66 Clin Pathol 1997;50:686-690

A comparative study of digoxigenin,2,4-dinitrophenyl, and alkaline phosphatase as

deoxyoligonucleotide labels in non-radioisotopicin situ hybridisation

S J Harper, E Bailey, C M McKeen, A S J Stewart, J H Pringle, J Feehally, T Brown

Richard Bright RenalUnit, SouthmeadHospital,Westbury-on-Trym,Bristol BSlO 5NB, UKS J Harper

Department ofNephrology, LeicesterGeneral Hospital,Gwendolen Road,Leicester LE5 4PW,UKE BaileyJ Feehally

Oswel Laboratory,Medical and BiologicalSciences Building,University ofSouthampton,Boldrewood, BassettCrescent East,Southampton SO167PX, UKC M McKeenA S J StewartT Brown

Department ofPathology, Universityof Leicester, ClinicalSciences Building,Leicester RoyalInfirmary, Leicester,UKJ H Pringle

Correspondence to:Dr Harper.

Accepted for publication13 May 1997

AbstractAim-To determine the optimum form oflabelling and the most efficient reportermolecule for non-radioisotopic in situhybridisation (ISH).Methods-Nine deoxyoligonucleotidescomplementary to histone mRNA weresynthesised and labelled either enzymati-cally or during solid-phase synthesis withthe reporter molecules digoxigenin, 2,4-dinitrophenyl (DNP), or alkaline phos-phatase. Pooled deoxyoligonucleotidecocktails were then used in non-radioisotopic ISH detection of histonemRNA in human tonsil. Hybrid detectionwas by nitroblue tetrazoleum/5-bromo-4-chloro-3-indolyl phosphate colorimetricdevelopment.Results-The use of a spacer in 3' enzy-matic labelling and when labelling withalkaline phosphatase significantly in-creased ISH signal. The 3' and 5' labellingof oligonucleotides with triple DNPgroups during solid-phase synthesis pro-duced the strongest signal as determinedby the highest cell signal intensity andshortest development time.Conclusions-3' and 5' solid-phase label-ling with triple DNP groups produced thebest labelling for non-isotopic ISH usingdeoxyoligonucleotide cocktails.(T Clin Pathol 1997;50:686-690)

Keywords: digoxigenin; 2,4-dinitrophenyl; alkalinephosphatase; in situ hybridisation

In situ hybridisation (ISH) oligonucleotideprobes may be labelled with radioactive ornon-radioactive reporter groups. After theinitial description of ISH in 19691 radioactivelabelling became the norm, allowing sensitivesemiquantitative detection of nucleotide se-quences. However the use of such labels (3H,32p, 35S, 1251) is hazardous, expensive, and timeconsuming, furthermore these labels have alimited shelf life. In contrast, the use ofnon-radioactive reporter groups (biotin, fluoro-scein, digoxigenin, dinitrophenyl) is simple andsafe, costs are comparable, scaling up is easy,development time short, and shelf life long.Non-radioisotopic ISH can also besemiquantitative.'' The main disadvantageremains the relatively lower sensitivity ofnon-radioactive reporter groups. The sensitivity

ofnon-radioisotopic ISH using oligonucleotideprobes can be increased in a number of ways:using oligonucleotides that form efficienthybrids,4 using cocktails of a number ofnon-overlapping oligonucleotides to the sametarget species, and by using multiple reportergroups per oligonucleotide.' This study consid-ers the last of these options.

Biotin has frequently been used as anon-radioactive reporter molecule. However, ithas a number of drawbacks including the pres-ence of high levels of endogenous biotin insome normal tissues. Fluorescein has beenused as an oligonucleotide label, as a haptenrather than a fluorochrome,6 but in ourexperience multiple additions of this reportergroup results in unacceptable backgroundstaining (unpublished data). The plant steroiddigoxigenin is becoming increasingly used tolabel oligonucleotides for ISH protocols andthe reporter molecule 2,4-dinitrophenyl(DNP) has been shown to give comparableresults to that of biotin, fluorescein, anddigoxigenin when bound to filters.7 8 The tech-nology also exists to label directly the oligonu-cleotides with enzymes-for example, alkalinephosphatase, although the chemistry is timeconsuming.'Our experience suggests that hybridisation

results on crude filters may not translate to ISHas some labels produce significant backgroundstaining when applied to tissue sections. Theaim of this work was to develop an oligonucleo-tide cocktail with a range of different labels toassess the most sensitive labelling-detectionsystem for use in non-radioisotopic ISH. Wechose to study the expression of histone mRNAin human tonsils. This is a model system for thedetection of native human mRNA in freshhuman tissue. It accurately identifies cells in Sphase of the cell cycle"0 and we have found thisprobe cocktail robust and reliable.'0 1' In humantonsil the probe cocktail identifies S phase cellsin the light and dark zones of the germinal cen-tres, within the basal layers of the epithelium,and in the interfollicular zone. Owing to thelimitations of biotin and fluorescein, we choseto study the labels digoxigenin, DNP, and alka-line phosphatase.

Aliquots of unlabelled oligonucleotides werelabelled enzymatically with digoxigenin- 1 1-dUTP or DNP-dUTP using terminal nucleo-tidyl transferase, both with and without anucleotide spacer (dATP). In addition, batches

686

on February 24, 2020 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.50.8.686 on 1 August 1997. D

ownloaded from

Digoxigenin, 2,4-DNP1 andAP as deoxyoligonucleotide labels in non-radioisotopic ISH

Table 1 Enzymatic deoxyoligonucleotide labelling

Without nucleotide spacer With nucleotide spacer

Nanopure H20 64.4 p1 Nanopure H20 61.1 1Probe cocktail (500 ng) 5 pt1 Probe cocktail (500 ng) 5 g110 mM MnCl, 10lO 10 mM MnCl2 10lOp1 mM digoxigemin-1 1-dUTP* or 1 mM digoxigenin- 1 1 -dUTP*or

1 mM DNP-dUTP" 1.6 1l 1 mM DNP-dUTP'5 1.6 p11 mM dATP - 1mM dATP 3.3 pl1.2 M Na cocodylate 10 p1 1.2 M Na cocodylate 10 p12 mM 2-deoxyuridine-5- tetrasodium 2 mM 2-deoxyuridine-5-

triphosphate 5-pl tetrasodium triphosphate 5 p1Terminal deocynucleotidyl transferaset 4 p1 Terminal deocynucleotidyl

transferaset 4 pil

*220582, Boehringer Mannheim; t27-0730-02, Pharmacia.

of probe were labelled by chemical synthesiswith digoxigenin (three groups at the 5' end,one at the 3' end); DNP (triple groups at both3' and 5' ends (previous work has suggestedthat triple DNP groups give optimal signal"2));or alkaline phosphatase at the 5' end (with andwithout a hydrophilic spacer).Each batch of labelled probe was then

optimised individually for concentration, strin-gency, antibody label detection, and colorimet-ric development time. The batches were

compared in parallel experiments: one in whicheach slide was incubated until optimal signalwas attained (experiment 1), the other in whicha fixed development time was used (one hour;experiment 2). Preliminary experiments indi-cated that this development time would allowcomparisons using subsaturated signal.

Materials and methodsTonsils from elective tonsillectomy cases were

cut into 0.5 x 1 cm sections at removal andfrozen in liquid nitrogen cooled isopentane.Subsequently all specimens were stored in liq-uid nitrogen before sectioning.

PROBE AND PROBE LABELLING

Deoxyoligonucleotides complementary to thehistone genes H2b, H3, and H4 were used.Final probe cocktails contained nine oligonu-cleotides, three for each gene, as previouslydescribed."13 All deoxyoligonucleotideswere 30 bases long, this length has been estab-lished as a practical compromise betweenhybrid stability, hybrid specificity, cost ofsynthesis and purification, and tissue penetra-tion efficiency.'4

Digoxigenin and DNP labelling was accom-

plished either enzymatically (unlabelled se-

quences were kindly donated by Pathway Serv-ices Ltd, Leicester, UK) or at synthesis.Oligonucleotides were labelled with alkalinephosphatase both directly and separated by a

hexaethylene glycol spacer. The hexaethyleneglycol monomer was prepared according to themethod of Durand et al."5Enzymatic 3' labelling was achieved using a

method adapted from Pringle et al"6 either withor without a nucleotide spacer (dATP) (table1). The labelling reaction took place at 37°Cfor two hours. The DNP deoxyuridine triphos-phate (5- [2,4 dinitrophenylaminopentylcar-bonnylarninopropargyl ] - 2 - deoxyuridine - 5 -

tetra-sodium triphosphate) was synthesised byestablished procedures.'7

Chemical synthesis and solid-phase chemi-cal labelling was performed using establishedchemical methods (OSWEL DNA Service,Southampton, UK).

EVALUATION OF OLIGONUCLEOTIDESNon-radioisotopic ISHThe protocol for non-radioisotopic ISH wasadapted from that previously reported.2 3 1 18 19RNAse free reagents and glassware (diethyl-pyrocarbonate (DEPC) treated; Sigma, UKD5758) were used throughout. Sequential8 jM sections were cut at -20°C onto silanecoated slides and fixed in freshly prepared 4%paraformaldehyde for 10 minutes at 4°C.Slides were washed in phosphate bufferedsaline (PBS)/DEPC 3 x 5 minutes. Probecocktail (15 p1) was added to the tissue and acoverslip applied. Hybridisation took place at37°C for two hours. Hybridisation solutioncontained 500-2000 ng/ml of oligonucleotideprobe, 600 mM NaCI, 50 mM Tris pH 7.5,0.2% bovine serum albumin, 1% sodiumdodecyl sulphate, 1% polyvinylpyrrolidone(40 kDa), 1% Ficoll (400 kDa), 0.1% sodiumpyrophosphate, 5 mM EDTA, 10% dextransulphate, and 30% formamide.

Post-hybridisation washes were as follows:2 x standard saline citrate solution (SSC)/30%formamide (2 x 10 minutes at 37°C); 2 x SSCat room temperature (2 x 10 minutes); and 15minutes in blocking solution at room tempera-ture. (SSC consisted of 150 mM NaCl, 15 mMtrisodium citrate pH 7.0; blocking solutionconsisted of TBS, 0.1% Triton-X100, 3%bovine serum albumin.) Slides were incubatedwith alkaline phosphatase labelled detectionantibodies for 30 minutes. Either polyclonalanti-digoxigenin F(ab)2 antibody (1093274;Boehringer Mannheim, Germany) diluted1/600 in blocking solution, or polyclonalanti-DNP F(ab)2 antibody (donated by Dako,High Wycombe, UK). Specimens were thenwashed in blocking solution (2 x 5 minutes),immersed in buffer 3 (0.1 M Tris HCI pH 9.5,0.1 M NaCI, 0.05 M MgCl2) (1 x 10 min-utes), and then in substrate solution (44 mlnitroblue tetrazoleum (NBT; Sigma, 75 mg/mlin 70% dimethylformamide), 33 ml 5-bromo-4-chloro-3-indolyl phosphate (BCIP; Sigma,50 mg/ml in dimethylformamide) in 10 mlbuffer 3). After development all slides werewashed in running tap water for 10 minutes,counterstained with Mayer's haematoxylin andmounted in aqueous mounting medium.

All slides and probe cocktails were coded.Slides were stopped (with frequent viewing)after development of optimal signal (experi-ment 1) or after one hour (experiment 2). Inexperiment 2 cell signal intensity of the first 50histone mRNA positive cells visualised at highpower (x40 objective) while scanning the longaxis of the same identified germinal centre wasmeasured using a Nikon Optiphot microscopeand an image analysis system using a HitachiTK-1280E Video Camera, an Apple-Mac7100/80AB, and NIH-Image software.Comparisons between probes could there-

fore be made both in terms of cell signal inten-sity and in optimal development time.

687

on February 24, 2020 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.50.8.686 on 1 August 1997. D

ownloaded from

Harper, Bailey, McKeen, Stewart, Pringle, Feehally, et al

Table 2 Development times

Time toOptimal optimalprobe signal

Label Probe labelling nglml* (hours) Backgroundt

Digoxigenin Enzymatic 3' without 500 18 0spacer

Digoxigenin Enzymatic 3' with spacer 500 3.5 0Digoxigenin Solid-phase 3' and 5' 2000 10 0DNP Enzymatic 3'with spacer 500 3 0DNP Solid-phase 3' and 5' 500 2 0Alkaline phosphatase Solid-phase 1000 8 0Alkaline phosphatase

with spacer Solid-phase 1000 3.5 0

*ng of oligonucleotide; tscale of 0 to +++.

PROBE SPECIFICITYSpecificity of this probe cocktail is well definedand has been confirmed by northern blotting'"and in this study by the use of appropriate ISHnegative controls." These were: probe omis-sion, antibody omission, RNAse Al pretreat-ment, random oligonucleotide, and non-homologous probe cocktail with the same G-Ccontent as the histone probe cocktail.

STATISTICAL ANALYSISCell signal intensities were compared usingone-way analysis of variance and Student's ttest with the Bonferroni correction.

ResultsIN SITU HYBRIDISATIONAll aliquots of labelled probe gave positiveresults. Sharp, clearly defined NBT/BCIP sitesof hybridisation were visible in germinalcentres, interfollicular zone, and epithelium.Background staining was negligible in all cases(table 2). Negative controls gave no visible sig-nal on any occasion, this equated to a mean

background pixel level of 1 13 AU.

Figure 1 Cell signal intensity in non-radioisotopic in situ hybridisation for histone mRNAin human tonsil. Mean and standard error of values taken from first 50 cells identified in auniform transection through the dark and light zones of the same germinal centre.

DigoxigeninTerminal deoxynucleotidyl transferase 3' label-ling with the addition of a nucleotide spacersignificantly increased the sensitivity of the ISHreducing the time for optimal development to< 20% expected (table 2) and increasing thecell signal intensity above baseline by 3.2 times(p < 0.00 1) (fig 1). The use of a spacer alloweda significant reduction in the probe concentra-tion to one fifth that normally required withreasonable maintenance of signal intensity(data not shown). 3' and 5' chemical labellingwith digoxigenin did not improve the resultsover 3' labelling with spacer (fig 1).

2,4-dinitrophenyl3' labelling with DNP and dATP spacer gave asignificantly stronger cell signal intensity thansimilar labelling with digoxigenin-11-dUTP(p < 0.001) (figs 1, 2B and C). Furthermore, 3'and 5' chemical labelling with DNP at synthesisproduced the strongest signal of all, withapproximately a 20% increase in cell signalintensity over 3' enzymatic labelling at one hour(p < 0.00 1) (figs 1, 2C and D), and correspond-ingly the shortest incubation time (table 2).

Alkaline phosphataseIn this study alkaline phosphatase labelled oli-gonucleotides produced comparable ISH sig-nal as produced by haptens (figs 1 and 2).Inclusion of a hexaethylene glycol spacerbetween the oligonucleotides and alkalinephosphatase significantly increased ISH signalcompared with direct labelling (p < 0.001)(figs 1, 2E and 2F)

DiscussionNon-radioisotopic ISH is well recognised as atechnique for identifying mRNA species ofmoderate to high copy number. In thesecircumstances non-radioisotopic ISH pro-duces a clear definition between positive andnegative cells with excellent resolution andpreserved morphology."4 This is true for paraf-fin wax embedded, frozen, and decalcified,fresh or archival material.3 1819 Furthermore, astrong ISH signal fascilitates double labellingof cells allowing simultaneous cellidentification.3 18 19 However, rarer mRNA spe-cies present a much more significant problemto the non-radioisotopic technique particularlywith the study of archival material. In this studywe considered the nature of the label and itseffect on the sensitivity of non-radioisotopicISH.The present study provides information

concerning the relative efficiency of threeoligonucleotide labels; one, digoxigenin, incommon use and two, as yet, infrequently used.The data presented suggest that the sensitivityof 3' labelled oligonucleotides for ISH can besignificantly improved by the inclusion of aspacer in the labelling reaction. This improve-ment was maintained even at low probeconcentrations. However, chemical 3' and 5'digoxigenin did not appear to improve cellsignal intensity or development time. Incontrast with DNP, there is no commerciallyavailable phosphoramidite for the addition of

688

on February 24, 2020 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.50.8.686 on 1 August 1997. D

ownloaded from

Digoxigenin, 2,4-DNP, andAP as deoxyoligonucleotide labels in non-radioisotopic ISH

A

4 4~~~~~~~4

.44 4 44

444 £4. .4

4, 4444.44 4 -. .,, 44.44 - .' 4: " .'.. ,44 4444 . -. '.. 4, 4 .744.4444,.*4 44444444 444w 4/4 .4 444'4 -.4. 4%44 444444 4 4.4 4444444. ¶444 44 444 .444 44.44. 444 .444444$44

444 4.4 444444 . -:.4 4,4 4 4 4 4

4 44 44

o 44 4, D

.44*;. - a

a .*444fl4 A 444)

(4 &.'d 4;t"'.'.C 9 44 1* LV!rt't)?\ a. 1

$ 7A'"'5..."j. I,444 4...4r¶ I

e '4-5aJ '444 'Fm:44. .25a.rn r4.q..#,' "4L. 4.4. 5 .4

44) 4''' 'S*4 ..r'-'. *1*7*4444

.4

'4

* A 9

*1

C4 4 4! 44. . 4'

4444) i4fl. ' tK*-' w4. '5.4

44.4' 444.. '44¶. . 444 St¶4 A 44 44 -USa

E

74t

q. 11

-..i$ .11t..P..1, k---' j."

'a.,,-,.

44

A ;i.

A.-.I

-.1

.1 -. 4.I 1.

%..

-R.. ir

r -, 4

-444

4 44 .3444 4¶ v-

* t,4r. '. * 444%t 4444

S4 .4. ,-. 4

.4.4.

.4 .44

44.4

A.*44.

.4*14. 4.4

V b

-.

4-p

A

F ,, 2 . 4

4444 44

t 4.44444A. w.½. ))4 444

* t4.t*4

4.4

4- *'

4.4S 44944

*44

1)'44

44t/ P444

-. .3 44 '44)¶ Ar:'

0

It4 "4 ..

44 . A.

44 7 4

* 4 4.4 4-

*4444

444* 44 44

Figure 2 In situ hybridisation demonstration of histone mRNA in tonsillar germinal centre. NBT/BCIP visualisation; haematoxylin counterstai'n.(A) Haematoxylin and eosin stain; (B) 3' digoxigenin labelling with nucleotide spacer; (C) 3'DNP labelling with nucleotide spacer; (D) 3' and 5'

chemical DNP labelling; (E) direct labelling of oligonucleotides with alkaline phosphatase; (F) alkaline phosphatase labelling with hydrophilic spacer.

689

4 4

I

I

VI'

u

.. I

,4I

4..

a

OA

6

44044t

..I

-1

4

F.

il

on February 24, 2020 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.50.8.686 on 1 August 1997. D

ownloaded from

Harper, Bailey, McKeen, Stewart, Pringle, Feehally, et al

digoxigenin to oligonucleotides. The digoxi-genin moieties have to be added to amino-derivatised oligonucleotides after solid-phasesynthesis. The reaction between the digoxi-genin derived active ester and the amino-functionalised oligonucleotide is neither quan-titative nor clean. High performance liquidchromatography purification removes the ma-jority of impurities and non-functionalised oli-gonucleotides, however, for multiple digoxi-genin labelling the product is not ashomogeneous as the equivalent DNP labelledoligonucleotide.

Oligonucleotides directly labelled with alka-line phosphatase gave comparable results tohapten labels despite the lack of an antibodydetection amplification step. Inclusion of ahydrophilic spacer in the alkaline phosphataselabel significantly improved the signal. Directlabelling may interfere with hybridisation ofnucleotides to target mRNA because of thelarge size of alkaline phosphatase.The best results were seen with the hapten

DNP. 3' labelling with DNP-dUTP gave betterresults than 3' labelling with digoxigenin-11-dUTP. Furthermore 3' and 5' labelling withtriple DNP groups gave better cell signalintensity and a shorter development time thanany other labelling protocol. In addition, 3' and5' chemical labelling provides a large quantityof uniformily labelled probe at a reasonableprice. This is in contrast to the labelling of oli-gonucleotides with haptens by enzymaticmethods that produces relatively small quanti-ties and has a variable efficiency. The DNPphosphoramidite monomer used in the presentstudy is structurally similar to those previouslyreported'2 but is easier to produce in largequantities.

Previous reports have shown DNP to givecomparable results to digoxigenin on testfilters7 and in the ISH detection of viralsequences and mycobacterial DNA.5 8We havenow used it in the detection of native mRNA.Our findings suggest that DNP is a superiorlabel to digoxigenin in ISH. The increased sen-sitivity advances routine non-radioisotopic ISHfor the detection of nucleic acid sequences ofrelatively low copy number and facilitates dou-ble labelling methods to identify the cell underinvestigation.

In summary these studies demonstrate that3' and 5' solid-phase labelling with triple DNPreporter groups produce the best labelling forsensitive non-radioisotopic ISH.

EB is supported by the National Kidney Research Fund NoR40\2\94. We are grateful to Pathway Services Ltd for the gift ofunlabelled oligonucleotides and to Dako for the anti-DNPF(ab)2 antibody.

1 John HA, Birnstiel ML, Jones KW. RNA-DNA hybrids atthe cytological level. Nature 1969;223:582-7.

2 Harper SJ, Pringle JH, Gillies A, Allen AC, Layward L, Fee-hally J, et al. Simultaneous in situ hybridisation of nativemRNA and immunoglobulin detection by conventionalimmunofluorescence in paraffin wax embedded sections.JClin Pathol 1992;45:114-19.

3 Harper SJ, Pringle JH, Wicks ACB, Hattersley J, Layward L,Allen AC, et al. Expression of J chain mRNA in duodenalIgA plasma cells in IgA nephropathy. Kidney Int 1994;45:836-44.

4 Southern EM, Case-Green SC, Elder JK, Johnson M, MirKU, Wang L, et al. Arrays of complementary oligonucle-otides for analysing the hybridisation behaviour of nucleicacids. Nucleic Acids Res 1994;22: 1368-73.

5 Stevenson K, Walker CA, Gryzbowski J, Brown T, SharpJM. Detection of PCR products from Mycobacteriumavium subspecies paratuberculosis using oligonucleotidescontaining multiple 2,4 dinitrophenyl reporter groups. Bio-medical Peptides, Proteins and Nucleic Acids 1995;1: 17-20.

6 Smith MD, Healy E, Thompson V, Morley A, Rees JL. Useof in situ detection of histone mRNA in the assessment ofepidermal proliferation: comparison with the Ki67 antigenand bromodeoxyuridine incorporation. Br J Dermatol1995;132:359-66.

7 Davison A, Duckworth G, Vaman Rao M, McClean J,Grzybowski J, Potier P, et al. Synthesis and antibody medi-ated detection of 2,4-dinitrophenyl (DNP) labelled oligo-nucleotides. Nucleosides Nucleotides 1995;14: 1049-52.

8 Cubie HA, Molyneaux PJ, Shearman MJ, Gryzbowski J,Brown T. Dot-blot hybridisation assay for detection of par-vovirus B19 infections using synthetic oligonucleotides.Mol Cell Probes 1995;9:59-66.

9 Urdea MS, Warner BD, Running JA, Stempien M, Clyne J,Horn T. A comparison of non-radioisotope hybridisationassay methods using fluorescent, chemiluminescent andenzyme labelled synthetic oligodeoxyribonucleotideprobes. NucleicAcids Res 1988;16:4937-56.

10 Alison M, Chaudry Z, Baker J, Lauder I, Pringle H. Liverregeneration: a comparison of in situ hybridisation for his-tone mRNA with bromodeoxyuridine labelling for thedetection of S-phase cells. JfHistochem Cytochem 1994;42:1603-8.

11 Jones PH, Harper SJ, Watts F. Stem cell patterning and fatein human epidermis. Cell 1995;80:83-93.

12 Grzybowski J, Will DW, Randall RE, Smith CA, Brown T.Synthesis and antibody mediated detection of oligonucle-otides containing multiple 2,4-dinitrophenyl reportergroups. Nucleic Acids Res 1993;21: 1705-12.

13 Bains MA, Agarwal R, Pringle JH, Hutchinson RM, LauderI. Flow cytometric quantitation of sequence specificmRNA in hemopoietic cell suspensions by primer inducedin situ (PAINS) fluorescent nucleotide labelling. Exp CellRes 1993;208:321-6.

14 Pringle JH, Ruprai AK, Primrose L, Keyte J, Potter L, CloseP, et al. In situ hybridisation of immunoglobulin light chainmRNA in paraffin sections using biotinylated or hapten-labelled oligonucleotide probes. J Pathol 1990;162:197-207.

15 Durand M, Chevrie K, Chassignol M, Thuong NT, Mauri-zot JC. Circular dichroism studies of an oligodeoxyridonu-cleotide containing a hairpin loop made of a hexaethyleneglycol chain: conformation and stability. Nucleic Acids Res1990;18:6353-9.

16 Pringle JH, Primrose L, Kind CN, Talbot IC, Lauder I. Insitu hybridisation demonstration of poly-adenylated RNAsequences in formalin fixed paraffin sections using a bioti-nylated oligonucleotide poly d(T) probe. JfPathol 1989;158:279-86.

17 Davison A, McClean JM, Vaman Rao M, Brown T. Synthe-sis of oligonucleotides labelled with 2,4-dinitrophenylgroups at thymidine sites. Biomedical Peptides, Proteins andNucleic Acids. 1996;2: 1-6.

18 Harper SJ, Allen AC, Bene M-C, Pringle JH, Faure G,Lauder I, et al. Increased dimeric IgA-producing B cells intonsils in IgA nephropathy determined by in situ hybridisa-tion for J chain mRNA. Clin Exp Immunol 1995;101:442-8.

19 Harper SJ, Allen AC, Pringle JH, Feehally J. Increaseddimeric IgA-producing B cells in the bone marrow in IgAnephropathy determined by in situ hybridisation forJ chainmRNA. _J Clin Pathol 1996;49:38-42.

690

on February 24, 2020 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.50.8.686 on 1 August 1997. D

ownloaded from