The retinoblastoma tumor-suppressor gene, the exception that proves the rule.

The retinoblastoma tumor-suppressor gene (Rb1) is centrally important in cancer research. Mutational inactivation of Rb1 causes the pediatric cancer retinoblastoma, while deregulation of the pathway in which it functions is common in most types of human cancer. The Rb1-encoded protein (pRb) is well known as a general cell cycle regulator, and this activity is critical for pRb-mediated tumor suppression. The main focus of this review, however, is on more recent evidence demonstrating the existence of additional, cell type-specific pRb functions in cellular differentiation and survival. These additional functions are relevant to carcinogenesis suggesting that the net effect of Rb1 loss on the behavior of resulting tumors is highly dependent on biological context. The molecular mechanisms underlying pRb functions are based on the cellular proteins it interacts with and the functional consequences of those interactions. Better insight into pRb-mediated tumor suppression and clinical exploitation of pRb as a therapeutic target will require a global view of the complex, interdependent network of pocket protein complexes that function simultaneously within given tissues.

Cell cycle regulation of c-Jun N-terminal kinase activity at the centrosomes.

The c-Jun N-terminal kinase (JNK), a subgroup of the mitogen-activated protein kinase (MAPK) family of serine/threonine kinases, has established functions in cell growth and apoptosis. While the mechanisms are unclear, JNK has also been also implicated in signaling pathways that initiate cell cycle checkpoints and cell cycle progression. By following the localization of active and inactive JNK during the cell cycle, we have found that the majority of cellular JNK is soluble and present in the cytoplasm and the nucleus. Interestingly, insoluble fractions of JNK are also localized in nuclear and cytoplasmic speckles, and to the centrosomes. While JNK is associated with the centrosome throughout the cell cycle, it is only active at the centrosome from S phase through anaphase. This novel localization of centrosomal JNK is a possible link between JNK-activating stimuli and centrosome or cell cycle events.

The nuclear death domain protein p84N5 activates a G2/M cell cycle checkpoint prior to the onset of apoptosis.

In contrast to extracellular signals, the mechanisms utilized to transduce nuclear apoptotic signals are not well understood. Characterizing these mechanisms is important for predicting how tumors will respond to genotoxic radiation or chemotherapy. The retinoblastoma (Rb) tumor suppressor protein can regulate apoptosis triggered by DNA damage through an unknown mechanism. The nuclear death domain-containing protein p84N5 can induce apoptosis that is inhibited by association with Rb. The pattern of caspase and NF-kappaB activation during p84N5-induced apoptosis is similar to p53-independent cellular responses to DNA damage. One hallmark of this response is the activation of a G(2)/M cell cycle checkpoint. In this report, we characterize the effects of p84N5 on the cell cycle. Expression of p84N5 induces changes in cell cycle distribution and kinetics that are consistent with the activation of a G(2)/M cell cycle checkpoint. Like the radiation-induced checkpoint, caffeine blocks p84N5-induced G(2)/M arrest but not subsequent apoptotic cell death. The p84N5-induced checkpoint is functional in ataxia telangiectasia-mutated kinase-deficient cells. We conclude that p84N5 induces an ataxia telangiectasia-mutated kinase (ATM)-independent, caffeine-sensitive G(2)/M cell cycle arrest prior to the onset of apoptosis. This conclusion is consistent with the hypotheses that p84N5 functions in an Rb-regulated cellular response that is similar to that triggered by DNA damage.

Adenovirus-mediated N5 gene transfer inhibits tumor cell proliferation by induction of apoptosis.

Gene therapy designed to initiate apoptotic cell death provides a potentially effective method to treat cancer. A prerequisite for this approach is the identification of genes that function in distinct apoptotic pathways. Although apoptotic pathways initiated by receptors such as tumor necrosis factor receptor-1 are well characterized, little is known about apoptotic pathways initiated within the nucleus in response to genotoxic stress. We have demonstrated previously that the nuclear, death domain-containing protein p84N5 can induce apoptosis upon transfection into cells, suggesting that it may play a role in an apoptotic pathway initiated within the nucleus. To test the possibility that N5 could be used in the gene therapy of cancer, we have generated a recombinant adenovirus engineered to express N5 and tested the effects of viral infection on the growth and tumorigenicity of tumor cells. N5 adenovirus infection significantly reduced the proliferation and tumorigenicity of breast, ovarian, and osteosarcoma tumor cell lines. Reduced proliferation and tumorigenicity were mediated by an induction of apoptosis as indicated by DNA fragmentation in infected cells. The results suggest that the N5 cDNA is a candidate for the gene therapy of cancer.

Apoptosis induced by the nuclear death domain protein p84N5 is associated with caspase-6 and NF-kappa B activation.

Although the mechanisms involved in responses to extracellular or mitochondrial apoptotic signals have received considerable attention, the mechanisms utilized within the nucleus to transduce apoptotic signals are not well understood. We have characterized apoptosis induced by the nuclear death domain-containing protein p84N5. Adenovirus-mediated N5 gene transfer or transfection of p84N5 expression vectors induces apoptosis in tumor cell lines with nearly 100% efficiency as indicated by cellular morphology, DNA fragmentation, and annexin V staining. Using peptide substrates and Western blotting, we have determined that N5-induced apoptosis is initially accompanied by activation of caspase-6. Activation of caspases-3 and -9 does not peak until 3 days after the peak of caspase-6 activity. Expression of p84N5 also leads to activation of NF-kappaB as indicated by nuclear translocation of p65RelA and transcriptional activation of a NF-kappaB-dependent reporter promoter. Changes in the relative expression level of Bcl-2 family proteins, including Bak and Bcl-Xs, are also observed during p84N5-induced apoptosis. Finally, we demonstrate that p84N5-induced apoptosis does not require p53 and is not inhibited by p53 coexpression. We propose that p84N5 is involved in an apoptotic pathway distinct from those triggered by death domain-containing receptors or by p53.

Glutamic acid mutagenesis of retinoblastoma protein phosphorylation sites has diverse effects on function.

The retinoblastoma tumor suppressor gene (Rb) has many functions within the cell including regulation of transcription, differentiation, apoptosis, and the cell cycle. Regulation of these functions is mediated by phosphorylation at as many as 16 cyclin-dependent kinase (CDK) phosphorylation sites in vivo. The contribution of these sites to the regulation of the various Rb functions is not well understood. To characterize the effect of phosphorylation at these sites, we systematically mutagenized the serines or threonines to glutamic acid. Thirty-five mutants with different combinations of modified phosphorylation sites were assayed for their ability to arrest the cell cycle and for their potential to induce differentiation. Only the most highly substituted mutants failed to arrest cell cycle progression. However, mutants with as few as four modified phosphorylation sites were unable to promote differentiation. Other mutants had increased activity in this assay. We conclude that modification of Rb phosphorylation sites can increase or decrease protein activity, that different Rb functions can be regulated independently by distinct combinations of sites, and that the effects of modification at any one site are context dependent.

Apoptosis induced by the nuclear death domain protein p84N5 is inhibited by association with Rb protein.

Rb protein inhibits both cell cycle progression and apoptosis. Interaction of specific cellular proteins, including E2F1, with Rb C-terminal domains mediates cell cycle regulation. In contrast, the nuclear N5 protein associates with an Rb N-terminal domain with unknown function. The N5 protein contains a region of sequence similarity to the death domain of proteins involved in apoptotic signaling. We demonstrate here that forced N5 expression potently induces apoptosis in several tumor cell lines. Mutation of conserved residues within the death domain homology compromise N5-induced apoptosis, suggesting that it is required for normal function. Endogenous N5 protein is specifically altered in apoptotic cells treated with ionizing radiation. Furthermore, dominant interfering death domain mutants compromise cellular responses to ionizing radiation. Finally, physical association with Rb protein inhibits N5-induced apoptosis. We propose that N5 protein plays a role in the regulation of apoptosis and that Rb directly coordinates cell proliferation and apoptosis by binding specific proteins involved in each process through distinct protein binding domains.

Histone H3 phosphorylation is required for the initiation, but not maintenance, of mammalian chromosome condensation.

The temporal and spatial patterns of histone H3 phosphorylation implicate a specific role for this modification in mammalian chromosome condensation. Cells arrest in late G2 when H3 phosphorylation is competitively inhibited by microinjecting excess substrate at mid-S-phase, suggesting a requirement for activity of the kinase that phosphorylates H3 during the initiation of chromosome condensation and entry into mitosis. Basal levels of phosphorylated H3 increase primarily in late-replicating/early-condensing heterochromatin both during G2 and when premature chromosome condensation is induced. The prematurely condensed state induced by okadaic acid treatment during S-phase culminates with H3 phosphorylation throughout the chromatin, but in an absence of mitotic chromosome morphology, indicating that the phosphorylation of H3 is not sufficient for complete condensation. Mild hypotonic treatment of cells arrested in mitosis results in the dephosphorylation of H3 without a cytological loss of chromosome compaction. Hypotonic-treated cells, however, complete mitosis only when H3 is phosphorylated. These observations suggest that H3 phosphorylation is required for cell cycle progression and specifically for the changes in chromatin structure incurred during chromosome condensation.

Cyclin D1/Cdk4 regulates retinoblastoma protein-mediated cell cycle arrest by site-specific phosphorylation.

The retinoblastoma protein (pRb) inhibits progression through the cell cycle. Although pRb is phosphorylated when G1 cyclin-dependent kinases (Cdks) are active, the mechanisms underlying pRb regulation are unknown. In vitro phosphorylation by cyclin D1/Cdk4 leads to inactivation of pRb in a microinjection-based in vivo cell cycle assay. In contrast, phosphorylation of pRb by Cdk2 or Cdk3 in complexes with A- or E-type cyclins is not sufficient to inactivate pRb function in this assay, despite extensive phosphorylation and conversion to a slowly migrating "hyperphosphorylated form." The differential effects of phosphorylation on pRb function coincide with modification of distinct sets of sites. Serine 795 is phosphorylated efficiently by Cdk4, even in the absence of an intact LXCXE motif in cyclin D, but not by Cdk2 or Cdk3. Mutation of serine 795 to alanine prevents pRb inactivation by Cdk4 phosphorylation in the microinjection assay. This study identifies a residue whose phosphorylation is critical for inactivation of pRb-mediated growth suppression, and it indicates that hyperphosphorylation and inactivation of pRb are not necessarily synonymous.

G1 arrest and down-regulation of cyclin E/cyclin-dependent kinase 2 by the protein kinase inhibitor staurosporine are dependent on the retinoblastoma protein in the bladder carcinoma cell line 5637.

The protein kinase inhibitor staurosporine has been shown to induce G1 phase arrest in normal cells but not in most transformed cells. Staurosporine did not induce G1 phase arrest in the bladder carcinoma cell line 5637 that lacks a functional retinoblastoma protein (pRB-). However, when infected with a pRB-expressing retrovirus [Goodrich, D. W., Chen, Y., Scully, P. & Lee, W.-H. (1992) Cancer Res. 52, 1968-1973], these cells, now pRB+, were arrested by staurosporine in G1 phase. This arrest was accompanied by the accumulation of hypophosphorylated pRB. In both the pRB+ and pRB- cells, cyclin D1-associated kinase activities were reduced on staurosporine treatment. In contrast, cyclin-dependent kinase (CDK) 2 and cyclin E/CDK2 activities were inhibited only in pRB+ cells. Staurosporine treatment did not cause reductions in the protein levels of CDK4, cyclin D1, CDK2, or cyclin E. The CDK inhibitor proteins p21(Waf1/Cip1) and p27 (Kip1) levels increased in staurosporine-treated cells. Immunoprecipitation of CDK2, cyclin E, and p2l from staurosporine-treated pRB+ cells revealed a 2.5- to 3-fold higher ratio of p2l bound to CDK2 compared with staurosporine-treated pRB- cells. In pRB+ cells, p2l was preferentially associated with Thrl6O phosphorylated active CDK2. In pRB- cells, however, p2l was bound preferentially to the unphosphorylated, inactive form of CDK2 even though the phosphorylated form was abundant. This is the first evidence suggesting that G1 arrest by 4 nM staurosporine is dependent on a functional pRB protein. Cell cycle arrest at the pRB- dependent checkpoint may prevent activation of cyclin E/CDK2 by stabilizing its interaction with inhibitor proteins p2l and p27.

Molecular characterization of the retinoblastoma susceptibility gene.

Retinoblastoma is recognized as a hereditary cancer. Genetic and epidemiological analysis of the disease has been incorporated into a two-hit mutational inactivation hypothesis of the origin of retinoblastoma. The molecular cloning and characterization of the retinoblastoma gene and gene product has allowed a critical testing of this two-hit hypothesis. All the predications of the model have been born out by experiment so far. These include inheritance of one mutated RB allele as the origin of hereditary retinoblastoma, subsequent loss of the remaining allele upon tumorigenesis, the involvement of the same RB gene in both sporadic and hereditary retinoblastoma, the somatic mutation of both RB alleles in sporadic retinoblastoma, the lack of RB expression in any retinoblastoma yet examined, and the recessiveness of mutated RB alleles. The RB gene exhibits functional properties consistent with its role as a suppressor of tumor formation. For example, re-expression of RB in tumor cells lacking endogenous RB leads to a loss of tumorigenic properties. RB protein can also inhibit progression through the cell division cycle, and it physically and/or functionally interacts with important cell cycle regulatory molecules. Although confirmation of the two-hit hypothesis seems complete, we can not rule out the possibility that other genes are involved in the genesis of this tumor. For example, there seems to be variable resistance to tumor development even in patients inheriting retinoblastoma susceptibility. Further, heterozygous RB null mice do not develop retinoblastoma, but develop a characteristic brain tumor instead. The molecular isolation of the RB gene is an important achievement in research on cancer. For the first time, it has become possible to examine, at the molecular level, genes that inhibit the growth of tumor cells. The precise mechanism of action of RB is unknown, but a broad outline is beginning to emerge. RB seems to negatively influence tumor cell growth by participating in regulation of the cell division cycle. RB has also been implicated in differentiation; its effect on the cell division cycle and its effects on differentiation may be different manifestations of the same function. Since RB is involved in oncogenesis, gene regulation, and cellular differentiation, it is obviously an attractive gene for intense study; understanding the function and mechanism of action of RB will impact the understanding of many, important cell processes.

Abrogation by c-myc of G1 phase arrest induced by RB protein but not by p53.

Inactivating mutations of the retinoblastoma gene (RB) are found in a wide variety of tumour cells. Replacement of wild-type RB can suppress the tumorigenicity of some of these cells, suggesting that the RB protein (Rb) may negatively regulate cell growth. As activation of c-myc expression promotes cell proliferation and blocks differentiation, it may positively regulate cell growth. The c-myc protein is localized in the nucleus and can physically associate with RB protein in vitro, hence c-myc may functionally antagonize RB function. Microinjection of Rb in G1 phase reversibly arrests cell-cycle progression. Here we co-inject RB protein with c-myc, EJ-ras, c-fos or c-jun protein. Co-injection of c-myc, but not EJ-ras, c-fos or c-jun, inhibits the ability of Rb to arrest the cell cycle. The c-myc does not inhibit the activity of another tumour supressor, p53 (ref. 12). Thus, c-myc and RB specifically antagonize one another in the cell.

Expression of the retinoblastoma gene product in bladder carcinoma cells associates with a low frequency of tumor formation.

Upon inactivation of both alleles of the retinoblastoma gene (RB), individuals develop the intraocular eye tumor, retinoblastoma. The gene encodes a Mr 110,000 phosphorylated nuclear protein that may be involved in regulation of the cell cycle. Besides retinoblastoma, mutations of the gene have been detected in several other types of tumors, including bladder carcinoma. Up to one-third of bladder carcinomas may contain mutations of the RB gene. Introducing the retinoblastoma gene into single retinoblastoma, osteosarcoma, or prostate carcinoma cell lines suppresses their tumorigenicity as assayed in nude mice. We have sought to extend these results by introducing the retinoblastoma gene into multiple bladder carcinoma lines, and analyzing several of the resulting, cloned lines. We have found that inhibition of tumorigenicity, as assayed by tumor growth in nude mice or growth of cells in soft agar, is the only consistent phenotype observed upon re-expression of RB in all bladder carcinoma cells examined. The effect of RB expression on growth and cellular morphology varied depending on the particular parental cell line. We conclude that RB expression generally correlates with reduced tumorigenicity, but not reduced growth rate, in bladder carcinoma cells.

The retinoblastoma gene product regulates progression through the G1 phase of the cell cycle.

The RB gene product is a nuclear phosphoprotein that undergoes cell cycle-dependent changes in its phosphorylation status. To test whether RB regulates cell cycle progression, purified RB proteins, either full-length or a truncated form containing the T antigen-binding region, were injected into cells. Injection of either protein early in G1 inhibits progression into S phase. Co-injection of anti-RB antibodies antagonizes this effect. Injection of RB into cells arrested at G1/S or late in G1 has no effect on BrdU incorporation, suggesting that RB does not inhibit DNA synthesis in S phase. These results indicate that RB regulates cell proliferation by restricting cell cycle progression at a specific point in G1 and establish a biological assay for RB activity. Neither co-injection of RB with a T antigen peptide nor injection into cells expressing T antigen prevents cells from progressing into S phase, which supports the hypothesis that T antigen binding has functional consequences for RB.

Evidence that retroviral transduction is mediated by DNA not by RNA.

Retroviral transduction of cellular nucleic acid sequences requires illegitimate RNA or DNA recombination. To test a model that postulates transduction via efficient illegitimate recombination during reverse transcription of viral and cellular RNAs, we have measured the ability of Harvey sarcoma viruses (HaSVs) with artificial 3' termini to recover a retroviral 3' terminus from helper Moloney virus (MoV) by illegitimate and homologous recombination. For this purpose, mouse NIH 3T3 cells were transformed with Harvey proviruses and then superinfected with MoV. The proviruses lacked the 3' long terminal repeat and an untranscribed region of the 5' long terminal repeat to prevent virus regeneration from input provirus. Only 0-11 focus-forming units of HaSV were generated upon MoV superinfection of 3 x 10(6) cells transformed by Harvey proviruses with MoV-unrelated termini. This low frequency is consistent with illegitimate DNA recombination via random Moloney provirus integration 3' of the transforming viral ras gene in the 10(6)-kilobase mouse genome. When portions of murine viral envelope (env) genes were attached 3' of ras, 10(2)-10(5) focus-forming units of HaSV were generated, depending on the extent of homology with env of MoV. These recombinants all contained HaSV-specific sequences 5' and MoV-specific sequences 3' of the common env homology. They were probably generated by recombination during reverse transcription rather than by recombination among either input or secondary proviruses, since (i) the yield of recombinants was reduced by a factor of 10 when the env sequence was flanked by splice signals and (ii) HaSV RNAs without retroviral 3' termini would be inadequate templates for provirus synthesis. We conclude that there is no efficient illegitimate recombination in retroviruses. In view of known precedents of illegitimate DNA recombination, the structure of known viral onc genes, and our evidence for illegitimate DNA recombination via provirus integration, we favor the DNA model of transduction over the RNA model.

Retroviral recombination during reverse transcription.

After mixed infection, up to half of related retroviruses are recombinants. During infection, retroviral RNA genomes are first converted to complementary DNA (cDNA) and then to double-stranded DNA. Thus recombination could occur during reverse transcription, by RNA template switching, or after reverse transcription, by breakage and reunion of DNA. It has not been possible to distinguish between these two potential mechanisms of recombination because both single-stranded cDNA and double-stranded proviral DNA exist in infected cells during the eclipse period. Therefore we have analyzed for recombinant molecules among cDNA products transcribed in vitro from RNA of disrupted virions. Since recombinants from allelic parents can only be distinguished from parental genomes by point mutations, we have examined the cDNAs from virions with distinct genetic structures for recombinant-specific size and sequence markers. The parents share a common internal allele that allows homology-directed recombination, but each contains specific flanking sequences. One parent is a synthetically altered Harvey murine sarcoma virus RNA that lacks a retroviral 3' terminus but carries a Moloney murine retrovirus-derived envelope gene (env) fragment 3' of its transforming ras gene. The other parent is intact Moloney virus. Using a Harvey-specific 5' primer and a Moloney-specific 3' primer, we have found recombinant cDNAs with the polymerase chain reaction, proving directly that retroviruses can recombine during reverse transcription unassisted by cellular enzymes, probably by template switching during cDNA synthesis. The recombinants that were obtained in vitro were identical with those obtained in parallel experiments in vivo.

The molecular genetics of retinoblastoma.

Retinoblastoma is a potentially hereditary cancer. Refinement of the genetic and epidemiological analysis of the disease has uncovered two distinct classes of retinoblastoma. Sporadic retinoblastoma is generally unilateral and unifocal, and is diagnosed at the late age of about two years. A few of these sporadic cases are probably due to a germ cell mutation inherited from a parent and hence can be classified as hereditary. Familial retinoblastoma is generally diagnosed at an earlier age, at 11 months, and is typically bilateral and/or multifocal. These observations have been incorporated into a 'two hit' mutational inactivation hypothesis of the origin of retinoblastoma. The molecular cloning and characterization of a candidate retinoblastoma susceptibility gene and its gene product has allowed a critical testing of this hypothesis. All of the predications of the model have been confirmed by experiment. These include inheritance of one mutated retinoblastoma susceptibility (RB) allele as the origin of hereditary retinoblastoma, subsequent loss of the remaining allele upon the genesis of the tumour, the involvement of the same RB gene in both sporadic and hereditary retinoblastoma, the somatic mutation of both RB alleles in sporadic retinoblastoma, the lack of evidence for expression of a normal RB gene product in any retinoblastoma yet examined, the inactivational nature of RB mutations and the recessiveness of these mutated alleles. The RB gene also exhibits suppression of neoplastic properties when introduced into retinoblastoma cells and also into some other tumour cells. These results mutually reinforce the two hit inactivation hypothesis as well as the cloned gene's correct identification as the retinoblastoma susceptibility locus. The confirmation of this hypothesis is, therefore, nearing completion. The definitive proof is achievable with the advent of chimeric mouse technology, which will allow construction of mice with one or both RB alleles that have been inactivated by mutation. Analysis of such mice may allow us to determine if inactivation of both RB alleles is necessary and sufficient for the development of retinoblastoma and possibly other tumour types. The molecular isolation of the RB gene is an important achievement in research on cancer. For the first time, it has become possible to examine, at the molecular level, genes which suppress the tumorigenicity of cancer cells. Analysis of such cloned genes should yield insight into mechanisms of oncogenesis, gene regulation and cellular differentiation complementary to the knowledge which has long been accumulating from the study of oncogenes.

Retroviral transduction of oncogenic sequences involves viral DNA instead of RNA.

We have studied whether the origin of retroviral onc genes, by transduction of sequences from cellular proto-onc genes, involves DNA or RNA recombination. By using altered Harvey sarcoma proviruses as models for transduction intermediates, we have investigated the mechanism of regeneration of transforming virus from truncated proviruses with only a single 5' long terminal repeat (LTR) but with a complete 5'-LTR-ras transforming gene. The Harvey proviruses were specifically altered to discriminate between virus regeneration by RNA template switching during reverse transcription, as has been postulated, and virus regeneration by DNA recombination with either helper virus or among elements of the defective provirus alone. For this purpose U3 elements of the Harvey proviral LTR, which are essential for replication but not for transcription, were deleted in vitro. Only proviral constructions with an intact or a nearly intact single LTR regenerated infectious Harvey sarcoma virus. Since all constructions transformed cells and produced identical RNAs, our results exclude a model of virus regeneration by switching of RNA templates during reverse transcription. We conclude that regeneration of infectious Harvey viruses from truncated provirus involved illegitimate recombination of cellular or cotransfected DNAs flanking the 5'-LTR-ras gene of Harvey sarcoma virus. Based on this and evidence from the literature, we propose that retroviral transduction proceeds by way of rare illegitimate recombinations between proviral and cellular DNAs.