Altered expression of G1 regulatory proteins in human soft tissue sarcomas. (Original Articles).

Soft tissue tumors constitute a highly heterogeneous group of tumors arising from the nonepithelial extraskeletal tissue of the body, exclusive of the reticuloendothelial system and glia. (1) Malignant soft tissue tumors (sarcomas) account for less than 1% of human malignancies and less than 2% of all cancer deaths. Although various physical and chemical factors, exposure to ionizing radiation, and inherited or acquired immunological defects are implicated as possible causes of sarcoma, the pathogenesis of most soft tissue tumors is still unknown. The findings of modern cellular and molecular biology have had a profound impact on the understanding of tumorigenesis. Many of the tumor suppressor genes and oncogenes directly participate in or regulate signal transduction pathways linking extracellular stimuli to cell-cycle progression and/or cell death, which are 2 critical areas where abnormalities occur in cancer cells. The genes (and their proteins) acting at the G1-S checkpoint represent one of the more frequent targets in molecular tumorigenesis. (2,3) The major protein interactions that occur during the G1 phase are summarized schematically in Figure 1.

[FIGURE 1 OMITTED]

p53 is a tumor suppressor gene located on chromosome 17. p53 mutations represent a frequent genetic alteration in human malignancies and produce proteins, sequence-specific transcription factors that act to induce or repress specific genes involved in multiple cellular functions, including progression through the cell cycle, DNA repair after damage, genomic instability, monitoring of gene amplification events, and commitment of cells to apoptosis. (4-6) This p53 pathway may also be disrupted by alteration of the cellular gene mdm2, which contains a p53 DNA binding site. (7,8) It has been demonstrated that an autoregulatory feedback loop regulates both the activity of p53 and the expression of the mdm2 gene. (9,10)

The Rb-cyclin D pathway is emerging as an important target for molecular 5 aberration at the G1-S checkpoint. (3) When activated by cyclin D1, cdk4 is able to phosphorylate the pRb with subsequent release of associated proteins that have the capability to activate genes necessary for the progression of cells through the G1 phase, resulting in a cell growth advantage and eventual tumor development. (11) The p16 protein can inhibit cell proliferation, possibly as a result of the inhibition of cdk4 and hence the regulation of pRb phosphorylation. (12)

The literature contains relatively limited data regarding the altered cell-cycle regulation that may underlie the development and/or progression of human sarcomas. In this article, we analyze a cohort of soft tissue tumors by investigating both the p53 and Rb-cyclin D pathways to detect the abnormal features of cell-cycle regulatory proteins in these tumors and to determine their potential role in clinical behavior.

MATERIALS AND METHODS

Sixty-seven samples were obtained from patients with soft tissue sarcoma. The tumors included 11 liposarcomas, 9 leiomyosarcomas, 14 malignant fibrous histiocytomas, 14 malignant peripheral nerve sheath tumors, and 19 rhabdomyosarcomas. All patients were surgically treated at the Catholic University Medical Center (Suwon, South Korea) between 1992 and 2000. None of the patients had received prior therapy. The tumor sizes ranged from 1.9 to 32 cm. Tissue samples were fixed in 10% buffered formalin. After routine embedding, light microscopy and immunohistochemistry and/or electron microscopy led to the final diagnosis. One of the authors (J.Y.) reviewed the histopathologic diagnosis or tumor typing according to the relevant World Health Organization classifications, tumor grade, and the quality of the tissue sections. Histologic evaluation of the tumor tissue assured that the specimens consisted of at least 75% tumor cells. Using the French Federation of Cancer Centers Sarcoma Group grading system, (13) 4 of the 67 sarcomas were identified as low grade, 39 as intermediate grade, and 24 as high grade.

Clinical information was obtained through a computerized prospective database. Follow-up data were available from 61 patients for a period ranging from 11 to 86 months, with a mean of 57 months. Eighteen patients developed either metastasis or local recurrence (relapsed group), whereas 43 had no evident signs of disease (disease-free group) during a minimum follow-up of 14 months.

Representative sections of archival, formalin-fixed, paraffin-embedded material were reacted in immunohistochemical procedures for the detection of the following antigens: p53 (Zymed Laboratories Inc, San Francisco, Calif), mdm2 (Zymed), pRb (PMG3-245; PharMingen, San Diego, Calif), p16 (PharMingen), cyclin D1 (NeoMarkers, Fremont, Calif), and cdk4 (Santa Cruz Biotechnology Inc, Santa Cruz, Calif). Each block was cut into 5-[micro]m-thick sections, which were deparaffinized in xylene and rehydrated in graded alcohols and water. The endogenous peroxidase activity was blocked by treating the sections with 0.3% hydrogen peroxide for 5 minutes at room temperature. The slides were then placed in a Coplin jar containing citrate buffer (2.1 g/ L, pH 6.0) and heated to 121 [degrees] C in an autoclave for 15 minutes for antigen retrieval. Sections were treated with protein-blocking reagent before incubation at 50 [degrees] C for 30 minutes with primary antibodies at the following conditions: 1:50 dilution for p53, mdm2, and cdk4; 2.0 [micro]g/mL concentration for cyclin D1; and 1.0 [micro]g/mL for pRb and p16. After extensive washing, slides were incubated at 50 [degrees] C for 30 minutes with biotinylated anti-mouse immunoglobulin antibodies (Dako Co, Ltd, Kyoto, Japan) at 1:20 dilution and for another 30 minutes with streptavidin-biotin peroxidase complexes at 1:25 dilution. The staining was developed in 3-amino-9-ethylcarbazole. The nuclei were counterstained with Mayer hematoxylin.

All series included positive and negative controls. The positive controls were tissue from colon tumor for p53 and pRb, tissue from breast tumor for mdm2 and cyclin D1, tonsil for p16, and normal skin for cdk4. As negative controls of the immunohistochemistry technique, the primary antibody was omitted. To perform a semiquantitative evaluation of immunostaining, all sections were screened by 3 independent observers who were not aware of the clinicopathologic data. Tumor cells showing a nuclear staining pattern were interpreted as positive. For the expression of p53, mdm2, cyclin D1, and cdk4 proteins, cell positivity was scored as follows: minus sign, less than 10% of tumor cells were stained; plus sign, 10% or more of the cells were reactive. (14,15) For pRb and p16, the immunoreactivities were classified into 3 categories, defined as follows: minus sign, less than 20% of cells were stained; heterogeneous, 20% to 80% were stained; plus sign, greater than 80% were reactive. (16,17) Samples showing either negative or heterogeneous staining were considered abnormal, whereas those with positive staining were interpreted as normal. Three observers initially reached agreement in 89% of the cases. For the discordant samples, slides were reviewed jointly and a consensus was reached. Statistical significance was evaluated by the [chi square] test and Fisher exact test with the criterion of P < .05.

RESULTS

Sixty-seven soft tissue sarcomas were classified, graded, and analyzed for the expression of p53, mdm2, pRb, p16, cyclin D1, and cdk4. Negative and positive controls showed the expected results. The results of the analysis are summarized in Tables 1 and 2.

Positive immunostaining for p53 was detected in 25 tumors (37%) (Figure 2). The distribution of p53-positive cells was random within a section. In relation to the histologic grade of sarcomas, p53 positivity was absent in all grade 1 tumors, but was present in 10 (26%) grade 2 and 15 (63%) grade 3 lesions. The frequency of p53 immunoreactivity significantly increased as the tumor grade advanced (P = .001).

[FIGURE 2 OMITTED]

Expression of mdm2 was observed in 16 (24%) of the 67 cases (Figure 3). The immunoreactivity was heterogeneous within the tumor area. A statistically significant correlation was found between mdm2 expression and tumor grade (P = .035). About half of the tumors (49%) exhibited a positive reaction for p53 and/or mdm2.

[FIGURE 3 OMITTED]

pRb was variably expressed in the nuclei of tumor cells with minimally detectable reactivities in their cytoplasm. Positive nuclear staining in more than 80% of tumor cells was considered normal, and was evident in 19 samples (28%), while altered patterns of pRb immunostaining were observed in 48 (72%). Altered staining patterns included heterogeneous expression with some cells (20%-80%) positive and others negative in 35 samples and a predominately unreactive nuclear pattern (<20%) in 13 (Figure 4). Abnormal pRb expression was detected in all (100%) of the grade 1 lesions, in 27 (69%) of the 39 grade 2 lesions, and in 17 (71%) of the 24 grade 3 lesions (P > .05).

[FIGURE 4 OMITTED]

For immunohistochemical evaluation of the p16 protein, the same criteria as for pRb were applied. Of the 67 sarcomas analyzed, 63 (94%) expressed p16 at abnormal levels, either heterogeneously (n = 34, 51%) or negatively (n = 29, 43%) (Figure 5). Only 4 samples (6%), 3 (21%) of 14 malignant fibrous histiocytomas and 1 (5%) of 19 rhabdomyosarcoma, displayed normal p16 expression. Altered patterns of p16 expression were detected in all grade 1 (100%), in 38 (97%) of 39 grade 2, and in 21 (88%) of 24 grade 3 tumors (P > .05).

[FIGURE 5 OMITTED]

The cyclin D1 protein was expressed in 14 samples (21%) (Figure 6). None of the grade 1 tumors were positive for cyclin D1, whereas 23% (9/39) of grade 2 tumors and 21% (5/24) of grade 3 tumors were positive for cyclin D1 (P > .05).

[FIGURE 6 OMITTED]

A strong and uniform nuclear immunostaining for the cdk4 protein was found in 64 (96%) of the samples (Figure 7). The cdk4 protein was demonstrated in 4 (100%) of 4 grade 1 tumors, 37 (95%) of 39 grade 2 tumors, and 23 (96%) of 24 grade 3 tumors (P > .05).

[FIGURE 7 OMITTED]

Table 3 presents the relationships among the expressions of different proteins. A statistically significant association was observed between cyclin D1 and pRb (P = .044). Eight of 14 cyclin D1-positive samples had detectable p53 protein compared with 17 of 53 cyclin D1-negative specimens. This difference approached statistical significance with a P value of .087 by the [chi square] method. All of the analyzed tumors showed altered expression of either pRb or p16, and 44 (66%) sarcomas were abnormal for both. However, no statistically significant correlation was found between these 2 proteins. Each of p53, mdm2, pRb, p16, cyclin D1, and cdk4 established no association with the patients' prognoses (Table 4) or other clinicopathologic parameters, including age, sex, tumor site, and tumor size (data not shown).

COMMENT

Our previous investigation of p53 gene alteration in soft tissue tumors suggested its role in the pathogenesis of these tumors, (18) which prompted us to further investigate other proteins regulating the cell-cycle progression for the evaluation of their functional status and possible correlations.

In the present study, p53 overexpression was observed in 37% of the tumors analyzed, whereas mdm2 immunoreactivity was identified in 24%. Alterations in the p53 pathway, evidenced by p53 and/or mdm2 overexpression, occurred in 49% of the cases, supporting their role in the development and/or progression of sarcomas. Because mdm2 protein inactivates p53 transcription activity, one might not expect mdm2 overexpression in a cell that overexpressed p53 protein. Leach et al (19) found no tumors containing an alteration of both p53 and mdm2 genes, suggesting that p53 and mdm2 genetic alterations are alternative mechanisms for inactivating the same regulatory pathway for suppressing cell growth. We, however, detected 8 (12%) cases exhibiting immunoreactivity for both proteins. Furthermore, all of these tumors were high-grade lesions (7 grade 3 tumors and 1 grade 2 tumor). These findings are in good agreement with those of Cordon-Cardo et al, (20) who reported that approximately 10% of sarcomas expressed both proteins and had poor prognoses. Our data also showed that p53 positivity was significantly correlated with the histologic grade of malignancy of sarcomas. The same tendency was observed in mdm2 immunoreactivity. These findings strongly indicate that the p53 pathway may be highly associated with the transition from low- to high-grade tumors and can be considered as a marker for tumor progression in sarcomas.

Reactivity for the p53 protein varied according to the types of tumors and may be of dissimilar significance in different sarcoma subtypes. The positive immunoreactivities were different even in the same type of sarcomas. Kawai et al (21) reported p53 expression in 100% of malignant peripheral nerve sheath tumors, whereas we found it in 36% of our samples with this tumor type. Differences in tumor differentiation, histologic features, clinical stages, and analytical procedures may have contributed to this difference. Additional study, including a large number of histologically homogeneous series of tumors, will be necessary to verify a certain degree of tumor specificity, p53 alterations can be detected by immunohistochemistry of p53 protein accumulation and p53 gene mutation. Several previous studies have described a discordance between these 2 methods. In our earlier investigation of soft tissue sarcomas, (18) the identification of p53 nuclear overexpression by immunohistochemistry did not always reflect the detection of p53 mutation, suggesting that mechanisms other than mutations may also play a role in p53 overexpression. The status of the p53 gene, however, was not evaluated in this series. Mousses et al (22) found that in primary human sarcomas, missense p53 mutations were associated with lower levels of p21 expression, whereas p53 truncation mutations were associated with higher expression, suggesting different effects of various types of p53 mutations on the transcriptional activation of the p21 promoter and on the expression of other downstream genes. In another recent investigation intending to elucidate the exact mechanisms by which p53 may regulate the cell-cycle progression,23 MAP kinase-dependent pathways were shown to regulate p53 levels by regulating the expression of p53 messenger RNA.

The retinoblastoma (Rb) gene, located on human chromosome 13q14, encodes a 110-kd nuclear phosphoprotein and is thought to function as a tumor suppressor gene by regulating cell cycle progression. (11,24,25) Its deletion in tumors of patients with heritable retinoblastoma and the high incidence of sarcomas in the survivors of this disease suggested that the deletion of the Rb gene might be associated with the development of sarcomas. (26,27) To date, altered Rb gene or pRb has been sporadically reported in sarcoma cell lines and some sarcomas. (16,17,28-32) Homozygous deletion was detected in 4.8% (3/63) of sarcomas, and loss of heterozygosity was observed in 22.7% (5/22) of tumors in which their matching controls were available. (28) In another study, the deletion of the Rb gene was seen in 3 (33.3%) of 9 osteosarcomas and in 4 (13.8%) of 29 soft tissue sarcomas. (29) However, the methods used in these studies were restriction fragment length polymorphism and Southern blot analysis, which are useful in detecting structural changes in the Rb gene itself, but inappropriate for examining such detailed changes as expression of the Rb gene product, thus the incidences were thought to be relatively low.

Using the immunohistochemical method, the present study demonstrated the expression of pRb to be normal in 19 (28%) of the total 67 soft tissue sarcomas and to be abnormal (either negative or heterogeneous) in 48 cases (72%). Cance et al (16) observed similar results (70.5% abnormal). However, all of the tumors analyzed in their study were high-grade lesions, including osteosarcomas and sarcomas that developed during childhood, which are better separated from adult soft tissue sarcomas. In a study by Karpeh et al, (17) only soft tissue sarcomas that developed in adults were included, and 75% of the samples showed abnormal expression. Other investigators described the incidence to be relatively low, from 53% to 23.7%. (30,31) These various results may have been due to the difference in tumor types and the nature of tumor samples analyzed. Other technical problems should also be considered, for example, whether paraffin-embedded or fresh tissue was used, whether heat treatment was applied, or what type of reagents were used. Four different antibodies to pRb used for the pRb immunostaining were reported to yield similar results, leading us to believe that the type of antibodies would not make a difference. (30) On the other hand, the criteria for interpreting the results as positive or negative may have attributed to this difference. Immunohistochemical staining of the pRb could show a heterogeneous pattern, which is believed to be due to the 2 clones, that is, one having the wild-type protein and the other having mutant protein. This heterogeneous pattern was considered either positive or negative according to the investigators using different percentage criteria (Table 5). (16,17,30,31) While only samples with more than 75% positive nuclei were regarded as having normal expression, immunoreaction in 25% to 75% of the cells was also considered as normal by others, thus the rate of abnormal expression was probably low. Karpeh et al (17) regarded more than 80% of the nuclei showing positive reaction as a normal expression and the rest as abnormal expression (<20% as negative, 20%-80% heterogeneous pattern), which has previously been defined by image analysis and confirmed by Western blotting. With the use of these criteria in the current study, pRb loss occurred in 72% of the cases, indicating that abnormal expression of the pRb in soft tissue sarcomas was more frequent than expected. It has previously been documented that aberrant pRb expression is different in the various types of sarcomas. Dei Tos et al (33,34) reported pRb abnormalities in 66% of liposarcomas and 90% of leiomyosarcomas without providing the information on their scoring scheme. Karpeh et al (17) observed high rates of altered pRb in liposarcomas, leiomyosarcomas, malignant fibrous histiocytomas, and rhabdomyosarcomas, whereas Cohen and Geradts (31) demonstrated infrequent pRb alterations in liposarcomas, fibresarcomas, and malignant peripheral nerve sheath tumors, although some of their normal samples would be considered abnormal in our scoring system. In this study, the lack of altered pRb in liposarcoma contrasts with the high rate of pRb loss detected in rhabdomyosarcoma, suggesting the possibility of tumor-specific patterns of molecular alterations. Since sarcomas are a group of heterogeneous tumors, the signal transduction pathway probably differs from one type to the other; thus, a histologically homogeneous series of sarcomas should be investigated using the identical scoring criteria in future studies.

Recent investigations demonstrated that wild-type p16 arrested normal diploid cells in late G1, whereas a tumor-associated mutant of p16 did not, suggesting p16 as a tumor suppressor. (12,35,36) Since the p16 protein was altered in most (94%) of our samples, its inactivation most likely plays a critical role in the formation of sarcomas. Interestingly, altered expression of either pRb or p16 was present in all sarcomas analyzed, which points to the significance of this pathway in the maintenance of the G1-S checkpoint.

Although cyclin D1 overexpression seemed to be an uncommon event, it showed a strong correlation with pRb and a possible correlation with p53, suggesting a key role of cyclin D1 in linking both pathways. In addition, our findings of low expression of cyclin D1 in the absence of pRb expression are in accordance with results of other investigators, (33) confirming the essential requirement of functional Rb gene in cyclin D1 expression. (37) The ability of p16 to induce cell-cycle arrest is also lost in cells lacking functional pRb. Thus, loss of pRb, loss of p16, and overexpression of cyclin D1 have similar effects on G1 progression and may represent a common pathway to tumorigenesis.

Overexpression of cdk4 in a significant number of tumors in the present study suggests a role of cdk4 upregulation in the process of tumor development. The cdk4 gene is localized to human chromosome segment 12q13-14, which also contains mdm2. (38,39) Overexpression of the products of both genes was observed in 22% of our cases, suggesting an existence of architectural factor in the cell-cycle regulation system.

In summary, our data emphasize the roles of the p53 and Rb-cyclin D pathways in soft tissue sarcoma pathogenesis. In particular, overexpression of p53 and mdm2 correlated with tumor grade, indicating their relationship to tumor progression. The presence of abnormalities in the Rb-cyclin D pathway in all of the tumors analyzed in this series raises the possibility of an early tumorigenic event. On the basis of the results of the correlation between overexpression of cyclin D1 and altered pRb or p53, we speculate that cyclin D1 is most likely responsible for linking both pathways in these tumors.

Table 1. Expression Status of p53, mdm2, pRb, p16, Cyclin D1, and
cdk4 in Different Types of Sarcomas *

                 Histologic Type, No. Positive (%)

               LS        LMS       MFH       MPNST
            (n = 11)   (n = 9)   (n = 14)   (n = 14)

p53          1 (9)     5 (56)     9 (64)     5 (36)
mdm2         3 (27)    4 (44)     3 (21)     4 (29)
pRb          1 (9)     2 (22)     2 (14)     3 (21)
p16          0 (0)     0 (0)      3 (21)     0 (0)
Cyclin D1    1 (9)     5 (56)     2 (14)     5 (36)
cdk4        11 (100)   8 (89)    14 (100)   13 (93)

             Histologic
             Type, No.
            Positive (%)

                RMS         Total
              (n = 19)     (n = 67)

p53            5 (26)      25 (37)
mdm2           2 (11)      16 (24)
pRb           11 (58)      19 (28)
p16            1 (5)        4 (6)
Cyclin D1      1 (5)       14 (21)
cdk4          18 (95)      64 (96)

* LS indicates liposarcoma; LMS, leiomyosarcoma, MFH, malignant
fibrous histiocytoma; MPNST, malignant peripheral nerve sheath
tumor; and RMS, rhabdomyosarcoma.
Table 2. Correlation of Expression of p53, mdm2,
pRb, p16, Cyclin D1, and cdk4 With Tumor Grade

                  Histologic Grade, No.
                      Positive (%)

               1         2          3
            (n = 4)   (n = 39)   (n = 24)   P Value *

p53         0 (0)     10 (26)    15 (63)      0.001
mdm2        1 (25)     5 (13)    10 (42)      0.035
pRb         0 (0)     12 (31)     7 (29)       NS
p16         0 (0)      1 (3)      3 (13)       NS
Cyclin D1   0 (0)      9 (23)     5 (21)       NS
cdk4        4 (100)   37 (95)    23 (96)       NS

* NS indicates not significant.
Table 3. Correlation Among the p53, mdm2, pRb, p16, Cyclin D1, and
cdk4 Proteins in Sarcomas *

                 mdm2           pRb           p16

            -         +    -         +    -         +

p53
  -         34         8   31        11   41        1
  +         17         8   17         8   22        3
  P value        NS             NS             NS
mdm2
  -                        38        13   49        2
  +                        10         6   14        2
  P value                       NS             NS
pRb
  -                                       44        4
  +                                       19        0
  P value                                      NS
p16
  -
  +
  P value
Cyclin D1
  -
  +
  P value

               Cyclin D1         cdk4

            -            +    -        +

p53
  -         36            6   3        39
  +         17            8   0        15
  P value        0.087            NS
mdm2
  -         41           10   2        49
  +         12            4   1        15
  P value         NS              NS
pRb
  -         41            7   2        46
  +         12            7   1        18
  P value        0.044            NS
p16
  -         49           14   3        60
  +          4            0   0         4
  P value         NS              NS
Cyclin D1
  -                           3        50
  +                           0        14
  P value                         NS

* NS indicates not significant.
Table 4. Correlation Between Grade, p53, mdm2, pRb, p16, Cyclin D1,
or cdk4 and Tumor Relapse *

                   Grade           p53            mdm2

               1     2     3   -         +    -         +

Relapsed       0     7    11   11         7   12         6
  P value          .012             NS             NS
Disease free   3    28    12   27        16   33        10

                   pRb            p16        Cyclin D1

               -         +    -         +   -         +

Relapsed       12         6   16        2   13        5
  P value           NS             NS            NS
Disease free   30        13   41        2   35        8

                   cdk4

               -        +

Relapsed       2        16
  P value          NS
Disease free   1        42

* NS indicates not significant.
Table 5. Summary of the Expression of pRb and p16 in Sarcomas
as Reported in the Literature *

                                          pRb            p16

                         No. of     No.     Abnormal,    No.
     Source, Year        Samples   Normal    No. (%)    Normal

Cance et al, (16) 1990      44       13     31 (70.5)     ND
Karpeh et al, 1995         108       27     81 (75)       ND
Wang et al, (30) 1995       68       16     35 (53)       ND
Cohen and Geradts,
  (31) 1997                 59       45     14 (23.7)     41
Current study               67       19     48 (72)        4

                           p16

                         Abnorm-
                         al, No.   Criteria for Abnor-    Correlation
     Source, Year          (%)     mal Immunostaining     Documented

Cance et al, (16) 1990     ND      <80% of tumor cells   pRb and
                                                           survival
Karpeh et al, 1995         ND      <80% of tumor cells   Absence
Wang et al, (30) 1995      ND      <75% of tumor cells   pRb and grade
Cohen and Geradts,
  (31) 1997              16 (28)   <25% of tumor cells   Absence
Current study            63 (94)   <80% of tumor cells   Absence

* ND indicates not done.

References

(1.) Enzinger FM, Weiss SW, eds. Soft Tissue Tumors. 3rd ed. St Louis, Mo: Mosby; 1988.

(2.) Cordon-Cardo C. Mutation of cell cycle regulators: biological and clinical implications for human neoplasia. Am J Pathol. 1995;147:545-560.

(3.) Sherr CJ. Cancer cell cycles. Science. 1996;274:1672-1677.

(4.) Vogelstein B, Kinzler KW. P53 function and dysfunction. Cell. 1922;70:523-526.

(5.) Clarke AR, Purdie CA, Harrison DJ, et al. Thymocyte apoptosis induced by p53-dependent and independent pathways. Nature. 1993;362:849-852.

(6.) Haffner R, Oren M. Biochemical properties and biological effects of p53. Curt Opin Genet Dev. 1995;5:84-90.

(7.) Momand J, Zambetti GP, Olson DC, George DL, Levine AJ. The mdm2 oncogene product forms a complex with the p53 protein and inhibits p53 mediated transactivation. Cell. 1992;69:1237-1245.

(8.) Oliner JD, Pietenpol JA, Thiagalingam S, Gyuris J, Kinzler KW, Vogelstein B. mdm2 conceals the activation domain of p53. Nature. 1993;362:857-860.

(9.) Barak Y, Juven T, Haffmer R, Oren M. mdm2 expression is induced by wild type p53 activity. EMBO J. 1993;12:461-468.

(10.) Zauberman A, Flusberg D, Haupt Y, Barak Y, Oren M. A functional p53-responsive intronic promoter is contained within the human mdm2 gene. Nucleic Acids Res. 1995;23:2584-2592.

(11.) Weinberg RA. The retinoblastoma protein and cell cycle control. Cell. 1995;81:323-330.

(12.) Dublin EA, Patel NK, Gillett CE, Smith P, Peters G, Barnes DM. Retinoblastoma and p16 proteins in mammary carcinoma: their relationship to cyclin D1 and histopathological parameters. Int J Cancer. 1998;79:71-75.

(13.) Coindre JM, Trojani M, Contesso G, et al. Reproducibility of a histopathologic grading system for adult soft tissue sarcoma. Cancer. 1986;58:306-309.

(14.) Bosari S, Viale G, Roncalli M, et al. p53 gene mutations, p53 protein accumulation and compartmentalization in colorectal adenocarcinoma. Am J Pathol. 1995;147:790-798.

(15.) Schneider BG, Gully ML, Eagan P, Bravo JC, Mera R, Geradts J. Loss of p16/CDKN2A tumor suppressor protein in gastric adenocarcinoma is associated with Epstein-Barr virus and anatomic location in the body of the stomach. Hum Pathol. 2000;31:45-50.

(16.) Cance WG, Brennan MF, Dudas ME, Huang CM, Cordon-Cardo C. Altered expression of the retinoblastoma gene product in human sarcomas. N Engl J Med. 1990;323:1457-1462.

(17.) Karpeh MS, Brennan MF, Cance WG, et al. Altered patterns of retinoblastoma gene product expression in adult soft-tissue sarcomas. Br J Cancer. 1995; 72:986-991.

(18.) Yoo J, Lee HK, Kang CS, Park WS, Lee JY, Shim SI. p53 gene mutations and p53 protein expression in human soft tissue sarcomas. Arch Pathol Lab Med. 1997;121:395-399.

(19.) Leach FS, Tokino T, Meltzer P. p53 mutation and mdm2 amplification in human soft tissue sarcomas. Cancer Res. 1993;52:2231-2234.

(20.) Cordon-Cardo C, Latres E, Drobnjak M, et al. Molecular abnormalities of mdm2 and p53 genes in adult soft tissue sarcomas. Cancer Res. 1994;54:794-799.

(21.) Kawai A, Noguchi M, Beppu Y, et al. Nuclear immunoreaction of p53 protein in soft tissue sarcomas: a possible prognostic factor. Cancer. 1994;73: 2499-2505.

(22.) Mousses S, Gokgoz N, Wunder JS, et al. p53 missense but not truncation mutations are associated with Low levels of [p21.sup.CIP1/WAF1] mRNA expression in primary human sarcomas. Br J Cancer. 2001;84:1635-1639.

(23.) Agarwal ML, Ramana CV, Hamilton M, et al. Regulation of p53 expression by the RAS-MAP kinase pathway. Oncogene. 2001;20:2527-2536.

(24.) Bernards R, Schackleford GM, Gerber MR, et al. Structure and expression of the murine retinoblastoma gene and characterization of its encoded protein. Proc Natl Acad Sci U S A. 1989;86:6474-6478.

(25.) Kamb A. Cell cycle regulators and cancer. Trends Genet. 1995;11:136-140.

(26.) Draper GJ, Sanders BM, Kingston JE. Second primary neoplasms in patients with retinoblastoma. Br J Cancer. 1986;53:661-671.

(27.) DerKinderen DJ, Koten JW, Nagelkerke NJD, Tan KE, Beener FA, Den Otter W. Non-ocular cancer in patients with hereditary retinoblastoma and their relatives. Int J Cancer. 1988;41:499-504.

(28.) Stratton MR, Williams S, Fisher C, et al. Structural alterations of the RB1 gene in human soft tissue sarcomas. Br J Cancer. 1989;60:202-205.

(29.) Reissmann PT, Simon MA, Lee WH, Slamon DJ. Studies of the retinoblastoma gene in human sarcomas. Oncogene. 1989;4:839-843.

(30.) Wang J, Coltrera MD, Gown AM. Abnormalities of p53 and p110RB tumor suppressor gene expression in human soft tissue tumors: correlations with cell proliferation and tumor grade. Mod Pathol. 1995;8:837-842.

(31.) Cohen JA, Geradts J. Loss of RB and MTS1/CDKN2 (p16) expression in human sarcomas. Hum Pathol. 1997;28:893-898.

(32.) Wunder JS, Czitrom AA, Kandel R, Andrulis IL. Analysis of alterations in the retinoblastoma gene and tumor grade in bone and soft-tissue sarcomas. J Natl Cancer Inst. 1991;83:194-200.

(33.) Dei Tos AP, Piccinin S, Doglioni C, et al. Molecular aberrations of the G1-S checkpoint in myxoid and round cell liposarcoma. Am J Pathol. 1997;151: 1531-1539.

(34.) Dei Tos AP, Maestro R, Doglioni C, et al. Tumor suppressor genes and related molecules in leiomyosarcoma. Am J Pathol. 1990;148:1037-1045.

(35.) Orlow I, Drobnjak M, Zhang ZF, et al. Alterations of INK4A and INK4B genes in adult soft tissue sarcomas: effect on survival. J Natl Cancer Inst. 1999; 91:73-79.

(36.) Medema RH, Herrera RE, Lam F, Weinberg RA. Growth suppression by p16ink4 requires functional retinoblastoma protein. Proc Natl Acad Sci U S A. 1995;2:6289-6293.

(37.) Bates S, Parry D, Bonetta L, Vousden K, Dickson C, Peters G. Absence of cyclin D/cdk complexes in cells lacking functional retinoblastoma protein. Oncogene. 1994;9:1633-1640.

(38.) Khatib ZA, Matushime H, Valentine M, Shapiro DN, Sherr CJ, Look AT. Coamplification of the CDK4 gene with MDM2 and GLI in human sarcomas. Cancer Res. 1993;53:5535-5541.

(39.) Berner JM, Forus A, Elkahloun A, Meltzer PS, Fodstad O, Myklebost O. Separate amplified regions encompassing CDK4 and MDM2 in human sarcomas. Genes Chromosom Cancer. 1996;17:254-259.

Accepted for publication November 9, 2001.

From the Department of Pathology, St Vincent's Hospital, Catholic University, South Korea.

Reprints: Seok Jin Kang, MD, PhD, Department of Pathology, St Vincent's Hospital, Catholic University, Suwon, Kyungkido 442-060, South Korea (e-mail: [e-mail]).