28 may be useful against a broad range of tumors. AV1Y28-mediated cell killing was not the result of redistributing cells into a more radiosensitive phase of the cell cycle and did not enhance IR-induced apoptosis. Given that Ras proteins transduce environmental signals to the nucleus, the effect of AV1Y28 on the IR-inducible transcription factor NF-κB were determined. Although AV1Y28 inhibited IR-induced NF-κB through the suppression of IKK, additional work established that NF-κB did not play a role in AV1Y28-mediated radiosensitization. However, a novel component of the signaling pathway responsible for IR-induced NF-κB was identified. Previous studies had suggested a relationship between mutant ras genes and IR-induced G2 delay; therefore the effects of AV1Y28 on the progression of cells from G2 to M after IR were determined. Pretreatment of cells with AV1Y28 prevented the IR-induced G2 arrest. AV1Y28-mediated abrogation of IR-induced G2 arrest correlated with those cell line lines that were sensitized by AV1Y28. Moreover, a significant increase in cells undergoing mitotic catastrophe was found after IR in AV1Y28 treated cells. The abrogation of G2 arrest by AV1Y28 was the result of maintaining the active form of cdc2, an inducer of mitosis, after exposure to IR. This study identified the mechanism of AV1Y28-mediated radiosensitization and has provided insight into the signal transduction pathways responsible for Ras-mediated radiation resistance. ^

Pathways leading to apoptosis resistance in a murine B -cell lymphoma cell system

The aberrant activation of signal transduction pathways has long been linked to uncontrolled cell proliferation and the development of cancer. The activity of one such signaling module, the Mitogen-Activated Protein Kinase (MAPK) pathway, has been implicated in several cancer types including pancreatic, breast, colon, and lymphoid malignancies. Interestingly, the activation of MAP-Kinase-Kinase-Kinase proteins often leads to the additional activation of NF-κB, a transcription factor that acts as a cell survival signal through its control of antiapoptotic genes. We have investigated the role of a specific dimer form of the NF-κB transcription factor family, NF-κB1 (p50) homodimers, in its control of the proto-oncogene, Bcl-2, and we have identified the MEK/ERK (MAPK) signaling cascade as a mediator of NF-κB1 activity. ^ Two murine B cell lymphoma cell lines were used for these studies: LY-as, an apoptosis proficient line with low Bcl-2 protein expression and no nuclear NF-κB activity, and LY-ar, a nonapoptotic line with constitutive p50 homodimer activity and 30 times more Bcl-2 protein expression than LY-as. Experiments modulating p50 activity correlated the activation of p50 homodimers with Bcl-2 expression and additional gel shift experiments demonstrated that the Bcl-2 P1 promoter had NF-κB sites with which recombinant p50 was able to interact. In vitro transcription revealed that p50 enhanced the production of transcripts derived from the Bcl-2 P1 promoter. These data strongly suggest that Bcl-2 is a target gene for p50-mediated transcription and suggest that the activation of p50 homodimers contributes to the expression of Bcl-2 observed in LY-ar cells. ^ Studies of upstream MAPK pathways that could influence NF-κB activity demonstrated that LY-ar cells had phosphorylated ERK proteins while LY-as cells did not. Treatment of LY-ar cells with the MEK inhibitors PD 98059, U0126, and PD 184352 led to a loss of phosphorylated ERK, a reversal of nuclear p50 homodimer DNA binding, and a decrease in the amount of Bcl-2 protein expression. Similarly, the activation of the MEK/ERK pathway in LY-as cells by phorbol ester led to Bcl-2 expression that could be blocked by PD 98059. Furthermore, treatment of LY-ar cells with TNFα, an IKK activator, did not change the suppressive effect of PD 98059 on p50 homodimer activity, suggesting an IKK-independent pathway for p50 homodimer activation. Lastly, all three MEK inhibitors sensitized LY-ar cells to radiation-induced apoptosis. ^ These data indicate that the activation of the MEK/ERK MAP-Kinase signaling pathway acts upstream of p50 homodimer activation and Bcl-2 expression in this B cell lymphoma cell system and suggest that the activation of MEK/ERK may be a key step in the progression of lymphoma to advanced-staged disease. Other researchers have used MEK inhibitors to inhibit cell growth and sensitize a number of tumors to chemotherapies. In light of our data, MEK inhibitors may additionally be useful clinically to radiosensitize cancers of lymphoid origin. ^

Function of GCIP (CCNDBP1), a helix -loop -helix leucine -zip protein, in cell differentiation and tumorigenesis

Cell growth and differentiation are complex and well-organized processes in which cells respond to stimuli from the environment by carrying out genetic programs. Transcription factors with helix-loop-helix (HLH) motif play critical roles in controlling the expression of genes involved in lineage commitment, cell fate determination, proliferation and tumorigenesis. This study has examined the roles of GCIP (CCNDBP1) in cell differentiation and tumorigenesis. GCIP is a recently identified HLH-leucine zipper protein without a basic region like the Id family of proteins. However, GCIP shares little sequence homology with the Id proteins and has domains with high acidic amino acids and leucine-rich regions following the HLH domain like c-Myc. Here we firstly demonstrate that GCIP is a transcription regulator related to muscle differentiation program. Overexpression of GCIP in C2C12 cells not only promotes myotube formation but also upregulates myogenic differentiation biomarkers, including MHC and myogenein. On the other hand, our finding also suggests that GCIP is a potential tumor suppressor related to cell cycle control. Expression of GCIP was significantly down-regulated in colon tumors as compared to normal colon tissues. Overexpression of GCIP in SW480 colon cancer cell line resulted in a significant inhibition on tumor cell colony formation on soft agar assays while silencing of GCIP expression by siRNA can promote cell proliferation and colony formation. In addition, results from transgenic mice specifically expressing GCIP in liver also support the idea that GCIP is involved in the early stage of hepatocarcinogenesis and decreased susceptibility to chemical hepatocarcinogenesis. ^

The Agrobacterium vitis T-6b oncoprotein induces auxin-independent cell expansion in tobacco

Among the Agrobacterium T-DNA genes, rolB, rolC, orf13, orf8, lso, 6b and several other genes encode weakly homologous proteins with remarkable effects on plant growth. The 6b oncogene induces tumors and enations. In order to study its properties we have used transgenic tobacco plants that carry a dexamethasone-inducible 6b gene, dex-T-6b. Upon induction, dex-T-6b plants develop a large array of morphological modifications, some of which involve abnormal cell expansion. In the present investigation, dex-T-6b-induced expansion was studied in intact leaves and an in vitro leaf disc system. Although T-6b and indole-3-acetic acid (IAA) both induced expansion and were non-additive, T-6b expression did not increase IAA levels, nor did it induce an IAA-responsive gene. Fusicoccin (FC) is known to stimulate expansion by increasing cell wall plasticity. T-6b- and FC-induced expansion were additive at saturating FC concentrations, indicating that T-6b does not act by a similar mechanism to FC. T-6b expression led to higher leaf osmolality values, in contrast to FC, suggesting that the T-6b gene induces expansion by increasing osmolyte concentrations. Metabolite profiling showed that glucose and fructose played a major role in this increase. We infer that T-6b disrupts the osmoregulatory controls that govern cell expansion during development and wound healing.

The FEZ1 gene at chromosome 8p22 encodes a leucine-zipper protein, and its expression is altered in multiple human tumors

Alterations of human chromosome 8p occur frequently in many tumors. We identified a 1.5-Mb common region of allelic loss on 8p22 by allelotype analysis. cDNA selection allowed isolation of several genes, including FEZ1. The predicted Fez1 protein contained a leucine-zipper region with similarity to the DNA-binding domain of the cAMP-responsive activating-transcription factor 5. RNA blot analysis revealed that FEZ1 gene expression was undetectable in more than 60% of epithelial tumors. Mutations were found in primary esophageal cancers and in a prostate cancer cell line. Transcript analysis from several FEZ1-expressing tumors revealed truncated mRNAs, including a frameshift. Alteration and inactivation of the FEZ1 gene may play a role in various human tumors.

Cross-lineage expression of Ig-β (B29) in thymocytes: Positive and negative gene regulation to establish T cell identity

Developmental commitment involves activation of lineage-specific genes, stabilization of a lineage-specific gene expression program, and permanent inhibition of inappropriate characteristics. To determine how these processes are coordinated in early T cell development, the expression of T and B lineage-specific genes was assessed in staged subsets of immature thymocytes. T lineage characteristics are acquired sequentially, with germ-line T cell antigen receptor-β transcripts detected very early, followed by CD3ɛ and terminal deoxynucleotidyl transferase, then pTα, and finally RAG1. Only RAG1 expression coincides with commitment. Thus, much T lineage gene expression precedes commitment and does not depend on it. Early in the course of commitment to the T lineage, thymocytes lose the ability to develop into B cells. To understand how this occurs, we also examined expression of well defined B lineage-specific genes. Although λ5 and Ig-α are not expressed, the μ0 and Iμ transcripts from the unrearranged IgH locus are expressed early, in distinct patterns, then repressed just before RAG1 expression. By contrast, RNA encoding the B cell receptor component Ig-β was found to be transcribed in all immature thymocyte subpopulations and throughout most thymocyte differentiation. Ig-β expression is down-regulated only during positive selection of CD4+CD8– cells. Thus several key participants in the B cell developmental program are expressed in non-B lineage-committed cells, and one is maintained even through commitment to an alternative lineage, and repressed only after extensive T lineage differentiation. The results show that transcriptional activation of “lymphocyte-specific” genes can occur in uncommitted precursors, and that T lineage commitment is a composite of distinct positive and negative regulatory events.

β-Catenin mutations in cell lines established from human colorectal cancers

β-catenin has functions as both an adhesion and a signaling molecule. Disruption of these functions through mutations of the β-catenin gene (CTNNB1) may be important in the development of colorectal tumors. We examined the entire coding sequence of β-catenin by reverse transcriptase–PCR (RT-PCR) and direct sequencing of 23 human colorectal cancer cell lines from 21 patients. In two cell lines, there was apparent instability of the β-catenin mRNA. Five different mutations (26%) were found in the remaining 21cell lines (from 19 patients). A three-base deletion (codon 45) was identified in the cell line HCT 116, whereas cell lines SW 48, HCA 46, CACO 2, and Colo 201 each contained single-base missense mutations (codons 33, 183, 245, and 287, respectively). All 23 cell lines had full-length β-catenin protein that was detectable by Western blotting and that coprecipitated with E-cadherin. In three of the cell lines with CTNNB1 mutations, complexes of β-catenin with α-catenin and APC were detectable. In SW48 and HCA 46, however, we did not detect complexes of β-catenin protein with α-catenin and APC, respectively. These results show that selection of CTNNB1 mutations occurs in up to 26% of colorectal cancers from which cell lines are derived. In these cases, mutation selection is probably for altered β-catenin function, which may significantly alter intracellular signaling and intercellular adhesion and may serve as a complement to APC mutations in the early stages of tumorigenesis.

Cell cycle-dependent colocalization of BARD1 and BRCA1 proteins in discrete nuclear domains

Germ-line mutations of the BRCA1 gene predispose women to early-onset breast and ovarian cancer by compromising the gene’s presumptive function as a tumor suppressor. Although the biochemical properties of BRCA1 polypeptides are not understood, their expression pattern and subcellular localization suggest a role in cell-cycle regulation. When resting cells are induced to proliferate, the steady-state levels of BRCA1 increase in late G1 and reach a maximum during S phase. Moreover, in S phase cells, BRCA1 polypeptides are hyperphosphorylated and accumulate into discrete subnuclear foci termed “BRCA1 nuclear dots.” BRCA1 associates in vivo with a structurally related protein termed BARD1. Here we show that the steady-state levels of BARD1, unlike those of BRCA1, remain relatively constant during cell cycle progression. However, immunostaining revealed that BARD1 resides within BRCA1 nuclear dots during S phase of the cell cycle, but not during the G1 phase. Nevertheless, BARD1 polypeptides are found exclusively in the nuclear fractions of both G1- and S-phase cells. Therefore, progression to S phase is accompanied by the aggregation of nuclear BARD1 polypeptides into BRCA1 nuclear dots. This cell cycle-dependent colocalization of BARD1 and BRCA1 indicates a role for BARD1 in BRCA1-mediated tumor suppression.

Syncytiotrophoblastic giant cells in teratocarcinoma-like tumors derived from Parp-disrupted mouse embryonic stem cells

The enzyme poly(ADP-ribose) polymerase (Parp) catalyzes poly(ADP-ribosyl)ation reaction and is involved in DNA repair and cell death induction upon DNA damages. Meanwhile, poly(ADP-ribosyl)ation of chromosome-associated proteins is suggested to be implicated in the regulation of gene expression and cellular differentiation, both of which are important in tumorigenesis. To investigate directly the role of Parp deficiency in tumorigenicity and differentiation of embryonic stem (ES) cells during tumor formation, studies were conducted by using wild-type J1 (Parp+/+) ES cells and Parp+/− and Parp−/− ES clones generated by disrupting Parp exon 1. These ES cells, irrespective of the Parp genotype, produced tumors phenotypically similar to teratocarcinoma when injected s.c. into nude mice. Remarkably, all tumors derived from Parp−/− clones contained syncytiotrophoblastic giant cells (STGCs), which possess single or multiple megalo-nuclei. The STGCs were present within large areas of intratumoral hemorrhage. In contrast, neither STGC nor hemorrhage was observed in tumors of both wild-type J1 cells and Parp+/− clones. Electron microscopic examination showed that the STGCs possess microvilli on the cell surface and contained secretory granules in the cytoplasm. Furthermore, the cytoplasms of STGCs were strongly stained with antibody against mouse prolactin, which could similarly stain trophoblasts in placenta. These morphological and histochemical features indicate that the STGCs in teratocarcinoma-like tumors derived from Parp−/− clones belong to the trophoblast cell lineage. Our findings thus suggest that differentiation of ES cells into STGCs was possibly induced by the lack of Parp during the development of teratocarcinoma.

NADPH-oxidase and a hydrogen peroxide-sensitive K+ channel may function as an oxygen sensor complex in airway chemoreceptors and small cell lung carcinoma cell lines

Pulmonary neuroepithelial bodies (NEB) are widely distributed throughout the airway mucosa of human and animal lungs. Based on the observation that NEB cells have a candidate oxygen sensor enzyme complex (NADPH oxidase) and an oxygen-sensitive K+ current, it has been suggested that NEB may function as airway chemoreceptors. Here we report that mRNAs for both the hydrogen peroxide sensitive voltage gated potassium channel subunit (KH2O2) KV3.3a and membrane components of NADPH oxidase (gp91phox and p22phox) are coexpressed in the NEB cells of fetal rabbit and neonatal human lungs. Using a microfluorometry and dihydrorhodamine 123 as a probe to assess H2O2 generation, NEB cells exhibited oxidase activity under basal conditions. The oxidase in NEB cells was significantly stimulated by exposure to phorbol esther (0.1 μM) and inhibited by diphenyliodonium (5 μM). Studies using whole-cell voltage clamp showed that the K+ current of cultured fetal rabbit NEB cells exhibited inactivating properties similar to KV3.3a transcripts expressed in Xenopus oocyte model. Exposure of NEB cells to hydrogen peroxide (H2O2, the dismuted by-product of the oxidase) under normoxia resulted in an increase of the outward K+ current indicating that H2O2 could be the transmitter modulating the O2-sensitive K+ channel. Expressed mRNAs or orresponding protein products for the NADPH oxidase membrane cytochrome b as well as mRNA encoding KV3.3a were identified in small cell lung carcinoma cell lines. The studies presented here provide strong evidence for an oxidase-O2 sensitive potassium channel molecular complex operating as an O2 sensor in NEB cells, which function as chemoreceptors in airways and in NEB related tumors. Such a complex may represent an evolutionary conserved biochemical link for a membrane bound O2-signaling mechanism proposed for other cells and life forms.

Estimating the probability of initiated cell death before tumor induction

The effects of cell toxicity are known to be inherent in carcinogenesis induced by radiation or chemical carcinogens. The event of cell death precludes tumor induction from occurring. A long standing problem is to estimate the proportion of initiated cells that die before tumor induction. No experimental techniques are currently available for directly gauging the rate of cell death over extended periods of time. The obstacle can be surmounted by newly developed theoretical methods of carcinogenesis modeling. In this paper, we apply such methods to published data on multiple lung tumors in mice receiving different schedules of urethane. Bioassays of this type play an important role in testing environmental chemicals for carcinogenic activity. Our estimates for urethane-induced carcinogenesis show that, unexpectedly, many initiated cells die early in the course of tumor promotion. We present numerical estimates for the probability of initiated cell death for different schedules (and doses) of urethane administration.

Natural killer cell lines kill autologous β2-microglobulin-deficient melanoma cells: Implications for cancer immunotherapy

Cancer vaccines used to generate specific cytotoxic T lymphocytes are not effective against tumor cells that have lost or suppressed expression of their class I major histocompatibility complex proteins. This loss is common in some cancers and particularly in metastatic lesions. We show that β2-microglobulin-deficient class I-negative melanoma variants derived from patients undergoing specific T cell therapy are lysed by heterologous as well as autologous natural killer (NK) lines and clones, but not by specific T cells. Moreover, the minor NK cell fraction but not the major T cell fraction derived from heterologous lymphokine activated killer cells kills those tumor cell lines. ICAM-1 expression by the different class I protein deficient tumors was correlated with their sensitivity to lysis by NK cells. Adoptive autologous NK therapy may be an important supplement to consider in the design of new cancer immunotherapies.