alterations expression p62c-myc in squamous cell carcinoma ·...

9 JClin Pathol 1997;50:896-903

Papers

Alterations in exon 1 of c-myc and expression ofp62C-Myc in cervical squamous cell carcinoma

J J O'Leary, R J Landers, M Crowley, I Healy,W F Kealy, J Hogan, C T Doyle

Nuffield Departmentof Pathology andBacteriology,University of Oxford,Oxford, UKJJ O'Leary

Department ofPathology, Universityof Sheffield, Sheffield,UKR J Landers

Department ofPathology, UniversityCoilege Cork, IrelandM CrowleyI HealyW F KealyJ HoganC T Doyle

Correspondence to:Dr O'Leary, Department ofPathology, Cornell UniversityMedical College, 1300 YorkAvenue, New York, NY10021, USA; email:[email protected]

Accepted for publication2 September 1997

AbstractAims-To examine human papillomavirus(HPV) positive and negative squamouscell carcinomas of the cervix for struc-tural alterations in exon 1 c-myc; and toinvestigate the expression pattern of p62,the protein product of c-myc.Material-Archival paraffin wax embed-ded tissues of cervical squamous cell car-cinomas, stage I and II, retrieved from thefiles of the department of pathology,University College Cork, Ireland: 40 caseswere examined for alterations in exon 1 ofc-myc; 57 cases were used for immunocy-tochemical p62 analysis.Methods-c-myc exon 1 PCR on HPVpositive and negative stage I and IIcervical squamous cell carcinomas wasperformed using primers designed tofragile sites in exon 1 of the c-myconcogene, which are frequently involvedin translocation phenomena and deletionsin other neoplasms. This region is bor-dered by two promoter sequences PI andP2. In addition, the expression of p62 wasevaluated using the monoclonal antibodyMycl-9E10.Results-Alterations in exon 1 of c-mycwere shown in 7.5% of squamous cell car-cinomas of the cervix. Changes in exon 1and 2 of c-myc were also found in COLO320 cells and Raji cells. These alterationswere due to small deletions within exon 1ofc-myc, but point polymorphisms occur-ring within the priming sites (in one case)may also have occurred. The alterationsuncovered appeared "clonal," as replicatesamples showed the same amplicon bandpattern. Expression of c-myc was vari-able, with cytoplasmic staining patternspredominating. All cases which showedexon 1 alterations were HPV positive andhad strong nuclear positivity on p62immunocytochemistry.Conclusions-Alterations in exon 1 ofc-myc occur in a minority ofcervical can-cers and there was increased expression ofp62 in a cohort ofHPV positive and nega-tive cervical squamous cell carcinomas.Exon 1 alterations may provide an alter-

native route to c-myc activation in earlysquamous cell carcinoma.(J Clin Pathol 1997;50:896-903)

Keywords: c-myc; exon-1 alterations; cervical cancer;human papillomavirus

The c-myc oncogene is expressed in almost alldifferentiated cell types and the protein isthought to act as a transcriptional regulator.'The gene has been shown to have animportant role in early embryogenesis,' thecontrol of cell growth, cellular differentiation,3tissue repair processes,4 and apoptosis.' 6Expression of the gene is considered to be amajor component in the regulatory processesassociated with normal cell proliferation anddifferentiation. This view is reinforced by thedemonstration of gene amplification or anincrease in the level of the gene product p62 ina variety of pathological states in which there isperturbation of cell proliferation and differen-tiation. These include several malignant andpremalignant conditions.7 8 The myc genecontains a transcriptional activation domainand a basic helix loop leucine zipper DNAbinding and dimerisation domain. Acting as aheterodimer with a structurally related pro-tein, Max, myc can bind DNA in a sequencespecific manner, which acts as a transcrip-tional activator for many genes critical inregulation of cell growth.5 Max is not fullyrestricted to acting with myc and indeed formshomodimers and heterodimers with Mad andMxi- 1, which then act as antagonists tomyc/Max heterodimers, by competing forsimilar DNA targets. In apoptosis, mycinteracts with p53, Bax, and Bcl-x. Mycinduced apoptosis appears to occur in cellsdeprived of certain growth factors or wheregrowth has been arrested by cytotoxic drugs.The role of c-myc role in apoptosis requiresbinding to Max.9

Expression of c-myc protein has been shownto be closely associated with the entry of a cellfrom the Go or resting phase into the cell cycle,while almost complete shut off of c-mycexpression accompanies the inhibition of pro-liferation associated with differentiation.'0However, this finding is not universal, withother reports suggesting that c-myc protein

896

on April 14, 2020 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.50.11.896 on 1 Novem

ber 1997. Dow

nloaded from

http://jcp.bmj.com/

Expression ofc-myc in cervical squamous cell carcinoma

A

200...bp

1 kb

.: cl <1; (N (N (N

E ro~ ZI cU) XU) U) CL fn

Human c-myc gene

= UFx C)Cn MU) u -

m_ _

( no X co finU)U)><U)m)Ut~~~~~~ -~~~~~~~~~; tm~~~~~~~~~

B P1 P2

C-1 C-2 C-3 C-4 C-0

Figure 1 Schematic representation of the human c-myc gene. Exons are indicated as solid black boxes. The position ofprimers in exon 1 is indicated in (B).

synthesis is independent of the cell cycle."Amplification of this oncogene has beenobserved in several human cancer cell lines(including cervix) and in solid tumourspecimens.'2 Structural changes in the c-myc

gene and abnormal expression of the c-mycproduct have been observed in numerous

tumours, particularly B and T cellmalignancies.'3The human cellular myc gene is made up of

three exons separated by two large introns(fig 1). The first exon codes for a leader whichcontains termination codons on all three read-ing frames. The c-myc gene has two promoterregions (P1 and P2) and two transcription-initiation sites. Cellular myc sequences are

transcribed in normal cells. Expression ofc-myc is controlled during the cell cycle innon-transformed cells, but this control is lostafter chemical transformation. Transcription ofthe c-myc gene increases markedly shortly afterstimulating resting cells to enter the GI phase ofthe cell cycle. In various types of carcinoma ithas been found that structural alteration ofc-myc is accompanied by increased quantitiesof the c-myc oncogene product. Some of thealterations more often affecting c-myc geneexpression in human cells are amplification,translocation, and somatic mutations.'4 '5 Geneamplification and high levels of c-myc expres-sion have been observed in HL-60 promyelo-cytic leukaemia cells, small cell lung carcino-mas, and in some breast tumours.'6Translocation activation ofthe c-myc oncogenehas been found in Burkitt's lymphoma and inan undifferentiated T cell lymphoma as a resultof specific chromosomal translocations. Thec-myc oncogene is translocated in differenttypes of tumours, the 8:14 translocation ofBurkitt's lymphoma being one of the beststudied. 17

In Raji cells (which contain the 8; 14 translo-cation), substantial alterations-deletions andpoint mutations-in the sequence of the firsttwo exons are seen. These changes probablyresult from somatic mutations occurring dur-ing or after the translocation event. Changesobserved in exon 2 ofc-myc result in an altered

c-myc protein product, whereas those in thefirst exon can result in either c-myc overexpres-

sion or higher stability of c-myc mRNA. 4

Exon-1 of the c-myc oncogene is important intranscriptional control of this gene and thusalterations at this site can lead to overexpres-sion of the c-myc gene or higher stability of thec-myc mRNA.

Interestingly, in COLO 320 cells, normaland truncated (lacking exon 1) c-myc genes

coexist and both are transcribed.'8 Turnoverstudies of c-myc mRNA show that the normalmRNA has an in vivo half life of approximately30 minutes, which appears to be approximatelysimilar to the turnover time of mRNA in lym-phoblastoid cells. However, the truncatedmRNA (in COLO 320 cells) is known to besubstantially more stable. This observation hasalso been made in Burkitt's lymphoma cell linewhich contains a translocated, truncated c-mycgene. Therefore truncation of the c-myc genecan cause the messenger RNA to be more sta-ble than the full size product, suggesting thatthis may indeed be a crucial factor in the acti-vation of the c-myc oncogene by exon 1 (suchas in chromosomal translocation). Exon 1 mayalso play a role in c-myc degradation mecha-nisms.

Amplification of c-myc oncogene remainsthe predominant route for c-myc activation incervical cancer'9"21 and appears to be importantin relation to prognosis in early carcinoma ofthe cervix.

In the light of this knowledge, we haveexamined a series of human papillomavirus(HPV) positive and negative squamous cellcarcinomas of the cervix for structural altera-tions in exon 1 c-myc and have investigated theexpression pattern of p62, the protein productof c-myc.

MethodsMATERIALS

Archival paraffin wax embedded tissues of cer-

vical squamous cell carcinomas, stage I and II,were retrieved from the files of the departmentof pathology, University College Cork, Ireland.Forty cases were examined for alterations in

v cn cuU U) U) U)

897

CZtn)



j.com/

J Clin P


ber 1997. Dow

nloaded from

http://jcp.bmj.com/

O'Leary, Landers, Crowley, Healy, Kealy, Hogan, et al

exon 1 of c-myc; 57 cases were used for immu-nocytochemical p62 analysis.

HPV IN SITU HYBRIDISATIONThree 5 ,um sections (selected from two differ-ent areas of the tumour) were cut onto APEScoated glass slides (PH106, C A Hendley,Essex, UK). Two areas from the included cer-vix were also examined.

Tissue dewaxing was carried out usingxylene and rehydration through a graded alco-hol series. Proteolytic digestion was done with0.5 mg/ml proteinase K in proteinase K buffer(50 mM Tris HC1 pH 7.6, 5 mM EDTA) at370C for 10 minutes. Slides were thenimmersed in phosphate buffered saline (PBS)containing 2 mg/ml glycine for five minutesand washed in PBS for a further five minutes.For alkaline phosphatase detection, slides wereimmersed in 20% (vol/vol) aqueous acetic acidat 40C for 15 seconds. For peroxidase detec-tion, slides were immersed in 3% sodiumazide/hydrogen peroxide to abolish endog-enous peroxidase activity. Slides were thenwashed in PBS for five minutes. A postfixationstep was not carried out. Dehydration of thesections then took place through gradedalcohols to water.22Cloned HPV probes (6, 11, 16, 18, 31, 33)

(genital types; gifts of Harald Zur Hausen)were labelled with biotin or digoxigenin using astandard nick translation protocol.22 Probeswere initially prepared at a concentration of200 ng/ml.The hybridisation buffer contained 2x SSC

(sodium chloride/sodium citrate), 5% dextransulphate, 0.2% dried milk powder containingno vegetable extracts, and 50% formamide.Approximately 10-20 ng of the appropriateprobe was used during standard in situhybridisation protocols.Medium stringency post-hybridisation

washings were initially applied using 2x SSC at600C for 20 minutes, followed by 0.2x SSC for420C at 20 minutes then O.lx SSC at roomtemperature for five minutes, and 2x SSC atroom temperature for five minutes. Higherstringency washes included 0.2x SSC at 55°Cand 600C for 10 minutes, and O.lx SSC atroom temperature, 42°C and 55°C.The hybridisation signal was detected using

one step, two step, or three step techniques, asdescribed previously.'8 Colorimetric detectionwash achieved using an NBT/BCIP substratefor alkaline phosphatase or aminoethylcarba-zole (Zymed kit for peroxidase, California,USA).22 Sections were counterstained with 2%methyl green for alkaline phosphatase detec-tion or haematoxylin for the peroxidase detec-tion system.

Tissue controls for non-isotopic in situhybridisation (NISH) included HPV positivecervical wart and myocardium (negative forHPV). Reaction controls included hybridisa-tion buffer on its own, biotin/digoxigeninlabelled plasmid sequences (pBR322), andirrelevant probe (herpes zoster virus, HZV).Labelled human placental DNA was used tocheck hybridisation efficiency.

Detection controls included omitting pri-mary or secondary antibody steps and additionof the colorimetric substrate only.

SOLUTION PHASE POLYMERASE CHAIN REACTIONHPV E6 DNA sequences were derived fromthe EMBL genetic sequence database.23 Oligo-nucleotide primers were synthesised on aPerkin Elmer Applied Biosystems DNA syn-thesiser (Perkin Elmer, Cheshire, UK), depro-tected, and stored in liquid ammonia at -20°C.During oligosynthesis, a biotin reporter mol-ecule with a 15 carbon atom linker arm wasadded to the 5' end of the oligonucleotideprobe which was subsequently used as aninternal probe to confirm product specificity.The nucleotide sequence of the primers are aspreviously detailed.'9

Controls included tissues HPV positive byNISH and HPV plasmids. Reaction controlsincluded omitting Taq DNA polymerase andrunning 3 actin PCR assays on each sample.

General HPV primers (which identify se-quenced and unsequenced virus) were alsoused for a pan HPV screen, unrestricted by thetype specific viruses chosen above.24Ten 5 ,um sections were cut from the tissue

blocks and placed in sterile Eppendorf tubes.Nucleic acid extraction was performed usingproteinase K (0.1-0.5 mg/ml) in proteinase Kbuffer (100 mM NaCl, 10 mM Tris HCI,25 mM EDTA, 0.5% sodium dodecyl sul-phate, pH 8.4). Proteinase K incubation wascarried out for three to five days at 37°C withadequate mixing of samples. Proteinase Kinactivation was then carried out at 94°C for 10minutes. DNA was purified using a standardphenol chloroform isoamylalcohol technique.Nucleic acid was precipitated using 3 Msodium acetate and ice cold ethanol. Strictanticontamination protocols were adoptedthroughout.For type specific HPV polymerase chain

reaction (PCR), the PCR solution consisted ofPCR buffer (50 mM KCI, 10 mM Tris HCIpH 8.3, 1.5 mM MgCl2, 0.01% gelatine,200 ,uM of each dNTP, 1.0 ,uM of each primer,2.5 units of AmpliTaq DNA polymerase, and100 ng of DNA template). Samples were thensubjected to 40 cycles of PCR in a PerkinElmer 480 DNA thermocycler. Cycling para-meters were as follows: 94°C for one minute,followed by 94°C for one minute, 55°C for twominutes, 72°C for three minutes x 40 cycles,with a final extension set for 72°C for five min-utes.For general primer PCR, amplification con-

ditions were similar except 3.5 mM MgCl2 wasused and an annealing temperature of 400Cwas applied.Type specific HPV PCR products were con-

firmed by dot blot hybridisation as previouslydescribed.23

PCR ANALYSIS OF EXON-1 C-MYCPCR for c-myc was performed using primersdesigned for fragile sites in exon 1 of the c-myconcogene which are frequently involved intranslocation phenomena and deletions in

898



j.com/

J Clin P


ber 1997. Dow

nloaded from

http://jcp.bmj.com/

Expression of c-myc in cervical squamous cell carcinoma

Table 1 Primer sequences for c-myc exon 1 polymerasechain reaction analysis

3' Universalprimer C-0 CCT GGT TTT CCA ATA CCC GAA

5' Primers C-1 ATG CGA GGG TCT GGA CGG CTGC-2 TCC TGC CTC GAG AAG GGC AGGC-3 AGC CGG T1TTCG GGG CTT TATC-4 AAG AGC CGG GCG AGC AGA GCT

Table 2 Amplifiedfragment size of exon I c-myc product

Size (use with universal 3'5'Amplimers Amplified region amplimer C-O) (bp)

C-i 2303-2882 580C-2 2388-2882 495C-3 2466-2882 417C-4 2550-2882 333

other neoplasms. This region is bordered bythe two promoter sequences P1 and P2 (fig 1).To eliminate fixation effects of formalde-

hyde, human placental DNA fixed in neutralbuffered formaldehyde was included in allassays. The samples had previously beenexamined with pyrurate dehydrogenase andglobin gene with no evidence of genome wideinstability.To test the validity of the exon-1 assay, we

used COLO 320 cells (ATCC), which containnormal and truncated exon-1 genes,'8 and Rajicells (ATCC) which contain three deletions inthe intervening sequence between exons 1 and2 and 25 base pair changes in exon 2 as

compared to published human c-myc

sequences.'5 However, the deletions and pointmutations described in Raji cells are not withinthe region amplified by outer primers C-0 andC-1.The PCR solution consisted of PCR buffer

(50 mM KCI, 10 mM Tris HCI pH 8.3,1.5 mM MgCl,2 0.01% gelatine, 200 jM ofeach dNTP, 1.0 ,uM of each primer (table 1),2.5 units of AmpliTaq DNA polymerase, and100 ng of DNA template). Samples were thensubjected to 40 cycles of PCR in a PerkinElmer 480 DNA thermocycler. Cycling para-meters were as follows: 94°C for two minutes,followed by 94°C for one minute, 62°C for oneminute, 72°C for one minute x 40 cycles, witha final extension set for 72°C for five minutes.The reaction was performed on two geographi-cal areas selected and in triplicate on eachsample to exclude PCR generated artefact.Twenty microlitres of the reaction product

were electrophoresed on a 2% agarose gel con-

taining 0.5 gg/ml ethidium bromide and visual-ised under ultraviolet transillumination, yield-ing product sizes as in table 2.

SEQUENCE ANALYSIS OF EXON-1 USING CYCLE

SEQUENCINGCycle sequencing using Taq FS has been foundto provide the best results when symmetricalPCR templates are used. When using sym-metrical PCR products, a purification step isneeded. In this study the Qiagen QIAquickprotocol system (Qiagen, West Sussex, UK)was used to purify PCR product according tothe following protocol.The amplified DNA fragment was excised

from the gel with a clean sharp scalpel. The gel

slice was weighed in a 1.5 ml Eppendorf tube.Approximately 100 mg of gel (100 gl equiva-lent gel volume) was used for each purification.Three volumes of buffer QX1 were added tothe gel slice. The mix was incubated at 50°Cfor 10 minutes. To aid dissolution of the slice,the tube was inverted and gently flicked, everytwo to three minutes during incubation. Onegel volume (100 gl) of isopropanol was addedto the mix and the pH of the solution checkedusing a pH strip. If pH was > 7.5, then 10 jt of3 M sodium acetate, pH 5.0, were added to themix and the pH rechecked. (The techniqueonly works when the pH of the solution is< 7.5.) A QIAquick column was placed in a2 ml collection tube and loaded with the sam-ple and centrifuged at 13 000 rpm for oneminute. The flow through was discarded andthe QIAquick column was replaced in the samecollection tube; 0.5 ml of buffer QIAquick wasadded to the column and centrifuged at 13,000rpm for one minute. The column was thenwashed with 0.75 ml of buffer PE (Qiagen kit)and centrifuged at 13 000 rpm for one minute.The flow-through was discarded and thecolumn was again spun for an additional oneminute at 13 000 rpm. The QIAquick columnwas then placed in a clean 1.5 ml microfugetube and then 30 p1 of 10 mM Tris-HClpH 8.8, or HPLC water was added to the cen-tre of the column and centrifuged at13 000 rpm for one minute. The eluted DNAas then stored at -20°C. For sequencing reac-tions, 50-100 ng of purified product weretaken.A reaction premix was prepared as follows;

5x sequencing buffer, dNTP mix, A dye termi-nator, C dye terminator, G dye terminator, Tdye terminator, and AmpliTaq DNA polymer-ase FS (10 units). The amount of terminatoradded to the pre-mix was adjusted to give anoptimal distribution of signal between 10 and700 bases. If a stronger signal was required atthe beginning of the sequencing run, then theamount of terminator was adjusted upwards to0.6 p1 per reaction.

Positive and negative strand sequencing wasperformed on selected cases. Cycle sequencingwas performed using 3.2 pmol ofthe positive ornegative strand primer in the Gene Amp PCRsystem 9600 (Perkin Elmer) and the followingthermocycling profile was adopted: rapid ther-mal ramp to 96°C, 96°C for 10 seconds, rapidthermal ramp to 50°C, 50°C for five seconds,rapid thermal ramp to 60°C, then 60°C forfour minutes. The above thermocycling profilewas repeated for 25 cycles. At end ofthermocycling, the solution was rapidly cooledto 4°C and held at that temperature. Extensionproducts were then purified as follows:

In a 1.5 ml Eppendorf tube, 2 gl of 3 Msodium acetate pH 4.6, and 50 p1 of 95% etha-nol were added. The entire cycle sequencingsolution was then transferred to the 1.5 mlEppendorf tube. The mix was vortexed and thetube placed on ice for 10 minutes. The tubewas then spun at maximum speed in a benchmicrofuge for 15 to 30 minutes. The ethanolsolution was carefully aspirated for the tube.The pellet at the bottom of the tube was rinsed

899



j.com/

J Clin P


ber 1997. Dow

nloaded from

http://jcp.bmj.com/


Table 3 Cellular localisation of c-myc protein product p62,using Mycl-9E10

Predominant staining pattern

N>C N=C C>N

Squamous cell carcinoma of thecervix 4 9 44

Normal cervix (margin of tumourresected) 8 5 8

N>C, nuclear staining exclusively present in 90%+ of positivelystaining cells.C>N, cytoplasmic staining present in 90%+ of positively stain-ing cells.C=N, areas of exclusively nuclear and cytoplasmic staining,both of which are present in different areas of the section, butneither pattern accounting for 90% or more of the staining pat-tern.

by the addition of 250 pl of 75% ethanol. Theethanol solution was aspirated off without dis-turbing the cell pellet. The pellet was thendried in a vacuum centrifuge and resuspendedin 5 pl of water.The ABI PRISM 377 (Perkin Elmer) was

used for identification and resolution ofsequenced product. A loading buffer contain-ing deionised formamide and 25 mM EDTApH 8.0, containing 50 mg/ml blue dextran in aratio of 5:1 formamide to EDTA/blue dextranwas prepared.

Samples were then spun in a microfuge andvortexed briefly, after which they were heatedto 90°C for two minutes to ensure completedenaturation. The samples were then loaded

Figure 2 Two per cent agarose gel electrophoresis ofc-myc exon I amplicons in cervicalcarcinoma, normal ectocervical margins, andfixed human placental DNA showingdifferential size patterns (see textfor details).

using 1.5 p1 approximately of the reactionsolution and loading buffer as above. In allcases, a 48 cm WTR 4.25% T 19:1acrylamide:bis-acrylamide gel format wasused, giving resolution up to 800 bases.Samples were run over a 12 hour period andthen analysed using sequence navigator soft-ware on the ABI PRISM 377 DNA sequencer.Automated sequence comparison for polymor-phism analysis was achieved using the ABIsequence navigator and auto-assembled on theABI PRISM 377 DNA sequencer.

IMMUNOHISTOCHEMICAL DEMONSTRATION OFTHE C-MYC ONCOGENE PRODUCT (P62) INCERVICAL SQUAMOUS CELL CARCINOMASFifty seven squamous cell carcinomas of thecervix were examined for the expression ofc-myc protein product p62. In 21 cases,adjacent histologically normal ectocervicalmucosa was available for examination, whichserves as a control for the immunocytochemi-cal staining (table 3). Two 5 gm sections werecut onto APES coated slides. Sections weredewaxed in xylene and washed in absolutealcohol. Endogenous peroxidase activity wasquenched by incubation in 3% H202 in metha-nol for 15 minutes. The sections were thenrehydrated through graded alcohols to Trisbuffered saline (TBS, pH 7.6). Sections werewashed for five minutes and were thenincubated in 10% normal rabbit serum (NRS)followed by incubation in the mouse mono-clonal antibody Mycl-9E10 at a dilution of1:400. (This antibody is raised against asynthetic peptide of the C-terminus 408-432of the human c-myc gene product and is a spe-cific and sensitive probe for the protein onimmunoblots.25) Following incubation for onehour, sections were washed in TBS and treatedagain with NRS for 10 minutes. Sections werethen incubated with biotin conjugated rabbitantimouse IgG (Dako, Bucks, UK) 1:400 in10% NRS for 30 minutes. Sections werewashed in TBS and then incubated for 30minutes with Dako AB complex (which isfreshly prepared). The sections were finallywashed with TBS and the peroxidase reactionwas developed by the application ofDAB solu-tion (1 mg/ml diaminobenzidine in 50 mMTris-HCl pH 7.6, containing 10 mM imida-zole, activated with 1 jl of 100 vol H202 imme-diately before use). The reaction was allowed toproceed for two minutes, after which it wasstopped by washing in tap water. Sections werecounterstained with Mayer's haematoxylin,dehydrated, cleared, and mounted.

ResultsHPV PREVALENCEOf the 40 cases examined for alterations inexon 1 c-myc, 35 were HPV positive (35 HPV16; one HPV 18) and five negative for any HPVtype by general primer HPV PCR. There wasno discordance between HPV results withNISH and type specific PCR or general primerHPV PCR (data not shown).Of the 57 cases examined by immunocyto-

chemistry for p62 localisation, an additional 17cases were added to the first study cohort

900



j.com/

J Clin P


ber 1997. Dow

nloaded from

http://jcp.bmj.com/

Expression of c-myc in cervical squamous cellcarcinoma90

A~~~~~~~~Ar

B

4~~~~2

fm

$- .4 1.

Ai~ tr,-I.0, 0'

* 9~~~~~~I,av

d0

,~~~~~~~

..

Figure 3 (A) Area of controlCGIN 2showing p62 staining using Mycl-9El0. antibody. (B) p62 cytoplasmic stainingpattern in invasive cervical squamous cell carcinoma. (C) p62 nuclear staining pattern in invasive cervical squamous cellcarcinoma.

above, including 16 HPV 16 positive and oneHPV negative case.

ALTERATIONS IN EXON 1 OF C-MYCWe found alterations in exon 1 of the c-myconcogene in 7.5% (three of 40 cases); all threecases were HPV 16 positive. The non-HPVcases showed no alterations.No alterations were seen in formalin fixed

human placental DNA, which therefore ex-cludes a "fixative related" phenomenon, and nogenomic instability was seen using f3 globin orPDH primers. The duplicate geographicalsamples from each case showed the samealteration, indicating the "clonality" of theobservation. Control adjacent cervix in allthree case had germline DNA configuration.The triplicate PCR results performed on allsamples were concordant.

Figure 2 shows the alterations in some of thesamples examined. Normal exon-1 c-mycyields amplifiable bands of 580, 495, 417, and333 base pairs with the primers selected (asdetailed in tables 1 and 2). The formation ofsmaller fragments than predicted suggests thatdeletions have occurred upstream of thetargeted amplification site. Background faintbands (of the normal predicted size) representthe normal background cell population (forexample, fibroblasts, inflammatory cells, and soon).Sample A showed loss of the normal ampli-

con in lane 3. However, a background ampli-con was also visible in lane 3 from backgroundcells (fibroblasts, inflammatory cells, etc)Smaller amplicons than the predicted sizeswere seen in lanes 1 and 2 of sample A. Thissuggests that deletions had occurred in theregion amplified by C-0 and C-3 as detailedearlier. Sequence analysis showed a deletion of62 base pairs, involving the C-3 priming site.Sample B demonstrated loss of the normal

amplicon in lane 2. Smaller amplicons in lane 1were then found. This again suggests that adeletion had occurred within the region ampli-fied by C-0 and C-2. Sequence analysis showedan 80 base pair deletion involving the C-2priming site.

Sample C showed amplification of adjacentnormal cervix in sample B, with the formationof normal predicted size amplicons.

In sample D (one of the replicates of sampleB), the normal amplicon was missing in lane 2.A smaller amplicon than predicted was foundin lane 1. A faint background normal ampliconwas also seen in lane 1. This is due to amplifi-cation of background cells as described forsample A.Sample E showed complete amplification of

neutral buffered formaldehyde fixed humanplacental DNA. Similar results were obtainedwith COLO 320 and Raji cells.The other case showing an abnormality of

exon 1 by PCR did not give convincingevidence of deletion on direct sequencing,because of the high proportion of normal con-taminating cells in the biopsy. Tumours with anormal exon- 1 PCR pattern revealed thegermline c-myc sequence on positive andnegative strand sequencing (n = 4).

PATT7ERN OF EXPRESSION OF P62 IN CERVICAL

SQUAMOUS CARCINOMASTable 3 shows the pattern of staining obtainedwith the monoclonal antibody Mycl-9E 10. Thepredominant staining pattern achieved wascytoplasmic in the tumour samples examined,with nuclear staining seen in some cases (fig 3).In the normal cervical margin, nuclear stainingpredominated, which is in keeping with thepattern of staining in normal differentiatedsquamous epithelium from cervix and othersites. The localisation of c-myc product p62 inthe cytoplasm of malignant squamous epithe-lial cells is statistically significant (p <.0,xanalysis with Yates correction).

DiscussionThe c-myc gene is apparently set up to be rap-idly modulated, since both the correspondingm-RNA and protein have a short half life.T'herefore any decreases in c-myc m-RNAconcentration would naturally be followed rap-idly by a loss of c-myc protein. Overproductionof c-myc messenger RNA can result from gene

- j..-

.- I

901

JOA,



j.com/

J Clin P


ber 1997. Dow

nloaded from

http://jcp.bmj.com/


amplification. However, it is unlikely thatc-myc gene activation is achieved only throughthis mechanism.

Recently, it has been suggested that c-mycexon 1 sequences play an important role ineither transcriptional control or in post-transcriptional mRNA degradation.'8 26 It isknown that c-myc rearrangements can lead toaccumulation of c-myc transcripts. While innormal cells it might be necessary to provideexternal growth stimulation to induce c-mycmRNA appearance after each round of celldivision, the accumulation of c-myc messengerRNA and subsequently the c-myc protein bydeletions, mutations, or single base changes,may effectively stimulate tumour cells tocontinuous growth. This fact is borne out inthe findings that c-myc has been foundrearranged and amplified in most tumours, butnot in normal tissues, and it therefore seemsclear that alternations of this gene may beimportant in the genesis or progression of car-cinomas in the uterine cervix, as pointed out byOcadiz and colleagues.26 Gariglio et al'4 haveproposed that human c-myc gene expression isregulated by a post-transcriptional mechanism,and comment on substantial alterations in thesequence of the first two exons (both singlebase pairs and small deletions) in Raji cells.Changes in exon 2 can theoretically result inaltered c-myc protein product, whereaschanges in the first exon (single base pairchanges or small deletions) can result in eitherc-myc overexpression or higher stability of itsmessenger RNA.

In this paper, we demonstrate alterations inexon- 1 of c-myc in a minority of squamous cellcarcinomas of the cervix. Other changes at dif-ferent sites within exon 1 and 2 of c-myc havebeen described in COLO 320 cells and Rajicells. The alterations described in this paperare due to small deletions within exon-1, butpoint polymorphisms occurring within thepriming sites could not be excluded in the thirdcase. The findings are concordant with those ofOcadiz et a1 26 and Huang et a1 21 who have pre-viously described alterations in c-myc in cervi-cal carcinoma. To our knowledge, this is thefirst report in relation to exon 1, however.Importantly, the alterations uncovered ap-peared "clonal" as replicate samples showedthe same amplicon band pattern. PCR replica-tive error has been excluded in this experiment(see Methods). In addition, the changes appeartumour specific, as no alterations were found inthe adjacent normal cervix in these cases. Allcases which showed alterations were HPVpositive, although drawing a link between HPVand such alterations is highly speculative. Pre-viously, the integration sites ofHPV in relationto c-myc and N-myc have been examined.28 Allcases which show exon-1 changes had strongnuclear positivity on p62 immunocytochemis-try, suggesting that such alterations mayprovide an alternative route to c-myc activationin early squamous cell carcinoma.

Secondly, we have examined the immunohis-tochemical localisation of c-myc protein prod-uct p62 using the monoclonal antibody Mycl-9E10. Increased expression of the cellular

oncogene c-myc has been described in a varietyof human tumours. The gene encodes apolypeptide with predicated molecular weightof 49 kDa, but which shows aberrant electro-phoretic mobility on polyacrylamide gel elec-trophoresis to give an apparent molecularweight of around 62 kDa (p62 c-myc). Bio-chemical and immunofluoresence studies ofcell lines in vitro have indicated that the proteinis normally located in the cell nucleus. Its bio-logical half life is approximately 15 to 20 min-utes. Although p62 c-myc in vitro bindsnon-specifically to DNA, in vivo the greaterproportion of the protein appears to remainunbound, having a rather weak association withthe nuclear matrix.25 During mitosis, p62becomes dispersed throughout the cytoplasm,having no association with the condensedchromosomes. Apart from this situation, verylittle p62 is normally found in cytoplasm.The monoclonal antibody Mycl-9E10 used

in this study was raised against a synthetic pep-tide of residues 408-432 of p62 c-myc and hasbeen shown by its originators to recognise anepitope of the human c-myc protein p62.25 Ithas been proved to be sensitive in immunoblot-ting experiments. However, immunohisto-chemical studies with monoclonal Mycl-9E10and its closely related Mycl-6E10 have pro-duced conflicting results when used withformalin fixed tissues. Mycl-6E10 appears torecognise a protein present in both the nucleusand the cytoplasm. Suggestions for this unex-pected distribution of p62 have included poortissue fixation condition or extraction ofp62 bysalt or low temperatures. Alternatively, it ispossible that this antibody may cross react withan unrelated antigen in the cell.29

In this study, the cellular localisation ofc-myc protein was examined in 57 cases ofsquamous cell carcinoma of the cervix (includ-ing the cases examined for exon- 1 alterations).The predominant staining pattern was cyto-plasmic in 44 cases (table 3, fig 3). Equalnuclear and cellular staining was seen in ninecases, which consisted of areas of exclusivelynuclear and cytoplasmic staining, both ofwhich were present in different areas of thesection, neither pattern accounting for 90% ormore of the staining pattern. Greater nuclearthan cellular staining was seen in four cases.To serve as an internal control, the normalcervical margin in some of the excisedsquamous carcinomas (where available) wasused to assess the predominant stainingpattern in these regions. The pattern of resultsin the normal cervical margin reflects those innormal tissues, where there is a definite corre-lation between p62 nuclear localisation andnormally proliferating cells. Interestingly,Hughes et al 9 showed higher c-myc positivityin normal epithelium (63%) than in specimenswith cervical intraepithelial neoplasia grades 2or 3 (20%) or invasive carcinoma (0%), asassessed by flow cytometry. The finding ofp62localisation in normal relative to dysplasticepithelium may be a reflection of increasedturnover of protein in dysplastic cells. Pulacarzet al'0 have previously shown pancellularc-myc expression in endocervical mucosa in

902



j.com/

J Clin P


ber 1997. Dow

nloaded from

http://jcp.bmj.com/

Expression of c-myc in cervical squamous cell carcinoma

90% of glandular intraepithelial neoplasia andin 100% of adenocarcinomas.The results of our study again emphasise

that fixation procedures in immunolocalisationof the c-myc product must be taken intoaccount when analysing results. There appearsto be a shift from the normal nuclearlocalisation ofp62 to the cytoplasmic compart-ment in the malignant squamous ectocervicalcell. The precise nuclear association of p62 incarcinomas may be more labile because of pHchanges within the malignant cell, ATP con-centration, or release of proteolyic enzymes orother undefined cellular/nuclear factors. Onepossible explanation for this cytoplasmic shiftphenomenon may be a biologically significantdifference representing accumulation of newlysynthesised p62 in the cytoplasm of thesemalignant cells, or a relative or absolutedeficiency of an undefined nuclear transportmechanism for p62. In addition, the newlyformed protein may be unable to bind in thenucleus or its degradation or inactivation maybe accentuated.

CONCLUSIONSIn summary, we have shown alterations in exon1 of c-myc in a minority of cervical cancers,and increased expression of p62 in a cohort ofHPV positive and negative cervical squamouscell carcinomas. Alterations in exon 1 may pro-vide the malignant cell with an additional routeto activation/stabilisation of c-myc, independ-ent of amplification. Larger studies are nowunder way to determine the precise prevalenceof such alterations and whether similar phe-nomena occur in exon 2.

1 Bishop JM. Viral oncogenes. Cell 1985;42:23-38.2 Muller R, Slamon DJ, Tremblay J, Cline N, Verma IM. Dif-

ferential expression of cellular oncogenes during pre andpost-natal development of the mouse. Nature 1982;299:640-4.

3 Lachman HM, Skoultchi AI. Expression of c-myc changesduring differentiation of mouse erythroleukaemia cells.Nature 1984;310:592-4.

4 Goyette M, Petropoulos GJ, Shank PR, Fausto N.Expression of a cellular oncogene during liver regeneration.Science 1983:219:510-17.

5 Vastrik I, Makela TP, Koskinen PJ, Klefstrom J, Alitalo K.Myc protein: partners and antagonists. Crit Rev Oncol1994;5:59-68.

6 Amati B, Land H. Myc-Max-Mad: a transcription factornetwork controlling cell cycle progression, differentiationand death. Curr Opin Genet Dev 1994;4:102-8.

7 Sikora K, Evan G, Watson J. Oncogenes and germ celltumours. IntjAndrol 1987;10:57-67.

8 Ciclitria PJ, McCartney JC, Evan G. Expression of c-myc innon-malignant and pre-malignant gastro-intestinal disor-ders. JPathol 1987;151:293-6.

9 Evan G, Harrington E, Fanidi A, Land H, Amati B, BennettM. Integrated control of cell proliferation and cell death bythe c-myc oncogene. Philos Trans R Soc Lond Biol Sci 1994;345:269-75.

10 Einat M, Resnitzky D, Kimchi A. Close link between reduc-tion of c-myc expression by interferon and Go/ GI arrest.Nature 1985;313:597-600.

11 Hann SR, Thompson CB, Eisenmann RN. c-myc Oncogeneprotein synthesis is independent of the cell cycle in humanand avian cells. Nature 1985;314:366-9.

12 Yokota J, Tsunetsugu-Yokota Y, Battifora H, Lefevie C,Cline MJ. Alterations of myc, myb and ras-Ha proto-oncogenes in cancers are frequent and show clinical corre-lation. Science 1986;231:261-5.

13 Kelly K, Siebenlist U. The regulation and expression ofc-myc in normal and malignant cells. Annu Rev Immunol1986;4:317-38.

14 Gariglio P, del Angel R, Herrera A, Bonilla M, Theoreticalmodel for the post-transcriptional regulation of the humanc-myc gene expression, involving double-stranded RNAprocessing. J7 Theor Biol 1987;125:83-92.

15 Rabbitts TH, Hamlyn PH, Baer R. Altered nucleotidesequences of a translocated c-myc gene in Burkittlymphoma. Nature 1983;306:760-5.

16 Dalla-Favera R, Wong-Staal F, Gallo RC. Onc gene amplifi-cation in promyelocytic leukaemia cell line HL-60 and pri-mary leukaemia cells of the same patient. Nature 1982;299:61-3.

17 Rabbitts TH, Forster A, Hamlyn, Baer R. Effect of somaticmutation within translocated c-myc genes in Burkitt'slymphoma. Nature 1984:309:592-7.

18 Rabbitts PH, Forster A, Stison MA, Rabbitts TH. Trunca-tion of exon 1 from the c-myc gene results in prolongedc-myc mRNA stability. EMBOJ 1985;4:3727-33.

19 Riou G, Le MG, Le Doussal V, Barrois M, George M, HaieC. c-myc Proto-oncogene expression and prognosis in earlycarcinoma of the uterine cervix. Lancet 1987;ii:761-3.

20 Ocadiz R, Sauceda R, Salcedo M, Ortega V, Rodriguez H,Gordillo C, et al. Occurrence ofhuman papillomavirus type16 and c-myc alterations in uterine cervix carcinoma. ArchInvest Med Mexico 1989;20:355-62.

21 Munzel P, Marx D, Kochel H, Scjauer A, Bock KW.Genomic alterations of the c-myc protoncogene in relationto the overexpression of c-erbB2 and Ki-67 in humanbreast and cervix carcinoma. J Cancer Res Clin Oncol 1991;117:603-7.

22 O'Leary JJ, Browne G, Landers RJ, Crowley M, Healy I,Bashir MS, et al. Non-isotopic detection ofDNA in tissues.In: Levy ER, Herrington CS, eds. In-situ hybridisation. Apractical approach series. Oxford: Oxford University Press,1994:51-83.

23 Arends MJ, Donaldson YK, Duvall E, Wyllie AH, Bird CC.HPV in full thickness cervical biopsies: high prevalence inCIN 2 and CIN 3 detected by a sensitive PCR method. JIPathol 1991;165:301-9.

24 Van den Brule AJC, Snijders PJF, Gordijn RLJ, Bleker OP,Meijer CJLM, Walboomers JMM. General primer-mediated polymerase chain reaction permits the detectionof sequenced and still unsequenced human papillomavirusgenotypes in cervical scrapes and carcinomas. Int J Cancer1990;45:644-9.

25 Evan GI, Lewis GK, Ramsay G, Bishop JM. Isolation ofmonoclonal antibodies specific for the human c-mycproto-oncogene product. Mol Cell Biol 1985;5:3610-16.

26 Ocadiz R, Sauceda R, Cruz M, GraefAM, Gariglio P. Highcorrelation between molecular alterations of the c-myconcogene and carcinoma of the uterine cervix. Cancer Res1987;47:4173-7.

27 Huang SL, Liu ZQ, Zhang XN, Li MJ. Human papillomavirus, human cytomegalovirus and oncogene c-myc in cer-vical carcinoma and cervicitis. Chin Med J 1993;106:208-10.

28 Couturier J, Saster-Garau X, Schneider-Manoury S, LabibA, Orth G. Integration of papillomavirus near myc genes ingenital carcinomas and its consequences for proto-onocogene expression. J Virol 1991;65:4534-8.

29 Hughes RG, Neill WA, Norval M. Papillomavirus andc-myc antigen expression in normal and neoplastic cervicalepithelium. J Clin Pathol 1989;42:46-5 1.

30 Polacarz SV, Darne J, Sheridan EG, Ginsberg R, Sharp FEndocervical carcinoma and precursor lesions: c-mycexpression and the demonstration of field changes. J ClinPathol 1991;44:896-9.

903



j.com/

J Clin P


ber 1997. Dow

nloaded from

http://jcp.bmj.com/

alterations expression p62c-myc in squamous cell carcinoma ·...

Documents