Recent advances in cancer research: mouse models of tumorigenesis

Over the past 20 years, cancer research has gained major insights into the complexity of tumor development, in particular into the molecular mechanisms that underlie the progressive transformation of normal cells into highly malignant derivatives. It is estimated that the transformation of a normal cell to a malignant tumor cell is dependent upon a small number of genetic alterations, estimated to be within the range of four to seven rate-limiting events. Critical events in the evolution of neoplastic disease include the loss of proliferative control, the failure to undergo programmed cell death (apoptosis), the onset of neoangiogenesis, tissue remodeling, invasion of tumor cells into surrounding tissue and, finally, metastatic dissemination of tumor cells to distant organs. In patients, the molecular analysis of these multiple steps is hampered by the unavailability of tumor biopsies from all tumor stages. In contrast, mouse models of tumorigenesis allow the reproducible isolation of all tumor stages, including normal tissue, which are then amenable to pathological, genetic and biochemical analyses and, hence, have been instrumental in investigating cancer-related genes and their role in carcinogenesis. In this review, we discuss mouse tumor models that have contributed substantially to the identification and characterization of novel tumor pathways. In particular, we focus on transgenic and knockout mouse models that closely mimic human cancer and thus can be used as model systems for cancer research.     

CELL KINETICS OF IRRADIATED EXPERIMENTAL TUMORS: RELATIONSHIP BETWEEN THE PROLIFERATING AND THE NONPROLIFERATING POOL

The role of nonproliferating cells in tumor regeneration has been studied after subcurative doses of low L.E.T. irradiation. Radiation was applied in a single dose at three different levels, 0–47, 0–94 and 1–88 krad. Studies included estimation of the absolute number of cells per tumor, differential cell counts, and autoradiographic determination of kinetic variables, employing transplantable mouse mammary adenocarcinoma DBAH. Quantitative changes of morphologically denned proliferating and non-proliferating cell pools were followed at different time intervals after irradiation. Irradiation resulted in reduction of the number of cells in both pools, with apparent sparing of nonproliferating cells. The regenerative period started with a gradual increase in the number of cells in the proliferating pool, whereas the number of cells in the nonproliferating pool continued to fall in tumors irradiated with 0–94 and 1 -88 krad. In the late phase of tumor regrowth, the increasing number of cells in the non proliferating pool corresponded to its replenishment by cell transition from the proliferating pool. In an effort to clarify whether cell transition from the nonproliferating to the proliferating pool may take place during the regrowth of radiation perturbed tumors, cell loss rates from both pools were estimated using experimental data. In addition to cell losses from the tumor as a whole, the ‘net loss rate’ of the non-proliferating pool reflects the rate of cell transition from the nonproliferating to the proliferating pool, minus the rate of transition in the opposite direction. A similar definition applies to cell loss rates from the proliferating pool. The results showed: (1) high losses in both pools, with excess losses in the proliferating during the early phase after irradiation; (2) in the early stage of regrowth after irradiation, the cell net loss rate for the nonproliferating pool increased, in contrast to the behavior of cell loss rate for the proliferating pool and the average cell loss rate for the tumor as a whole; (3) in the late stage of regrowth a decrease in net loss rate for the nonproliferating pool reflects the excess production of nonproliferating cells over control tumors. These results suggest that cell transition from the nonproliferating to the proliferating pool takes place at the beginning of tumor regrowth after subcurative single-dose irradiation.     

The role of alpha 6 integrin in prostate cancer migration and bone pain in a novel xenograft model

Of the estimated 565,650 people in the U.S. who will die of cancer in 2008, almost all will have metastasis. Breast, prostate, kidney, thyroid and lung cancers metastasize to the bone. Tumor cells reside within the bone using integrin type cell adhesion receptors and elicit incapacitating bone pain and fractures. In particular, metastatic human prostate tumors express and cleave the integrin A6, a receptor for extracellular matrix components of the bone, i.e., laminin 332 and laminin 511. More than 50% of all prostate cancer patients develop severe bone pain during their remaining lifetime. One major goal is to prevent or delay cancer induced bone pain. We used a novel xenograft mouse model to directly determine if bone pain could be prevented by blocking the known cleavage of the A6 integrin adhesion receptor. Human tumor cells expressing either the wildtype or mutated A6 integrin were placed within the living bone matrix and 21 days later, integrin expression was confirmed by RT-PCR, radiographs were collected and behavioral measurements of spontaneous and evoked pain performed. All animals independent of integrin status had indistinguishable tumor burden and developed bone loss 21 days after surgery. A comparison of animals containing the wild type or mutated integrin revealed that tumor cells expressing the mutated integrin resulted in a dramatic decrease in bone loss, unicortical or bicortical fractures and a decrease in the ability of tumor cells to reach the epiphyseal plate of the bone. Further, tumor cells within the bone expressing the integrin mutation prevented cancer induced spontaneous flinching, tactile allodynia, and movement evoked pain. Preventing A6 integrin cleavage on the prostate tumor cell surface decreased the migration of tumor cells within the bone and the onset and degree of bone pain and fractures. These results suggest that strategies for blocking the cleavage of the adhesion receptors on the tumor cell surface can significantly prevent cancer induced bone pain and slow disease progression within the bone. Since integrin cleavage is mediated by Urokinase-type Plasminogen Activator (uPA), further work is warranted to test the efficacy of uPA inhibitors for prevention or delay of cancer induced bone pain.     

Cell kinetics of irradiated experimental tumors: relationship between the proliferating and the nonproliferating pool.

The role of nonproliferating cells in tumor regeneration has been studied after subcurative doses of low L.E.T. irradiation. Radiation was applied in a single dose at three different levels; 0-47, 0-94 and 1-88 krad. Studies included estimation of the absolute number of cells per tumor, differential cell counts, and autoradiographic determination of kinetic variables, employing transplantable mouse mammary adenocarcinoma DBAH. Quantitative changes of morphologically defined proliferating and nonproliferating cell pools were followed at different time intervals after irradiation. Irradiation resulted in reduction of the number of cells in both pools, with apparent sparing of nonproliferating cells. The regenerative period started with a gradual increase in the number of cells in the proliferating pool, whereas the number of cells in the nonproliferating pool continued to fall in tumors irradiated with 0-94 and 1-88 krad. In the late phase of tumor regrowth, the increasing number of cells in the non proliferating pool corresponded to its replenishment by cell transition from the proliferating pool. In an effort to clarify whether cell transition from the nonproliferating to the proliferating pool may take place during the regrowth of radiation perturbed tumors, cell loss rates from both pools were estimated using experimental data. In addition to cell loses from the tumor as a whole, the 'net loss rate' of the nonproliferating pool reflects the rate of cell transition from the nonproliferating to the proliferating pool, minus the rate of transition in the opposite direction. A similar definition applies to cell loss rates from the proliferating pool. The results showed: (1) high losses in both pools, with excess losses in the proliferating during the early phase after irradiation; (2) in the early stage of regrowth after irradiation, the cell net loss rate f-or the nonproliferating pool increased, in contrast to the behavior of cell loss rate for the proliferating pool and the average cell loss rate for the tumor as a whole; (3) in the late stage of regrowth a decrease in net loss rate for the nonproliferating pool reflects the excess production of nonproliferating cells over control tumors. These results suggest that cell transition from the nonproliferating to the proliferating pool takes place at the beginning of tumor regrowth after subcurative single-dose irradiation.     

Distribution of HLA class I altered phenotypes in colorectal carcinomas: high frequency of HLA haplotype loss associated with loss of heterozygosity in chromosome region 6p21

HLA class I loss or down-regulation is a widespread mechanism used by tumor cells to avoid tumor recognition by cytotoxic T lymphocytes, and thus favor tumor immune escape. Multiple mechanisms are responsible for these HLA class I alterations. In different epithelial tumors, loss of heterozygosity (LOH) at chromosome region 6p21.3, leading to HLA haplotype loss, occurs in 6–50% of all cases depending on the tumor entity. In this paper we report the frequency of LOH at 6p21 in 95 colorectal carcinomas (CRC) previously analyzed for altered HLA class I expression with immunohistological techniques. We used PCR microsatellite amplification of selected STR markers located on Chromosome 6 to identify LOH with DNA from microdissected tumor tissues and the surrounding stroma. Sequence-specific oligonucleotide analysis was performed in microdissected stroma and tumor cells for HLA typing, and to detect HLA haplotype loss. A high frequency (40%) of HLA haplotype loss was found in CRC. Eight tumors showed microsatellite instability. We sometimes observed two or more mechanisms responsible for HLA alteration within the same HLA-altered phenotype, such as LOH and HLA class I total loss. In 25 tumors (26%) no HLA class I alteration could be identified. These data are potentially relevant for CRC patients undergoing T-cell-based immunotherapy.     

Inferring the location of tumor suppressor genes by modeling frequency of allelic loss

Allelic loss is often part of a multistep process leading to tumorigenesis. Analysis of genomic markers highlights regions of elevated allelic loss, which in turn suggests a nearby tumor suppressor. Furthermore, pooling published analyses to combine evidence can increase the power to detect a tumor suppressor gene. If the pattern of loss for each tumor, or allelotype, is known, a stochastic model proposed by Newton et al. (1998, Statistics in Medicine 17, 1425-1445) can be used to analyze the correlated binary data. Many studies report only incomplete allelotypes, augmented with frequencies of allelic loss (FAL) at each marker, in which the number of informative tumors showing allelic loss is provided along with the number of informative tumors. We describe an extension of the allelotype model to handle FAL data, using a hidden Markov model or a normal approximation to compute the likelihood. The FAL model is illustrated using data from a study of colorectal cancer.     

Tumor volume estimation by the percentage carcinoma method in uterine cervix carcinoma

OBJECTIVE: To assess the capacity of the percentage carcinoma method to predict tumor volume in cervical carcinoma and generate a mathematical equation for calculation of tumor volume. STUDY DESIGN: Thirteen radical hysterectomy specimens were studied. The actual tumor volume was assessed by stereology. A factor was generated by the formula tumor volume = factor x percentage carcinoma x number of blocks. The percentage carcinoma was calculated by the grid method. Then tumor volume was estimated by the same formula. The relationship between actual and estimated tumor volumes was analyzed by Spearman's correlation. RESULTS: Correlation of the tumor volumes assessed by the two methods was excellent (r = .945) and statistically significant (P = .01). CONCLUSION: Percentage carcinoma assessed by the grid ratio method is highly predictive of tumor volume, but care must be taken not to overstate the importance of our results due to the small number of patients. The prediction of outcome in cervical carcinoma by means of percentage carcinoma should be evaluated in large clinical studies.     

Mathematical analysis of cancer chemotherapy

This paper presents a mathematical analysis of a tumor model first proposed by Skipper and Zubrod. The tumor model is comprised of three compartments, a proliferative compartment, a nonproliferative but viable compartment, and a dead compartment. By the suitable selection of functions describing loss of cells from the proliferative and nonproliferative compartments, the model is capable of describing tumor behavior during periods of growth and drug treatment. The loss functions during treatment are related to pharmacokinetic functions and may be chosen according to known drug properties. Tumor properties may be simulated by the appropriate choice of cell cycle parameters. It therefore seems feasible to simulate tumor behavior for scheduled treatment with chemotherapeutic agents. Another important result of this analysis is the derivation of a fraction labelled mitosis function which incorporates the nonproliferative compartments.     

Non-hematopoietic expression of IDO is integrally required for inflammatory tumor promotion

Indoleamine 2,3-dioxygenase (IDO) is generally considered to be immunosuppressive but recent findings suggest this characterization oversimplifies its role in disease pathogenesis. Recently, we showed that IDO is essential for tumor outgrowth in the classical two-stage model of inflammatory skin carcinogenesis. Here, we report that IDO loss did not exacerbate classical inflammatory responses. Rather, IDO induction could be elicited by environmental signals and tumor promoters as an integral component of the inflammatory tissue microenvironment even in the absence of cancer. IDO loss had limited impact on tumor outgrowth in carcinogenesis models that lacked an explicit inflammatory tumor promoter. In the context of inflammatory carcinogenesis where IDO was critical to tumor development, the most important source of IDO was radiation-resistant non-hematopoietic cells, consistent with evidence that loss of the IDO regulatory tumor suppressor gene Bin1 in transformed skin cells facilitates IDO-mediated immune escape by a cell autonomous mechanism. Taken together, our results identify IDO as an integral component of ‘cancer-associated’ inflammation that tilts the immune system toward tumor support. More generally, they promote the concept that mediators of immune escape and cancer-associated inflammation may be genetically synonymous.     

Loss of the Keratin Cytoskeleton Is Not Sufficient to Induce Epithelial Mesenchymal Transition in a Novel KRAS Driven Sporadic Lung Cancer Mouse Model

Epithelial-to-mesenchymal transition (EMT), the phenotypical change of cells from an epithelial to a mesenchymal type, is thought to be a key event in invasion and metastasis of adenocarcinomas. These changes involve loss of keratin expression as well as loss of cell polarity and adhesion. We here aimed to determine whether the loss of keratin expression itself drives increased invasion and metastasis in adenocarcinomas and whether keratin loss leads to the phenotypic changes associated with EMT. Therefore, we employed a recently described murine model in which conditional deletion of the Keratin cluster II by Cre-recombinase leads to the loss of the entire keratinmultiprotein family. These mice were crossed into a newly generated Cre-recombinase inducible KRAS-driven murine lung cancer model to examine the effect of keratin loss on morphology, invasion and metastasis as well as expression of EMT related genes in the resulting tumors. We here clearly show that loss of a functional keratin cytoskeleton did not significantly alter tumor morphology or biology in terms of invasion, metastasis, proliferation or tumor burden and did not lead to induction of EMT. Further, tumor cells did not induce synchronously expression of vimentin, which is often seen in EMT, to compensate for keratin loss. In summary, our data suggest that changes in cell shape and migration that underlie EMT are dependent on changes in signaling pathways that cause secondary changes in keratin expression and organization. Thus, we conclude that loss of the keratin cytoskeleton per se is not sufficient to causally drive EMT in this tumor model.     

Mathematical Models of Gene Amplification with Applications to Cellular Drug Resistance and Tumorigenicity

An increased number of copies of specific genes may offer an advantage to cells when they grow in restrictive conditions such as in the presence of toxic drugs, or in a tumor. Three mathematical models of gene amplification and deamplification are proposed to describe the kinetics of unstable phenotypes of cells with amplified genes. The models differ in details but all assume probabilistic mechanisms of increase and decrease in gene copy number per cell (gene amplification/deamplification). Analysis of the models indicates that a stable distribution of numbers of copies of genes per cell, observed experimentally, exists only if the probability of deamplification exceeds the probability of amplification. The models are fitted to published data on the loss of methotrexate resistance in cultured cell lines, due to the loss of amplified dihydrofolate reductase gene. For two mouse cell lines unstably resistant to methotrexate the probabilities of amplification and deamplification of the dihydrofolate reductase gene on double minute chromosomes are estimated to be approximately 2% and 10%, respectively. These probabilities are much higher than widely presumed. The models explain the gradual disappearance of the resistant phenotype when selective pressure is withdrawn, by postulating that the rate of deamplification exceeds the rate of amplification. Thus it is not necessary to invoke a growth advantage of nonresistant cells which has been the standard explanation. For another analogous process, the loss of double minute chromosomes containing the myc oncogene from SEWA tumor cells, the growth advantage model does seem to be superior to the amplification and deamplification model. In a more theoretical section of the paper, it is demonstrated that gene amplification/deamplification can result in reduction to homozygosity, such as is observed in some tumors. Other applications are discussed.     

Quantification of tumor cellularity and mitotic index in invasive ductal carcinoma of the breast

OBJECTIVE: To evaluate the use of stereologically estimated tumor cell counts in the mitotic index as well as to investigate its correlation with the currently used method and test the reproducibility of the method. STUDY DESIGN: The stereologic method described by Simpson et al was used to estimate tumor cellularity in 50 invasive ductal carcinomas. Mitotic counts were also performed, and the mitotic index was calculated by the use of estimated tumor cell counts. Estimated cell counts and the mitotic index calculated were compared statistically with the actual cell counts and the traditional mitotic grades, respectively. Interobserver reproducibility of the method was also tested. RESULTS: Stereologically estimated tumor cell counts had a good correlation with actual cell counts (r = .891, P < .001). Besides, the mitotic indices calculated with tumor cell counts (calculated with both estimated and actual cell counts) in the denominator of the fraction of the mitotic index were in agreement with the currently used method (P < .01 for both). There was no statistically significant difference between the counts of two observers (P = .068). CONCLUSION: The suggested method, considering tumor cellularity as an influencing factor, was practical, reproducible and in agreement with the traditional method. This method should be studied in a large group of patients with follow-up data to determine the threshold values for different grades and determine its prognostic value during the disease course.     

LOH at 6p21.3 region and HLA class altered phenotypes in bladder carcinomas

Alterations in HLA class I antigen expression have been frequently described in different epithelial tumors and are thought to favor tumor immune escape from T lymphocyte recognition. Multiple molecular mechanisms are responsible for these altered HLA class I tumor phenotypes. Some are structural defects that produce unresponsiveness to treatment with interferons. Others include alterations in regulatory mechanisms that can be switched on by treatment of tumor cells with different cytokines. One important mechanism belonging to the first group is loss of heterozygosity (LOH) at chromosome region 6p21.3, which can lead to HLA haplotype loss. In this investigation, the frequency of LOH at 6p21 chromosome region was studied in 69 bladder carcinomas. Short tandem repeat analysis showed that 35% of cases had LOH in this chromosome region. By considering these results together with immunohistological findings previously published by our group, we identified a distribution pattern of HLA class I altered phenotypes in bladder cancer. The most frequently altered phenotype in bladder carcinomas was total loss of HLA class I expression (17 cases, 25%), followed by phenotype II associated with HLA haplotype loss (12 cases, 17.5%), and HLA allelic loss (ten cases, 14.5%). Nine cases (13%) were classified as having a compound phenotype, five cases (7%) as having HLA locus loss, and in 16 cases (23%) no alteration in HLA expression was detected. An important conclusion of this report is that a combination of different molecular and immunohistological techniques is required to precisely define which HLA alleles are lost during tumor progression and to characterize the underlying mechanisms of these losses. These studies should be performed when a cancer patient is to be included in an immunotherapy protocol that aims to stimulate different immune effector mechanisms.     

Osteocyte biology: its implications for osteoporosis

Osteocyte viability may play a significant role in the maintenance and integrity of bone. Bone loss due to osteoporosis may be due in part to osteocyte cell death. We have identified a factor that will protect both osteoblasts and osteocytes from cell death due to agents known to be responsible for various forms of osteoporosis. Not only does estrogen preserve osteoblast and osteocyte viability, but so does a molecule called CD40Ligand. This molecule is expressed on activated T lymphocytes, human dendritic cells, and human vascular endothelial cells, whereas its receptor CD40 is expressed on normal epithelium, B cells, and dendritic cells. CD40Ligand protects osteoblasts and the MLO-Y4 osteocyte-like cells against apoptosis induced by glucocorticoids, tumor necrosis factor alpha or etoposide. As tumor necrosis factor a has been shown to be responsible for post-menopausal bone loss and glucocorticoids induce dramatic bone loss, this finding has important implications with regards to potential therapy for both post-menopausal and steroid-induced osteoporosis.     

Genetic mechanisms of tumor-specific loss of 11p DNA sequences in Wilms tumor.

Wilms tumor, a common childhood renal tumor, occurs in both a heritable and a nonheritable form. The heritable form may occasionally be attributed to a chromosome deletion at 11p13, and tumors from patients with normal constitutional chromosomes often show deletion or rearrangement of 11p13. It has been suggested that a germinal or somatic mutation may occur on one chromosome 11 and predispose to Wilms tumor and that a subsequent somatic genetic event on the normal homologue at 11p13 may permit tumor development. To study the frequency and mechanism of such tumor-specific genetic events, we have examined the karyotype and chromosome 11 genotype of normal and tumor tissues from 13 childhood renal tumor patients with different histologic tumor types and associated clinical conditions. Tumors of eight of the 12 Wilms tumor patients, including all viable tumors examined directly, show molecular evidence of loss of 11p DNA sequences by somatic recombination (four cases), chromosome loss (two cases), and recombination (two cases) or chromosome loss and duplication. One malignant rhabdoid tumor in a patient heterozygous for multiple 11p markers did not show any tumor-specific 11p alteration. These findings confirm the critical role of 11p sequences in Wilms tumor development and reveal that mitotic recombination may be the most frequent mechanism by which tumors develop.     

Paradoxical role of apoptosis in tumor progression

Tumors frequently acquire resistance to apoptosis that is expected to contribute to malignant phenotype and reduce sensitivity to treatment. In fact, inactivation of p53 tumor suppressor gene resulting in suppression of apoptosis serves as a negative prognostic marker. Surprisingly, expression of a strong anti-apoptotic protein Bcl-2, another mechanism to avoid apoptosis, was found to be associated with a favorable prognosis. This paradoxical anti-progressor function of Bcl-2 has been explained in literature based on the negative effect of Bcl-2 on cell proliferation. Here, by analyzing accumulated experimental and clinical data, we provide evidence supporting another hypothesis that defines apoptosis as an accelerator of tumor progression. The mechanism of anti-progressor function of Bcl-2 is based on creation of tumors that maintain control of genomic stability by eliminating selective advantages for the cells that acquire resistance to apoptosis through loss of p53. Thus, inhibition of apoptosis does not lead to loss of genomic stability and creates tumor environment that no longer supports further tumor progression and inhibitors of apoptosis can be considered as factors suppressing tumor progression.     

Loss of PTEN expression in neuroendocrine pancreatic tumors

PTEN (phosphatase and tensin homologue deleted from chromosome 10) is a well established tumor suppressor gene, which was cloned to chromosome 10q23. PTEN plays an important role in controlling cell growth, apoptosis, cell adhesion, and cell migration. In various studies, a genetic change as well as loss of PTEN expression by different carcinomas has been described. To date, the role of PTEN as a differentiation marker for neuroendocrine tumors (NET) and for the loss of PTEN expression is still unknown. It is assumed that loss of PTEN expression is important for tumor progression of NETs. We hypothesize that PTEN might be used as a new prognostic marker. We report 38 patients with a NET of the pancreas. Tumor tissues were surgically resected, fixed in formalin, and embedded in paraffin. PTEN expression was evaluated by immunohistochemistry and was correlated with several clinical and pathological parameters of each individual tumor. After evaluation of our immunohistochemistry data using a modified Remmele Score, a widely accepted method for categorizing staining results for reports and statistical evaluation, staining results of PTEN expression were correlated with the clinical and pathological parameters of each individual tumor. Our data demonstrates a significant difference in survival with existence of lymph node or distant metastases. Negative patients show a significant better survival compared with positive patients. Furthermore, we show a significant difference between PTEN expression and WHO or TNM classification. Taken together, our data shows a positive correlation between WHO classification and the new TNM classification of NETs, and loss of PTEN expression as well as survival. These results strongly implicate that PTEN might be helpful as a new prognostic factor.     

Yes-associated protein (YAP) functions as a tumor suppressor in breast

Yes-associated protein (YAP) has been shown to positively regulate p53 family members and to be negatively regulated by the AKT proto-oncogene product in promoting apoptosis. On the basis of this function and its location at 11q22.2, a site of frequent loss of heterozygosity (LOH) in breast cancer, we investigated whether YAP is a tumor suppressor in breast. Examination of tumors by immunohistochemistry demonstrated significant loss of YAP protein. LOH analysis revealed that protein loss correlates with specific deletion of the YAP gene locus. Functionally, short hairpin RNA knockdown of YAP in breast cell lines suppressed anoikis, increased migration and invasiveness, inhibited the response to taxol and enhanced tumor growth in nude mice. This is the first report indicating YAP as a tumor suppressor, revealing its decreased expression in breast cancer as well as demonstrating the functional implications of YAP loss in several aspects of cancer signaling.     

Selective loss of HLA-A or HLA-B antigen expression in colon carcinoma

A panel of colorectal carcinomas has been examined immunohistologically with mAb specific for allodeterminants of HLA-A or HLA-B, and with mAb reactive with HLA-A,B,C framework determinants or beta 2-microglobulin. Out of 85 carcinomas, 5 cases were completely negative for all class I Ag in the tumor cell population. Fifteen cases exhibited a differential expression of HLA-A and HLA-B products in all tumor cells or in a subset. Such tumor cells showed strong staining with monomorphic HLA antibodies but failed to stain with certain allospecific mAb. Negativity of tumor cells was scored only when normal stromal cells within the tumor mass were clearly labeled, indicating expression of the particular allodeterminant(s) in the given tissue. Four carcinomas showed a selective loss of HLA-A2 or a non-A2 HLA-A allospecificity in all tumor cells or a major subset, whereas HLA-B Ag were expressed. Seven cases showed a selective loss of the HLA-Bw6 superspecificity in the tumor cell population; HLA-Bw4 and HLA-A Ag were present. Three cases exhibited a combined loss of HLA-A2 and HLA-Bw6 Ag, with non-A2 HLA-A and HLA-Bw4 Ag being expressed. A further case was HLA-Bw6/Bw4 negative in a major tumor cell subset and HLA-A negative in the complementary minor subset. The results show that a considerable proportion of colorectal carcinomas is heterogeneous with regard to HLA-A and HLA-B expression. The reasons for the appearance of tumor subsets with complete loss of class I Ag or selective loss of only HLA-A or HLA-B Ag are not clear, but it is conceivable that some of them arose because they have escaped immunoselection.