Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

Lung cancer is the result of deregulation (activation or inactivation) of specific genes involved in the process of proliferation, differentiation, apoptosis, and DNA repair. Lung neuroendocrine (NE) tumors embrace a spectrum of four clinicopathologic entities with varying degrees of differentiation and aggressiveness. Typical carcinoids are low-grade, well-differentiated NE tumors with excellent prognosis. Atypical carcinoids, as described by Arrigoni and coworkers (1), keep morphologic characteristics of NE differentiation but show foci of necrosis and/or mitotic activity (more than 5 mitoses per 10 high-power fields [HPF]). High-grade NE lung tumors include small-cell lung carcinoma (SCLC) and large-cell neuroendocrine carcinoma (LCNEC) with architectural maintenance of NE pattern, the expression of NE markers but marked cellular pleomorphism, prominent necrosis, and high mitotic activity (2). SCLC is the less differentiated and the most poorly aggressive tumor type at the end of this spectrum.

It is estimated that about 10 to 20 mutations occur during the pathogenesis of lung cancers (3). Among these mutations, inactivation of tumor suppressor genes seems to be a necessary step in the process of tumorigenesis, especially inactivation of p53 and retinoblastoma (RB) genes. One of these genes, the RB gene, whose protein product is the main terminal effector of p53-dependent G1 arrest, plays a major role in the control of cell cycle progression and seems to be an obligatory target in high-grade NE lung tumors (4). The RB gene encodes a nuclear phosphoprotein of 110 kD (7) whose phosphorylation level changes depending on cell cycle: low in G1 and high in S phase and early mitosis (8). This phosphorylation is mediated by specific cyclin-cyclin-dependent kinase (cdk) complexes such as cyclin D1-cdk4 in G1 phase and cyclin E-cdk2 at the G1-S transition (9, 10), the activity of which is regulated by cdk inhibitors such as p21, p16^INK4, and p15^INK4 (11, 12). The hypophosphorylated form of the Rb protein has been shown to be inactivated through binding to the transforming proteins of several DNA tumor viruses such as the large T antigen of simian virus 40 (SV40), E1A of adenovirus, and E7 of human papillomavirus (13- 15). Furthermore, the Rb protein is also able to interact with a set of cellular transcription factors such as E2F (16), DRTF1 (17), c-myc, and N-myc (18). All of these proteins bind to a specific domain in the Rb protein called the RB pocket; this portion of the RB protein is encoded by exons 13 to 18 and 20 to 24 of the RB gene (19, 20). Whereas the hypophosphorylated form of Rb protein can bind to the transcription factor E2F, blocking the cells in G1 phase (16), the phosphorylation of Rb protein leads to the dissociation of this complex, allowing cell cycle progression (21). Thus Rb protein acts as a key protein in cell cycle control at the G1 checkpoint, explaining its function as a tumor suppressor gene. RB inactivation can result at a genetic level from mutation of its DNA sequence or defect of gene transcription, or at epigenetic level through inappropriate phosphorylation. Both p16^INK4 inactivation (22) and cyclin D1 overexpression (23) are responsible for RB metabolic inactivation.

The RB gene is inactivated in all retinoblastomas indicating its major role in the pathogenesis of this cancer (24). Knock-out RB transgenic mice die in utero with multiple defects before the sixteenth embryonic day (25); animals heterozygous for RB function fail to develop retinal tumors as in humans, but have a high incidence of pituitary adenomas (25). These discordances are probably due to tissue specificity of the functional redundancy of the Rb family products (p107 and pRb2/p130) in these mouse models (28). Inactivation of both p53 and RB genes is observed in 70% of SCLCs whereas mice deficient in both p53 and RB develop tumors, primarily of endocrine origin, and bronchial neuroendocrine hyperplasia, but fail to show neuroendocrine lung tumors (29). These experiments indicate the phenotypic differences observed between mice and humans with analogous genetic mutations, and the limits of the models of human diseases in the mouse (28). Human tumors are thus necessary to study, not only using cell lines but also and especially freshly excised tumor tissues, in order to understand the specific role of the RB gene in human carcinogenesis.

Inactivation of the RB gene is involved in the carcinogenesis of lung cancers (4, 5) and especially NE lung tumors (4, 30, 31). Loss of RB protein is observed in 90% of SCLC cell lines (5, 30, 32). In the majority of these cases, this inactivation comes under the loss of one allele (33) and the inactivation of the remaining allele by an unknown mechanism leading to loss of mRNA expression (4, 30, 32). Major structural abnormalities are rare (4, 30). In contrast with neuroendocrine carcinomas, Rb is mainly inactivated through loss of p16 function (33, 34) and/or constitutive cyclin D1 activation (35) in nonneuroendocrine carcinomas.

Mechanisms of RB gene inactivation have been investigated in only a few cases of freshly excised SCLCs (4, 30). We have previously shown that 87% of high-grade NE carcinomas exhibited an abnormal expression of Rb protein that was highly correlated with a loss of heterozygosity at the RB locus (36). To analyze the role of the RB gene as a tumor suppressor in the carcinogenesis of NE lung tumors of various grades of differentiation and malignancy, and to try to clarify the molecular mechanism that led to altered expression of Rb protein, genetic abnormalities and expression of the RB gene were investigated in a large series of freshly excised NE lung carcinomas of various histologic types. Our results showed that inactivation of Rb gene is significantly linked with a high grade of malignancy in NE tumors (SCLCs and LCNECs), in keeping with that observed in SCLC cell lines, and seems unnecessary for the development of typical carcinoids.

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

Resected tumors were obtained from 37 patients with NE lung carcinomas. Tumor samples were fixed and embedded in paraffin for conventional microscopic analysis and histologic classification. A sample was immediately frozen in liquid nitrogen at surgery and stored at 80°C until use. Histologic classification of all specimens was assessed according to the World Health Organization (WHO) classification of lung cancer (1981). These 37 neuroendocrine carcinomas included 8 carcinoids, 1 atypical carcinoid, 7 LCNECs as described by Travis and coworkers (2), and 21 classic SCLCs. Neuroendocrine markers have been studied using chromogranin A, leu7, synaptophysin, and anti-N-CAM (2, 37) and at least three of them were expressed in every case. Thirty normal lung tissue samples corresponding to 30 tumors analyzed were also studied.

Immunohistochemical Analysis

Immunohistochemical detection of the Rb protein was performed on frozen tumor sections. Five specific primary Rb antibodies (Rb1F8, Rb1 [clone 84-b3-1], C36, PMG3-245, and G99-549), which recognized the unphosphorylated form were used (Table 1). Immunostaining was performed as previously described (36).

Southern Blotting Analysis

DNA was extracted by a standard method as previously described (36). DNAs (20 µg) were digested with HindIII restriction endonuclease for structural abnormality analysis and with SacI-SacII restriction endonucleases for hypermethylation analysis and electrophoresed on a 0.7% agarose gel. They were then transferred to a nylon membrane (Hybond N; Amersham, Buckinghamshire, United Kingdom) according to the method of Southern (38). The probes used for DNA structural abnormality study were the 3.8- and 0.9-kb EcoRI fragments (called RB 3.8 kb and 0.9 kb) of the cloned 4.7-kb RB gene cDNA (39). The p123M1.8 probe (1.8-kb EcoRI-HindIII fragment) was used for hypermethylation analysis (40). Probes were radiolabeled by random priming using [-³²P]dCTP. Hybridization was performed at 42°C in 50% formamide over 12 h and the membranes were washed twice for 30 min in 2× SSC (1× SSC is 0.15 M NaCl plus 0.015 M sodium citrate)-0.1% sodium dodecyl sulfate (SDS) at room temperature and for a further 15 min in 0.1× SSC-0.1% SDS at 55°C.

Single-Strand Conformation Polymorphism and Sequence Analyses

Amplification of exons 13 to 18 of the RB gene and single-strand conformation polymorphism (SSCP) analysis were carried out as previously described (41).

For the study of exons 20 to 24, 100 ng of genomic DNA was mixed with dATP, dCTP, dGTP, and dTTP (200 µM each), 0.4 pmol of each primer, 2 units of Taq polymerase (Promega, Madison, WI), 1 µCi of [-³³P]dATP (3,000 Ci/mmol; Amersham), 10 mM Tris-HCl (pH 8.3), 50 mM KCl, 1.5 mM MgCl₂, and 0.001% gelatin (wt/vol). Amplification of the genomic DNA consisted of denaturation at 94°C for 5 min followed by 30 cycles of 94°C for 30 s, annealing at a specific temperature for 30 s (see the following) and extension at 72°C for 90 s each, and was ended by a 5-min extension at 72°C.

Primers used for DNA amplification of exons 20 to 24 of the RB gene were as follows: exon 20 (sense 5' AAAATGACTAATTTTTCTTATTCCC 3', antisense 3' CGTGAAGTGGAAGAGAGGAGGGATG 5', annealing at 45°C), exon 21 (sense 5' ATAAAATTCTGACTACTTTTACATC 3', antisense 3' CTATTTAGGTATAGGTATTGTATTG 5', annealing at 38°C), exon 22 (sense 5' AATTCATTTAACAAGTAAATTTTAC 3', antisense 3' CCC-AGGTGGTTTTGTAATTTATTTA 5', annealing at 47°C), exon 23 (sense 5' GATTGGAAAAATCTAATGTAATGGG 3', antisense 3' CACACAAAAGAGAAATCCCTTCATC 5', annealing at 50°C), and exon 24 (sense 5' ATGATGTATTTATGCTCATCTCCTGC 3', antisense 3' GAAAGATACTTTATATTATCATACG 5', annealing at 44°C).

For SSCP analysis of exons 20-24, 1 µl (10%) of the polymerase chain reaction (PCR) product was diluted with 9 µl of a loading buffer containing 95% formamide, 1 N NaOH, 0.05% bromophenol blue, and 0.05% xylene cyanol. Samples were denatured at 95°C for 5 min and placed immediately on ice before loading on a 6 or 8% polyacrylamide nondenaturing gel containing 10% glycerol. The DNAs were electrophoresed in TBE (0.045 M Tris base, 0.045 M boric acid, and 1.5 mM EDTA [ethylenediaminetetraacetic acid]) running buffer for 5 h at 30 W at 4°C. Gels were dried and exposed to Xar-5 film for 12-72 h at 80°C.

The cases with abnormal profiles as determined by SSCP were cloned in a pTAg vector (pUC backbone derived) using a LigATor kit (R&D Systems, Minneapolis, MN) and sequenced using a T7 sequencing kit (Pharmacia, Uppsala, Sweden).

Reverse Transcription PCR

Total RNA was isolated as described by Chomczynski and Sacchi (42). One microgram of RNA was reverse transcribed with avian myeloblastosis virus (AMV) reverse transcriptase (Promega) at 37°C for 1 h using oligo(dT) in a total volume of 50 µl. Two microliters of the cDNA was amplified with a 10 µM concentration each of primers RB22ase and RB19se previously described (40), 0.2 mM dNTPs, and 2.5 units of Taq DNA polymerase in the buffer recommended by the supplier (Promega). Amplification consisted of denaturation at 94°C for 1 min followed by 30, 35, 40, and 45 cycles at 94°C for 30 s, at 52°C for 1 min, and at 72°C for 2 min each and was ended by a 2-min extension at 72°C. The PCR reaction consisted of the amplification of an RB cDNA fragment of 517 bp corresponding to exons 19 to 22 of the RB gene. Amplification of a fragment of the cDNA of -actin (200 bp) was performed in the same PCR reaction as internal control, with the following primers: sense (F, 5' CCTTCCTGGGATGGAGTCCTG 3') and antisense (R, 5' GGAGCAATGATCTTGATCTTG 3'). Half of each reaction mixture was analyzed by agarose gel electrophoresis. Each PCR was repeated at least once.

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

Immunohistochemical Analysis

Nuclear staining only was interpreted. Tumors were considered as negative for Rb protein expression when staining of normal cells (especially endothelial cells) was retained, whereas all or large areas of tumors were negative with all antibodies. Altered Rb protein expression was rare and attributed only to cases in which some antibodies gave focal positive reaction with tumor cells, whereas others gave negative reaction, suggesting an abnormality of some protein epitopes. In this respect, it is of note that the normal cells of the stroma and normal structures (type II pneumocytes, bronchial epithelial cells) were always strongly reactive with all antibodies and were exploited as internal positive controls.

Loss or altered expression of Rb protein was detected in 1 of 8 typical carcinoids (12.5%), 6 of 7 LCNECs (86%), and 17 of 21 SCLCs (81%) (Table 2) (Figure 1). Abnormal Rb expression was more frequently observed in high-grade NE lung carcinoma (23 of 28, 82%) than in typical and atypical carcinoids (1 of 9, 11%; P < 0.001). DNA and RNA alterations were studied in the 24 tumors with abnormal RB protein expression.

View larger version (114K): [in this window] [in a new window]   View larger version (118K): [in this window] [in a new window]   Figure 1.   Immunohistochemical staining of Rb protein using Rb1F8 antibody in NE lung carcinomas. (A) Typical carcinoid: Tumor cells and stromal cells show positive nuclear staining. (B) Small-cell lung carcinoma: Absence of nuclear staining in tumor cells as compared to endothelial cells. (Immunoperoxidase; original magnification, ×400; HE counterstained.)

DNA Structural Rearrangements

Southern blot analysis of the 24 NE tumors with loss or altered Rb protein expression was performed with the RB 3.8 kb and RB 0.9 kb probes. Structural alteration was observed in only one SCLC (tumor 17), which showed a deletion of the 10-, 6.2-, 4.5-, and 2.1-kb restriction fragments when compared to the 10-, 7.5-, 6.2-, 4.5- and 2.1-kb HindIII restriction fragments usually detected with the RB 3.8 kb probe (Figure 2). This deletion comprised approximately exons 24 to 27 of the RB gene. Structural abnormalities were never detected with the RB 0.9 kb probe.

View larger version (43K): [in this window] [in a new window]   Figure 2.   Southern blot analysis of the RB gene in NE lung carcinomas. DNA (20 µg) was digested with HindIII restriction endonuclease and hybridized with RB 3.8 kb probe. This probe detected the 10.0-, 7.5-, 6.2-, 4.5-, and 2.1-kb HindIII fragments in normal lung (N) and in tumors 1 and 5 (T1 and T5). Tumor 17 (T17) presented a deletion of the 10.0-, 6.2-, 4.5-, and 2.1-kb HindIII fragments.

SSCP Analysis

Minor DNA alterations were analyzed using SSCP in the 24 Rb-negative NE tumors. Exons 13 to 18 and 20 to 24 of the RB gene were studied.

Mobility shifts were observed in one case (tumor 12) for exons 15 and 16, one case for exon 20 (tumor 11), two cases for exon 21 (tumor 15 and tumor 10), one case for exon 23 (tumor 8), and one for exon 24 (tumor 5).

In total, 6 of 24 (25%) tumors exhibited aberrant profiles for exons 13-18 or 20-24 of the RB gene, using SSCP. As expected on the basis of Southern blot analysis, no PCR product could be observed for exon 24 in tumor 17.

Sequence analysis of these abnormal cases revealed a single-base insertion or deletions leading to frameshifts or a premature stop codon (Table 3) (Figure 3).

View larger version (34K): [in this window] [in a new window]   Figure 3.   Examples of PCR-SSCP and sequencing analyses in exons 16 and 21 of the RB gene. Exon 16: Tumor 12 exhibited a mobility shift using SSCP in exons 15 and 16 of the RB gene when compared to normal lung (N) and tumor 9 (T9). Sequencing analysis showed a 2-bp deletion (*) in codon 485 in tumor 12 (T12) when compared with the normal lung. Exon 21: Tumor 10 exhibited a mobility shift using SSCP in exon 21 of the RB gene when compared to normal lung (N) and tumor 2 (T2). Sequencing analysis showed a single base-pair insertion (*) in codon 721 in tumor 10 (T10) when compared with the normal lung.

Expression of RB mRNA

The study of the transcriptional activity of the RB gene was carried out using RT-PCR in the 17 NE tumors without detected genetic alterations as determined by Southern blotting and SSCP analyses. The level of RB mRNA was studied after 30, 35, 40, and 45 cycles of amplification and analyzed as follows: the mRNA expression was considered to be normal when a PCR product was detected after 30 cycles of amplification as was the case in normal lung tissue. A low level of transcription was defined when the PCR product appeared only after 40 cycles, and the expression was considered to be absent when any PCR product could be detected even after 45 cycles of amplification.

Among the 17 tumors analyzed, 3 showed a normal level of expression of RB mRNA, 4 exhibited a low level of expression, and 10 had an absence of RB mRNA expression (Figure 4). Overall, abnormalities in RB mRNA expression were observed using RT-PCR in 58% (14 of 24) of tumors with abnormal Rb protein expression.

View larger version (41K): [in this window] [in a new window]   Figure 4.   RT-PCR analysis of RB mRNA expression in NE lung tumors. Expression of RB mRNA was analyzed after 30, 35, 40, and 45 cycles of PCR amplification of exons 19 to 22 of the RB cDNA, corresponding to a fragment of 517 bp. For each reaction a fragment of 200 bp of the actin fragment was simultaneously amplified as internal control of PCR amplification. Tumor 2 presented an apparently normal expression of RB mRNA as compared with the normal tissue. A low level of RB mRNA expression was detected in tumor 18 and absence of RB mRNA was observed in tumor 14.

Methylation Status of the RB Gene

In a search for a mechanism to explain RB silencing, the methylation status of the CpG-rich island at the 5' end of the RB gene was analyzed using the methylation-sensitive restriction enzyme SacII in the 14 NE lung carcinomas with negative or aberrant expression of RB mRNA as determined by RT-PCR. In normal lung, the 6.1-kb SacI fragment from the 5' end of the RB gene was cleaved into 4.1-, 1.7-, and three 0.2-kb fragments. These last three small fragments were not detectable by Southern blot hybridization (40). Hypermethylation was never observed in the tumors analyzed. However, a rearrangement of the methylation-sensitive SacII restriction site, which was not shown by Southern analysis, was detected in tumor 16 (Figure 5).

View larger version (60K): [in this window] [in a new window]   Figure 5.   Study of methylation of the 5' CpG island of RB in NE lung tumors. DNA was digested with SacI (S) and SacII (II) enzymes and hybridized with p123M1.8 probe (40). Tumor 3 showed a normal pattern after SacI and SacII digestion. A rearrangement was observed in tumor 16 after SacI restriction (fragment at 5.2 kb) and SacI and SacII restriction (fragments at 5.2 and 4.2 kb).

Correlation Between Negative Immunohistochemistry and Molecular Analysis

Table 4 summarizes the results of the molecular analyses performed in an attempt to explain the abnormal expression of Rb protein detected in 24 NE lung tumors using immunohistochemistry (IHC). Of the 24 NE tumors with abnormal expression of Rb protein, 14 showed a low level or absence of RB mRNA expression, 6 exhibited mutations in exons 13-18 or 20-24 of the RB gene, and 1 had a rearrangement. One tumor exhibited both a rearrangement and a loss of RB RNA expression. Either of these mechanisms was detected in 100% of the SCLCs with abnormal Rb immunostaining whereas only 50% of the Rb-negative LCNECs analyzed carried one of these genetic alterations.

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

We have previously shown a loss of or aberrant Rb protein expression using IHC in 87% of high-grade NE lung carcinomas (SCLCs and LCNECs) (36). Moreover, loss of Rb protein was highly associated with loss of heterozygosity at the RB locus in these NE carcinomas, suggesting that the RB gene was specifically targeted as a tumor suppressor gene in these tumors. In the present study, Rb protein expression was analyzed using IHC with five Rb antibodies in a larger series of human NE lung carcinomas including low- and high-grade tumors and the genetic alterations leading to an abnormal Rb protein expression were investigated.

Of the 37 NE lung tumors analyzed, 11% of typical and atypical carcinoids and 82% of highly malignant LCNECs and SCLCs displayed loss or altered Rb protein expression. Our results on SCLCs are in agreement with previous reports on SCLC cell lines, wherein loss of Rb protein expression was detected in 35 to 90% of cases (5, 30, 32, 43). In accord with one report (31), loss of Rb protein expression was more frequent in high-grade SCLCs and LCNECs than in low-grade carcinoids. However, altered expression was never described, probably owing to the use of only one Rb antibody in most reports, and because it is far less frequent in SCLCs than in LCNECs, which were not previously studied. This, of course, underestimates the incidence of altered expression, which could rely on abnormal conformation of the Rb protein and differences in the accessibility of some epitopes, as compared with the normal Rb protein. Accordingly, the large variation in the frequency of Rb loss between previous series in SCLCs (35% to 90%) could be related to the different affinities of the Rb antibodies used (5, 43). In the present series, only 18% of the high-grade NE lung carcinomas retained Rb protein expression. Although mutations in exons 13 to 18 and 20 to 24 of the RB gene leading to the expression of an Rb protein with altered function have been described, such mutations are rare (44, 45), and we found that p16^INK4 was inactivated in the majority of these cases (data not shown) suggesting a normal RB function. We thus focused our attention on the cases with abnormal Rb staining and sought the genetic mechanisms for RB gene inactivation.

Structural abnormalities, detectable by Southern blot analysis, were rare in the present study, in agreement with previous studies on a small number of freshly excised SCLC tumors (4, 30).

The mutations were analyzed in exons 13-18 and 20-24 of the RB gene because these specific regions had been described as key regions (RB pocket) for the binding of Rb protein with viral (13) and cellular proteins (16), and were previously reported as the targets for mutations in several types of tumors and in some SCLC cell lines (19, 24, 46). We demonstrated the existence of frameshifts or deletions leading to the premature introduction of a stop codon, which explained the defect in Rb protein expression. Such genetic alterations have been previously reported in the small number of SCLC cell lines analyzed (24, 46). However, at this time, our study is the first one analyzing a large series of NE lung tumors for mutations and showing the implication of such mutations (25% of the cases with abnormal Rb staining) in the loss of Rb protein expression.

Abnormality of RB transcription was detected using RT-PCR with a high frequency of low level mRNA or an absence of mRNA similar to that observed in SCLC cell lines by Northern blot analysis (4, 30). Because methylation was described as a means of transcriptional silencing of other tumor suppressor genes such as p16^INK4 (49) and had been previously described in some sporadic retinoblastomas (tumors and cell lines) (40), we looked for methylation at the 5' end of the RB gene in the tumors with absent or low expression of RB mRNA. However, methylation of the RB gene was never found. This result suggests the existence of other inactivating genetic mechanisms such as mutations outside the RB pocket leading to abnormal splicing, mutations in the promotor region, or deregulation of transcription factor activities.

Overall, we could detect rearrangement, mutation in the RB pocket, or transcriptional deficiency in 100% of SCLCs and in only 50% of LCNECs with loss of or altered Rb protein expression. These results suggest that different mechanisms of RB gene inactivation are involved in both tumor types or that yet undescribed mutations are present in other exons of the RB gene in LCNECs. Such mutations outside the RB pocket have been previously described in some retinoblastomas and non-small-cell lung carcinomas (48, 50, 51) but have never been studied in NE lung tumors. Furthermore, data have shown the importance of the amino terminus of Rb (52, 53), and somatic frameshift mutations, leading to stop codons and truncated proteins, have been reported in the N terminal-encoding region of the RB gene in retinoblastomas (48). Because the antibodies used in our IHC analysis detect epitopes in mid- or carboxy-terminal regions, we investigated the possibility that truncated Rb protein, restricted to that portion of the gene encoding the NH₂ terminal part of the protein, exists. However, any Rb protein could be detected by immunoprecipitation with an NH₂-terminal antibody, indicating that if such mutations occur in this region, they probably lead to unstable or rapidly degraded proteins (data not shown).

In summary, this study indicates that Rb protein expression is lost or altered at a high frequency in high-grade NE lung tumors, including SCLCs and LCNECs, and that inactivation of the RB gene is required for SCLC growth and phenotype. The inactivation of the RB gene is more frequently related to altered expression of RB mRNA than to mutations in the RB pocket in SCLCs. The LCNEC phenotype could rely in half of the cases on other mechanisms of genetic inactivation of the RB gene, which deserves further study. The loss of Rb protein in one carcinoid indicates that RB inactivation, quite exceptional (31), can be observed in well-differentiated low-grade NE proliferation, but increases considerably in frequency with grade. Interestingly, the only case of Rb-negative carcinoid included in this study had a higher proliferation rate as reflected by five mitoses per high power field (1) and 11% of labeled cells with Ki 67, a cell cycle proliferation antigen (53). With regard to the threshold value of the mitotic index given by Arrigoni and coworkers (1) for discrimination between typical and atypical carcinoids, this case could be considered as marginally situated between typical and atypical carcinoid.

Thus, restoration of G1 arrest by RB transfer under strong promotion could have therapeutic effects in most high-grade NE lung tumors in which the RB gene is inactivated at the genetic level, as already shown in SCLC cell lines (54, 55).