THYROIDVolume 11, Number 12, 2001Mary Ann Liebert, Inc.

Expression of Eukaryotic Translation Initiation Factors 4E and 2a Correlates with the Progression

of Thyroid Carcinoma

Songtao Wang,1 Ricardo V. Lloyd,2 Michael J. Hutzler,1 Igor B. Rosenwald,1 Marjorie S. Safran,3

Nilima A. Patwardhan,4 and Ashraf Khan1

Cell growth and proliferation depend on protein synthesis that is regulated, in part, by two eukaryotic trans-lation initiation factors, eIF-4E and eIF-2a. These factors are transiently increased as normal cells respond togrowth factors and are constitutively elevated in transformed cells. In cultured cells, eIF-4E facilitates cell cy-cle progression by increasing the expression of cell cycle promoting proteins including cyclin D1. Our previ-ous study revealed elevated cyclin D1 expression in histologically more aggressive thyroid carcinomas as com-pared to conventional papillary carcinoma. We hypothesized that the increased cyclin D1 expression mightcorrelate with increased eIF-4E expression. We, therefore studied the expression of eIF-4E by immunohisto-chemistry in 25 cases of conventional papillary carcinoma (CPC) and 28 cases of aggressive thyroid carcino-mas (ATC), the latter included 11 tall cell/columnar cell variant of papillary carcinoma, 5 insular carcinomas,and 12 anaplastic carcinomas. We also analyzed the expression of eIF-2a in the same samples as this factor isusually regulated similarly to eIF-4E in cell culture models. Of the 25 CPC, 13 were eIF-4E positive (11 weaklyand 2 strongly), and 19 were eIF-2a positive (14 weakly and 5 strongly). Conversely, of the 28 ATC, 25 wereeIF-4E positive (4 weakly and 21 strongly), and 23 were eIF-2a positive (4 weakly and 19 strongly). There wasa significantly increased expression of both eIF-4E (p , 0.001) and eIF-2a (p , 0.001) in ATC compared to CPC,suggesting that these translation initiation factors may play a role in the progression of thyroid cancer.

1101

Introduction

NORMAL CELLS proliferate transiently in response to ap-propriate extracellular growth factors or mitogens and

readily withdraw into quiescence on cessation of growthstimuli. In contrast, transformed cells often become inde-pendent of extracellular growth factors and proliferate con-tinuously. Cell growth and proliferation rates depend criti-cally on the rate of protein synthesis. It has been shown thata 50% inhibition of protein synthesis is sufficient to com-pletely arrest cell replication (1–3). The upregulation of netprotein synthesis after mitogenic stimulation is because inpart, to increased expression and function of the eukaryotictranslation initiation factors, 4E and 2a (eIF-4E and eIF-2a)(4–10). The least abundant component of the eIF-4F complexis eIF-4E, which is responsible for binding the 59 cap struc-

ture that is present in virtually all eukaryotic mRNAs andtransferring mRNAs to the ribosomes. The two other sub-units of the eIF-4F complex are eIF-4A, a helicase responsi-ble for unwinding mRNA secondary structures, and eIF-4G,which holds the complex together and is responsible for ri-bosome binding (4). The eIF-2 initiation factor complex,which transfers initiator methionyl tRNA to the 40S riboso-mal subunit, is composed of eIF-2a associated with two otherproteins, eIF-2b and eIF-2g. The eIF-2a subunit is the rate-limiting component of the protein synthesis initiation factor2 (eIF-2) and the phosphorylation of eIF-2a by the interferon-inducible kinase (PKR) inactivates the whole eIF-2 complex(11). The expression of both eIF-4E and eIF-2a is transientlyincreased in normal cells when they leave the resting G0 pe-riod and proliferate in response to extracellular growth stim-uli (4–10). The expression of these factors is constitutively

1Department of Pathology, University of Massachusetts Medical School, Worcester, Massachusetts.2Department of Laboratory Medicine & Pathology, Mayo Clinic, Rochester, Minnesota.3Department of Medicine, University of Massachusetts Medical School, Worcester, Massachusetts.4Department of Surgery, University of Massachusetts Medical School, Worcester, Massachusetts.Presented as an abstract and platform presentation at the US and Canadian Academy of Pathology meeting, March 1999, San Francisco,

California.

high in oncogene-transformed and tumor cells (12–14). Im-portantly, overexpression of either eIF-4E or eIF-2a is suffi-cient to transform cultured cells to a malignant phenotype(4,14–17). Conversely, the tumorigenic properties of cancercells can be inhibited by antisense-RNA against eIF-4E, or byoverexpression of its inhibitor proteins, 4E-BP (see Sonen-berg [4], Rosenwald [14] and Sonenberg and Gingras [18] forreviews).

It has been found earlier that overexpression of eIF-4Ecauses increased expression of cyclin D1 in NIH 3T3 cells(9,10), and cyclin D1 is upregulated in majority of cases ofcolonic neoplasms that display elevated levels of eIF-4E (19).Cyclin D1 forms a multimer complex with cyclin-dependentkinases (CDK), which facilitate transition through the re-striction point of cell cycle by hyperphosphorylation and in-activation of the retinoblastoma tumor-suppressor protein(pRb). Hyperphosphorylation of pRb leads to the release ofE2F transcription factors and other proteins from their com-plexes with pRb, which activates genes coding for positiveregulators of cell proliferation (see Dirks and Rutka [20], Sherr[21], and Bartek et al. [22] for review). Therefore, increase incyclin D1 and induction of D1-dependent events may be oneof the multiple effects caused by elevated eIF-4E.

Conventional papillary carcinoma of the thyroid is a well-differentiated thyroid carcinoma, which usually has an in-dolent behavior and excellent prognosis. The tall cell andcolumnar cell variants, however, tend to occur in older pa-tients, grow to a larger size, and have greater incidence ofvascular invasion, extrathyroidal extension, and mortality(23,24). Anaplastic carcinoma of thyroid often presents as afast-growing neck mass and is a highly malignant tumor. In-sular carcinoma is a poorly differentiated carcinoma with anintermediate morphology and clinical behavior between con-ventional papillary carcinoma and anaplastic carcinoma ofthe thyroid (25,26). Our recent study revealed elevated cy-clin D1 expression in histologically aggressive carcinomas ascompared to conventional papillary carcinoma of the thy-roid (27). We hypothesized that an increased cyclin D1 ex-pression may correlate with the increased expression of eIF-4E. We therefore studied the expression of eIF-4E, and alsoof eIF-2a by immunohistochemistry in conventional papil-lary carcinoma and aggressive thyroid carcinomas includingtall cell and columnar cell variants of papillary carcinoma,insular and anaplastic carcinomas, because both these trans-lation factors are regulated similarly in cell culture models(13,14).

Materials and Methods

Tissue specimens studied

Twenty-five cases of conventional papillary carcinoma,and 28 cases of more aggressive carcinoma of the thyroid in-cluding 11 tall cell/columnar cell variant of papillary carci-noma, 5 insular carcinoma, and 12 cases of anaplastic carci-noma were selected from the surgical pathology files ofUMass Memorial (University Campus) Medical Center,Worcester, Massachusetts (diagnosed between 1992 and1997), and Mayo Clinic, Rochester, Minnesota (diagnosed be-tween 1960 and 1993). All cases were reviewed by at leasttwo pathologists to confirm the diagnoses. The diagnosis ofpapillary carcinoma and aggressive carcinomas of the thy-roid was based on the growth patterns, nuclear, and cyto-

plasmic features (28). Paraffin blocks selected for immuno-histochemical staining required both tumor and a rim of nor-mal thyroid tissue around it, the latter to serve as a control.Therefore, normal thyroid tissue was seen in all 53 cases. ForWestern blot analysis, we obtained frozen tissue from nor-mal thyroid, two papillary carcinomas, one columnar vari-ant of papillary carcinoma, one insular carcinoma, and twoanaplastic carcinomas. Total protein from each specimen wasseparated by gel electrophoresis and analyzed with anti-eIF-4E and anti-eIF-2a antibodies.

Western blot protein analysis

Western blot analysis was performed as described previ-ously (19,29). Briefly, protein lysates were obtained by pul-verizing portions of normal thyroid and thyroid tumorsfrozen in liquid nitrogen. To prepare protein lysates, thefrozen tissue powder was resuspended with lysis buffer(0.5% Nonidet P-40, 420 mmol/L of NaCl, 20 mmol/L of Tris,pH 7.5, 2 mmol/L of phenyl-methylsulfonyl fluoride, and0.02 mmol/L of leupeptin). The suspension was forcefullypassed 10 times through an 18-gauge needle to disrupt cel-lular material. Each lysate was centrifuged at 10 3 103 rpmat 4°C for 15 minutes. The protein content in the supernatantswas analyzed by the Bradford assay (BioRad, Hercules, CA).Forty micrograms of total protein from each lysate were runon 8% sodium dodecyl sulfate (SDS)-polyacrylamide gel andblotted for 18 hours at 20 V at 4°C onto Immobilonpolyvinylidene difluoride (PVDF) membrane (Amersham,Arlington Heights, IL) using a liquid transfer solution con-taining 25 mmol/L Tris, pH 7.5, 192 mmol/L glycine, 10%methanol, and 0.01% SDS. The monoclonal mouse antibodyto eIF-4E (1:1000 dilution; Transduction Laboratories, Lex-ington, KY) and a monoclonal mouse antibody to eIF-2a(1:4000, gift from Dr. E. Henshaw, University of Rochester;Scorsone et al. [30]) were used sequentially, followed byhorseradish-peroxidase conjugated anti-mouse immu-noglobulin G (IgG) (1:3000 dilution; Promega, Madison, WI)to detect the corresponding proteins using the EnhancedChemiluminescence (ECL) developing system (Amersham).NIH 3T3 cells known to express high levels of eIF-4E andeIF-2a were used as a positive control.

Immunohistochemistry

All thyroid resection specimens in this study were fixedin 10% buffered formalin and paraffin-embedded by rou-tinely processing through a VIP Tissue Tek processor (MilesScientific, Naperville, IL). Sections were cut at a thickness of4 mm, heated at 60°C for 30 minutes, then deparaffinized,and hydrated through a series of xylene and alcohol prior tostaining. The slides were microwaved with a proprietaryantigen retrieval solution (citrate buffer) (BioTek Solutions,Santa Barbara, CA) for 5 minutes in an 800-W microwaveoven. After replenishment of this solution the slides weremicrowaved again for an additional 5 minutes and then al-lowed to cool for 20 minutes. Immunohistochemical stain-ing was performed with a monoclonal mouse antibody toeIF-4E (1:100 dilution; Transduction Laboratories) and amonoclonal mouse antibody to eIF-2a (1:2000; gift from Dr.E. Henshaw, University of Rochester, Scorsone et al. [30]). Astandard avidin/biotin complex (ABC) method as imple-mented on a Techmate 1000 (BioTek, Santa Barbara, CA) au-

WANG ET AL.1102

tomated immunostainer was used. The staining procedureconsisted of 45-minute incubation in the primary antibodyfollowed by brief buffer washes, and then incubation in acocktail of biotinylated anti-mouse IgG/IgM (BioTek) for 30minutes. The slides were then washed, incubated inavidin/biotin complex (BioTek) for 30 minutes, washed, andthen reacted with diaminobenzidine and hydrogen peroxideto visualize the end product. The sections were counter-stained with hematoxylin. For negative control, nonimmuneserum was substituted for primary antibody.

Evaluation of immunostaining

The immunostained sections were examined on an Olym-pus microscope (Tokyo, Japan) at 4003. The normal thyroidtissue in the same specimen usually showed barely de-tectable staining with antibodies to either eIF-4E or eIF-2a.Thus, these cells were used as an internal negative control.The cytoplasmic staining was graded based on the intensityof the immunoperoxidase staining as follows: 0, barely de-tectable staining, comparable to normal thyroid tissue; 11(weakly positive), if cells showed slightly more intense stain-

ing than normal thyroid tissue; 21 (strongly positive), if cellsstained much more intensively than the residual thyroid tis-sue. The blood vessels, which are strongly positive for eIF-2a, served as internal positive control for eIF-2a. Scatteredplasma cells, demonstrating strong staining for eIF-4E,served as internal positive control for eIF-4E.

Data analysis

The nonparametric Wilcoxon rank sum test was used toassess the different expression of eIF-4E and eIF-2a betweenpapillary carcinoma and aggressive carcinomas of the thy-roid (31). Statistical significance was set at the level of p ,0.05 (two-sided).

Results

Characterization of expression eIF-4E and eIF-2a innormal thyroid and thyroid carcinomas by Western blot

As illustrated in Figure 1, the antibodies used in this studyare specific for their targets. The anti-eIF-4E antibody reactswith a 25-kd protein and anti-eIF-2a antibody reacts with a36-kd protein (15,30). The specific recognition of eIF-4E andeIF-2a by the above antibodies was demonstrated in lym-phoid and colonic tissues as well (19,29). Because of prefer-ential expression of actin in the stromal component of ep-ithelial tumors (Rosenwald et al. [19] and data not shown)and variable expression of cytokeratin in thyroid tumors(data not shown), normalization to the total protein extractedfrom thyroid carcinoma cells and normal thyroid epitheliumcannot be achieved by using antiactin and anticytokeratinantibodies. Because both eIF-4E and eIF-2a are expressed instromal compartment weakly, and ratio of total protein ex-tracted from epithelial component to stromal componentmay be variable in cell lysates, the density of the Westernblot bands may not represent the true concentration of the

eIF-4E AND 2a IN THYROID CANCER PROGRESSION 1103

FIG. 1. Western blot analysis of eIF-4E and eIF-2a in nor-mal thyroid, conventional papillary carcinoma, and aggres-sive carcinomas of the thyroid. 3T3, NIH 3T3 cells with ac-cumulation of eIF-4E and eIF-2a; A; anaplastic carcinoma; I,insular carcinoma; C, columnar variant of papillary carci-noma; P, papillary carcinoma; NT, normal thyroid.

FIG. 2. Immunostaining for eIF-4E showing weak staining (11) in papillary carcinoma (A), and strong staining (21) inthe tall cell variant of papillary carcinoma (B), insular carcinoma (C), and anaplastic carcinoma (D). The residual thyroidtissue (B, top right corner, and C, bottom right corner) demonstrates negative (0) staining. A few benign glands adjacent tothe tall variant of papillary carcinoma show strong expression of eIF-4E (B, arrow). Immunoperoxidase 1003.

initiation factors in normal thyroid and its neoplasms. There-fore, Western blot results do not allow us to draw a reliableconclusion regarding the levels of eIF-4E and eIF-2a in thy-roid tumors. However, the results do validate the specificrecognition of eIF-4E and eIF-2a in thyroid epithelium andthyroid tumors.

Immunohistochemical Expression of eIF-4E and eIF-2a inpapillary and aggressive carcinomas of the thyroid

We studied the expression of translation initiation factorsin histologically indolent thyroid carcinoma (conventionalpapillary carcinoma). Thirteen and 19 of 25 conventionalpapillary carcinomas were eIF-4E and eIF-2a positive, re-spectively. Of these, 11 (44%) were 11 (Fig. 2A) and 2 (8%)

were 21 eIF-4E positive; 14 (56%) were 11 (Fig. 3A) and 5(20%) were 21 eIF-2a positive.

When the expression of eIF-4E and eIF-2a in several typesof histologically aggressive carcinomas of the thyroid wasstudied, 7 and 8 of 11 cases of tall cell/columnar cell variantof papillary carcinoma were strongly positive (21) for eIF-4E (Fig. 2B) and eIF-2a (Fig. 3B), respectively. Only 3 and 1of 11 cases were weakly positive (11) for eIF-4E and eIF-2a,respectively. It is interesting to note the paracrine effect ofthe tumor on eIF-4E and eIF-2a in normal tissue close to tallcell variant of papillary carcinoma as occasional residual nor-mal thyroid follicles closest to the tumor demonstrated ele-vated expression of both initiation factors (Figs. 2B and 3B).All the cases of insular carcinoma were strongly positive foreIF-4E (Fig. 2C) and 4 of 5 cases were strongly positive foreIF-2a (Fig. 3C). Only one insular carcinoma weakly ex-

WANG ET AL.1104

TABLE 1. EXPRESSION OF eIF-4E AND 2a IN THYROID CARCINOMAa

ConventionalStaining papillary

Proteins intensity carcinoma TCV/CCV IC AC

0 12 1 0 2eIF-4E 1 11 3 0 1

2 2 7 5 90 6 2 0 3

eIF-2a 1 14 1 1 12 5 8 4 7

11 5 12

# cases 25 28

aThe nonparametric Wilcoxon rank-sum test (31) was used to assess the difference in expression of eIF-4E (p, 0.001) and eIF-2a (p , 0.001) between conventional papillary carcinoma and aggressive carcinomas of the thyroid, the latter group include TCV/CCV, IC and AC.TCV/CCV, tall cell/columnar variants of papillary carcinoma; IC, insular carcinoma; AC, anaplastic carcinomaStaining Intensity: 0, barely detectable staining comparable to normal thyroid tissue; 11 (weakly positive), if cells showed slightly more

intense staining than normal thyroid tissue; 21 (strongly positive), if cells stained much more intensively than the residual thyroid tissue.

Aggressive thyroid carcinoma

FIG. 3. Immunostaining for eIF-2a showing weak staining (11) in papillary carcinoma (A), and strong staining (21) inthe tall cell variant of papillary carcinoma (B), insular carcinoma (C), and anaplastic carcinoma (D). The residual thyroidtissue (B, top right corner) demonstrate negative (0) staining. A few benign glands adjacent to the tall variant of papillarycarcinoma showed strong expression of eIF-2a (B, arrow). Immunoperoxidase 1003.

pressed eIF-2a. Nine of 12 and 7 of 12 of anaplastic carcino-mas were strongly positive for eIF-4E (Fig. 2D) and for eIF-2a (Fig. 3D), respectively. One and 2 of 12 cases were weaklypositive for eIF-4E and eIF-2a, respectively. Table 1, sum-marizes the results of immunostaining that shows that, mostcases of histologically aggressive carcinomas of the thyroiddemonstrated strong expression (grade 2) of eIF-4E and eIF-2a. On the other hand, most cases of conventional papillarythyroid carcinoma that are known to behave in an indolentfashion only displayed either similar (grade 0) or slight in-crease (grade 1) in the expression of these factors comparedto the residual normal thyroid. There was a significant dif-ference in both eIF-4E (p , 0.001) and eIF-2a (p , 0.001) ex-pression between papillary carcinoma and aggressive carci-nomas of the thyroid. It was noteworthy that tumor cells ofaggressive carcinomas of the thyroid often demonstrated un-even expression of eIF-4E and eIF-2a with stronger expres-sion in the periphery than in the central areas of the tumor(data not shown).

Discussion

Western blot analysis of normal thyroid and representa-tive cases of indolent and aggressive carcinomas of the thy-roid (Fig. 1) demonstrates that both eIF-4E and eIF-2a are ex-pressed in normal and neoplastic thyroid tissues and thatantibodies used in this study specifically recognize these fac-tors in thyroid tissue. Our immunohistochemical study in-dicates that aggressive carcinomas of the thyroid usually dis-play a strong increase in the expression of eIF-4E and eIF-2aas compared to normal thyroid epithelium. Levels of eIF-4Eand eIF-2a are higher in the periphery of aggressive carci-nomas of the thyroid than the main bulk of the tumors. Thisis consistent with the fact that the cells at the periphery ofmalignant tumors that also represent their invasive edge pro-liferate actively, while the central areas of the tumor usuallydo not proliferate as actively and may exhibit necrosis dueto a lack of sufficient supply of nutrients (32).

Our previous study has demonstrated that cyclin D1 ex-pression strongly correlates with aggressiveness of thyroidneoplasm (27). In this present study most of the aggressivethyroid carcinoma that showed strong (21) eIF2-a and eIF4estaining also showed strong (more than 50% of tumor cellnuclei staining positive) cyclin D1 staining (cyclin D1 dataof Wang et al. 27). Earlier studies have shown that overex-pression of eIF-4E causes elevation in the synthesis of cyclinD1 in cultured NIH 3T3 cells (9,10). Therefore, it is likely thatincreased expression of eIF-4E in thyroid tumors leads to el-evation of cyclin D1, activation of cyclin-dependent kinaseand hyperphosphorylation-dependent inactivation of pRb.Inactivation of pRb would allow release of E2F transcriptionfactor and activation of cell cycle promoting genes (20–22).Importantly, it has recently been discovered that pRb nega-tively regulates both transfer RNA (tRNA) and ribosomalRNA (rRNA) gene transcription (33). Therefore, eIF-4E–de-pendent elevation of cyclin D1 has a potential to bring aboutincreased production of rRNA and tRNA, further amplify-ing protein synthetic activity of the cell during tumor pro-gression. This hypothesis will be a subject of further inves-tigations.

Most aggressive thyroid carcinomas including the insularand anaplastic carcinoma arise from a preexisting well-dif-

ferentiated carcinoma, both papillary and follicular type (28).Several oncogene products including bcl-2 and the mutationor allelic loss of tumor suppressor gene p53 have been shownto be associated with the malignant progression of thyroidcarcinoma (34–38). The expression of bcl-2 has been observedin well-differentiated thyroid carcinoma and insular carci-noma, but significantly less in anaplastic carcinoma. Fur-thermore, in a subset of tumors having both differentiatedand anaplastic carcinoma, bcl-2 expression was restricted tothe residual differentiated component only; the anaplasticcomponent was negative (34,35). Of note, it has been shownthat eIF-4E prevents apoptosis in response to both growthfactor withdrawal and c-myc activation in cultured fibro-blasts without elevating bcl-2 protein level (39). It remainsto be determined if the initiation factor eIF-4E or eIF-2a pre-vent apoptosis in tumors, including aggressive carcinomasof the thyroid.

Mutation of p53 gene has been found in a high percent-age of anaplastic carcinomas but not in the residual papil-lary carcinoma component, suggesting that p53 mutation oc-curred after the development of the papillary carcinoma andmay play a key role in malignant progression of the tumor(34–38). Interestingly, wild-type p53 has been recently im-plicated in the negative regulation of ribosomal RNA pro-duction (40,41). Therefore, thyroid tumor progression maydepend on upregulation of multiple elements of protein syn-thetic machinery.

Our previous studies have demonstrated upregulation ofeIF-4E in colonic adenomas as an early event in malignanttransformation of the colon (19). In non-Hodgkin’s lym-phoma, however, increased expression of eIF-4E and eIF-2acorrelates with higher histologic grade (29). It appears thatthe upregulation of translation initiation factors eIF-4E andeIF-2a may, depending on the tissue, occur during the in-dolent stage or aggressive stage of tumorigenesis. The roleseIF-4E and eIF-2a play in the malignant progression of thy-roid tumors deserves further investigation. In various cellculture models eIF-4E preferentially increases synthesis ofspecific growth-promoting proteins, including cyclin D1,myc, ornithine decarboxylase, fibroblast growth factor (FGF)as well as proteins responsible for tumor angiogenesis (vas-cular permeability factor, FGF) and metastasis (v6 splice vari-ant of CD44 surface glycoprotein and collagenase type IV)(9,10,42–45). On its own, the general increase in protein syn-thesis caused by constitutively elevated eIF-4E and eIF-2aactivity, would facilitate accumulation of cellular proteins,accelerating growth and division rates. Previous findings re-veal that increased transcription of eIF-4E and 2a is inducedby c-myc expression in cultured fibroblasts (8,46). Furtherstudies are needed to determine which oncoproteins are re-sponsible for the elevated expression of eIF-4E and 2a in thy-roid carcinoma.

It has been suggested that increased expression and func-tion of initiation factors and the resulting constitutive in-crease in protein synthesis are central events in tumorigen-esis (14). The analysis of these factors in human neoplasmsremains limited. However, it has been found that eIF-4E maybe elevated in breast, and head and neck carcinomas (47,48)and both eIF-4E and eIF-2a are increased in colonic and lym-phoid neoplasms (19,29). Our present findings establish thatthe progression of the thyroid neoplasms is associated withelevated expression of eIF-4E and eIF-2a. It remains to be es-

eIF-4E AND 2a IN THYROID CANCER PROGRESSION 1105

tablished if these translation initiation factors could serve asmarkers for grading and progression of thyroid carcinomaand as targets for therapy.

References

1. Brooks RE 1977 Continuous protein synthesis is required tomaintain the probability of entry in to S phase. Cell12:311–317.

2. Epifanova OI 1977 Mechanisms underlying the differentialsensitivity of proliferating and resting cells to external fac-tors. Intern Rev Cytol 5(suppl):303–335.

3. Pardee AB, Dubrow R, Hamlin JL, Kletzien RF 1978 Animalcell cycle. Annu Rev Biochem 47:715–750.

4. Sonenberg N 1996 mRNA 59 cap-binding protein eIF4E andcontrol of cell growth. In: Hershey JWB, Mathews MB, So-nenberg N (eds) Translational Control. Cold Spring HarborLaboratory Press, Cold Spring Harbor, New York, pp245–269.

5. Morris DR 1995 Growth control of translation in mammaliancells. Prog Nucl Acid Res Mol Biol 51:339–363.

6. Cohen RB, Boal TR, Safer B 1990 Increased eIF-2 expressionin mitogen-activated primary T-lymphocytes. EMBO J9:3831–3837.

7. Mao X, Green JM, Safer B, Lindsten T, Frederickson RM,Miyamoto S, Sonenberg N, Thompson CB 1992 Regulationof translation initiation factor gene expression during hu-man T cell activation. J Biol Chem 267:20444–20450.

8. Rosenwald IB, Rhoads DB, Callanan L, Isselbacher KJ,Schmidt EV 1993 Increased expression of eukaryotic trans-lation initiation factors eIF-4E and eIF-2a in response togrowth induction by c-myc. Proc Natl Acad Sci USA90:6175–6178.

9. Rosenwald IB, Lazaris-Karatzas A, Sonenberg N, SchmidtEV 1993 Elevated expression of cyclin D1 protein in responseto increased expression of eukaryotic initiation factor 4E.Mol Cell Biol 13:7358–7363.

10. Rosenwald IB, Kaspar R, Rousseau D, Gehrke L, LeBoulchP, Chen J-J, Schmidt EV, Sonenberg N, London IM 1995 Eu-karyotic translation initiation factor 4E regulates expressionof cyclin D1 at transcriptional and post-transcriptional lev-els. J Biol Chem 270:21176–21180.

11. Trachsel H 1996 Binding of initiator methyionyl-tRNA to ri-bosomes. In: Hershey JWB, Mathews MB, Sonenberg N (eds)Translational Control. Cold Spring Harbor Laboratory Press,Cold Spring Harbor, New York, pp 113–139.

12. Rosenwald IB 1995 Growth factor-independent expressionof the gene encoding eukaryotic translation initiation factor4E in transformed cell lines. Cancer Letters 98:77–82.

13. Rosenwald IB 1996 Upregulated expression of the genes en-coding translation initiation factors eIF-4E and eIF-2a intransformed cells. Cancer Lett 102:113–123.

14. Rosenwald IB 1996 Deregulation of protein synthesis as amechanism of neoplastic transformation. Bioessays18:243–250.

15. Lazaris-Karatzas A, Montine KS, Sonenberg N 1990 Malig-nant transformation by an eukaryotic initiation factor sub-unit that binds to mRNA 59 cap. Nature 345:544–547.

16. De Benedetti A, Rhoads RE 1990 Overexpression of eu-karyotic protein synthesis initiation factor 4E in HeLa cellsresults in aberrant growth and morphology. Proc Natl AcadSci USA 87:8212–8216.

17. Donze O, Jagus R, Koromilas AE, Hershey JWB, SonenbergN 1995 Abrogation of translation initiation factor eIF-2 phos-phorylation causes malignant transformation of NIH 3T3cells. EMBO J 14:3828–3834.

18. Sonenberg N, Gingras AC 1998 The mRNA 59 cap-bindingprotein eIF4E and control of cell growth. Curr Opin Cell Biol10:268–275.

19. Rosenwald IB, Chen J-J, Wang S, Savas L, London IM, Pull-man J 1999 Upregulation of protein synthesis initiation fac-tor eIF-4E is an early event during colon carcinogenesis.Oncogene 18:2507–2517.

20. Dirks PB, Rutka JT 1997 Current concepts in neuro-oncol-ogy: The cell cycle. A review. Neurosurgery 40:1000–1015.

21. Sherr CJ 1996 Cancer cell cycles. Science 274:1672–1677.22. Bartek J, Bartkova J, Lukas J 1996 The retinoblastoma pro-

tein pathway and the restriction point. Curr Opin Cell Biol8:805–814.

23. Johnson TL, Lloyd RV, Thompson NW, Beierwaltes WH, Sis-son JC 1988 Prognostic implications of the tall cell variantof papillary thyroid carcinoma. Am J Surg Pathol 12:22–27.

24. Akslen LA, Varhaug JE 1990 Thyroid carcinoma with mixedtall-cell and columnar-cell features. Am J Clin Pathol94:442–445.

25. Carcangiu ML, Zampi G, Rosai J 1984 Poorly differentiated(“insular”) thyroid carcinoma. A reinterpretation of Lang-hans’ “Wuchernde Struma.” Am J Surg Pathol 8:655–668.

26. Aldinger KA, Samaan NA, Ibanez M, Hills CS Jr 1978Anaplastic carcinoma of the thyroid: A review of 84 casesof spindle and giant cell carcinoma of the thyroid. Cancer41:2267–2275.

27. Wang S, Lloyd RV, Hutzler MJ, Safran MS, Patwardhan NA,Khan A 2000. The role of cell cycle regulatory protein cyclinD1 in the progression of thyroid cancer. Mod Pathol13:882–887.

28. Rosai J, Carcangiu ML, DeLellis RA 1992 Atlas of TumorPathology, Third Series. In: Rosai J, Sobin LH (ed) Tumorsof the Thyroid Gland. Armed Forces Institute of Pathology,Bethesda, MD pp 65–159.

29. Wang S, Rosenwald IB, Hutzler MJ, Pihan GA, Savas L, ChenC-C, Woda BA 1999 Expression of the eukaryotic initiationfactors 4E and 2a in non-Hodgkin’s lymphomas. Am JPathol 155:247–255.

30. Scorsone KA, Panniers R, Rowlands AG, Henshaw EC 1987Phosphorylation of eukaryotic initiation factor 2 duringphysiological stresses which affect protein synthesis. J BiolChem 262:14538–14543.

31. Siegel S, Castellan NJ 1997 Nonparametric Statistics for theBehavioral Sciences. McGraw-Hill, Boston, pp 128–137,135–244.

32. Kuwano H, Saeki H, Kawaguchi H, Sonoda K, Kitamura K,Nakashima H, Toh Y, Sugimachi K 1998 Proliferative activ-ity of cancer cells in front and center areas of carcinoma insitu and invasive sites of esophageal squamous-cell carci-noma. Int J Cancer 78:149–152.

33. Larminie CGC, Alzuherri HM, Cairns CA, McLees A, WhiteRJ 1998 Transcription by RNA polymerases I and III: A po-tential link between cell growth, protein synthesis andretinoblastoma protein. J Mol Med 76:94–103.

34. Pilotti S, Collini P, Rilke F, Cattoretti G, Del Bo R, PierottiMA 1994 Bcl-2 protein expression in carcinomas originatingfrom the follicular epithelium of the thyroid gland. J Pathol172:337–342.

35. Pilotti S, Collini P, Del Bo R, Cattoretti G, Pierotti MA RilkeF, 1994 A novel panel of antibodies that segregates im-munocytochemically poorly differentiated carcinoma fromundifferentiated carcinoma of the thyroid gland. Am J SurgPathol 18:1054–1064.

36. Fagin JA, Matsuo K, Karmakar A, Chen DL, Tang SH, Ko-effler HP 1993 High prevalence of mutations of the p53 gene

WANG ET AL.1106

in poorly differentiated human thyroid carcinomas. J ClinInvest 91:179–184.

37. Ito T, Seyama T, Mizuno T, Tsuyama N, Hayashi T, HayashiY, Dohi K, Nakamura N, Akiyama M 1992 Unique associa-tion of p53 mutations with undifferentiated but not with dif-ferentiated carcinomas of the thyroid gland. Can Res52:1369–1371.

38. Nakamura T, Yana I, Kobayashi T, Shin E, Karakawa K, Fu-jita S, Miya A, Mori T, Nishisho I, Takai S 1992 p53 genemutations associated with anaplastic transformation of hu-man thyroid carcinomas. Jpn J Cancer Res 83:1293–1298.

39. Polunovsky VA, Rosenwald IB, Tan AT, White J, Chiang L,Sonenberg N, Bitterman PB 1996 Translational control ofprogrammed cell death: Eukaryotic translation initiation fac-tor 4E blocks apoptosis in growth-factor-restricted fibro-blasts with physiologically expressed or deregulated Myc.Mol Cell Biol 16:6573–6581.

40. Cairns CA, White RJ 1998 p53 is a general repressor of RNApolymerase III transcription. EMBO J 17:3112–3123.

41. Budde A, Grummt I 1999 p53 represses ribosomal gene tran-scription. Oncogene 18:1119–1124.

42. Shantz LM, Hu R-H, Pegg AE 1996 Regulation of ornithinedecarboxylase in a transformed cell line that overexpressestranslation initiation factor eIF-4E. Cancer Res 56:3265–3269.

43. Kevil C, Carter P, Hu B, De Benedetti A 1995 Translationalenhancement of FGF-2 by eIF-4 factors, and alternate uti-lization of CUG and AUG codons for translation initiation.Oncogene 11:2339–2348.

44. Kevil CG, De Benedetti A, Payne DK, Coe LL, Laroux FS,

Alexander JS 1996 Translational regulation of vascular per-meability factor by eukaryotic initiation factor 4E: Implica-tions for tumor angiogenesis. Int J Cancer 65:785–790.

45. Graff JR, Boghaert ER, De Benedetti A, Tudor DL, ZimmerCC, Chan SK, Zimmer SG 1995 Reduction of translation ini-tiation factor 4E decreases the malignancy of ras-trans-formed cloned rat embryo fibroblasts. Int J Cancer60:255–263.

46. Jones RM, Branda J, Johnston KA, Polymenis M, Gadd M,Rustgi A, Callanan L, Schmidt EV 1996 An essential E boxin the promoter of the gene encoding mRNA cap-bindingprotein (eukaryotic initiation factor 4E) is a target for acti-vation by c-myc. Mol Cell Biol 16:4754–4764.

47. Li BDL, Liu L, Dawson M, De Benedetti A 1997 Overex-pression of eukaryotic initiation factor 4E (eIF-4E) in breastcarcinoma. Cancer 79:2385–2390.

48. Nathan C-AO, Liu L, Li BDL, Abreo FW, Nandy I,DeBenedetti A 1997 Detection of the proto-oncogene eIF-4Ein surgical margins may predict recurrence in head and neckcancer. Oncogene 15:579–584.

Address reprint requests to:Ashraf Khan, M.D., FRCPath

Department of PathologyUniversity of Massachusetts Medical School

55 Lake Avenue NorthWorcester, MA 01655

E-mail: KhanA@ummhc.org

eIF-4E AND 2a IN THYROID CANCER PROGRESSION 1107