Quantitative in situ hybridization with enhanced sensitivity in soft , bone and tooth tissue using digoxigenin tagged RNA probes

Uvod: Kvantitativna neradioaktivna in situ hibridizacija je moćna tehnika za lokalizaciju ekspresije transkripata mRNA. Ove metode omogućavaju otkrivanje mRNA uz visoku rezoluciju na razini jedne jedine stanice. Dosad najšire opisani neradioaktivni protokoli rabili su deblje kriostatske rezove mekog tkiva i lokalizirali visoko zastupljene gene bez kvantifi ciranja. Materijali i metode: Mi smo razvili metodu neradioaktivne in situ hibridizacije pomoću tankih rezova demineraliziranog koštanog tkiva uklopljenog u parafi n, uz otkrivanje nisko zastupljenih gena i kvantifi kaciju in situ signala. Naš protokol zasniva se na optimalnoj sintezi digoksigeninom obilježenih DNA probi za vizualiziranje vrlo nisko zastupljenih gena u mekom, koštanom i zubnom tkivu uklopljenom u parafi n. Naša nova tehnika prikazuje in situ signal uz umnožavanje i pojačanje boje kroz reakciju alkalne fosfataze. Rezultati: Osjetljivost je na razini radioaktivnih protokola, a rezolucija je na razini pojedine stanice, što je bolje nego kod radioaktivnih protokola. Kvantifi ciranje in situ hibridizacijskog signala, što se ranije radilo pomoću radioaktivnog ili fl uorescentnog obilježavanja, sad je moguće s alkalnom fosfatazom pomoću naše tehnike i programa ImageJ. Zaključak: Prikazani primjeri pokazuju kako slijedeća metoda ima bolju dokazanu rezoluciju i jednaku osjetljivost kao radioaktivno obilježene metode u različitim tkivima s mogućom kvantifi kacijom; to pak pokazuje da se ova metoda općenito može rabiti u istraživačkim kao i u kliničkim laboratorijima. Ključne riječi: kvantitativna in situ hibridizacija, digoksigeninske probe, koštano tkivo uklopljeno u parafi n Abstract


Introduction
Experiments in clinical and research laboratories often require mRNA in situ localization of tissue specifi c biomarkers and their quantifi cation within specifi c cells in the tissue (1,2).Thirty-seven years have elapsed since the fi rst protocol for in situ was published (3).Since then, numerous papers have been published on this technique, which is now widely applied in basic and clinical research (4-7).There are publications describing that non-radioactive in situ hybridization can exhibit equivalent sensitivity as radioactive hybridization (8).Ideally, such techniques would allow for detection of very low abundant mRNAs at the cellular level but with sensitivity and quantifi cation capabilities of radioactive labeled probes 33 P or 35 S. Several laboratories prefer detection with colorimetric methods rather than fl uorescent; the reason is that the result is not sensitive to light, it will not fade away with time, and there is no rush with the pictures which can also be retaken without a fl uorescent microscope.The advantage of fl uorescent methods is co-localization of mRNA in situ since these methods permit very good detection of multiple probes and easy quantifi cation of hybridization signal.On the other hand, fl uorescence in situ hybridization has traditionally been conducted with diffi culties involving low signal, high background, and auto-fl uorescence, especially in bone tissue (9).Most of the published protocols for non-radioactive in situ hybridization were developed using thick cryosections (10)(11)(12), and few protocols with paraffi n embedded tissue (13,14).This approach was applied for temporal and spatial patterns in soft tissue with high abundant genes.For high resolution signal in thin sections of bone tissue and teeth using low abundant genes at a single cell level, more sensitive hybridization protocols and solutions were needed.Fortunately, the advent of the tyramide signal amplifi cation (TSA, PerkinElmer), a system which we applied to increase sensitivity, has resolved some of these issues (15).In this report, we provide a detailed protocol for detection of mRNA using signal amplifi cation and enhanced color developing techniques compared to two other different non-radioactive techniques and one radioactive in situ hybridization technique also developed and applied in our laboratory.These techniques are very sensitive and can be used on thin paraffi n sections of soft as well as demineralized tooth and bone tissues.These techniques can detect signals of low abundant genes with the same sensitivity of radioactive methods but at the level of single cell, which can be quantifi ed.

Materials and methods
Preparing the RNase-free DNA templates RNA probes for in situ hybridization were transcribed with an in vitro transcription reaction.Each diff erent cDNA probe construct inserted in a plasmid vector for in vitro transcription contained a T 3 , T 7 or SP 6 promoter.To transcribe each probe only from the desired cDNA (not from the whole vector) we cut our vectors and made linearized DNA templates.To prepare RNase-free DNA templates, approximately 20 µg of plasmid construct were digested, containing cDNA for probe, with the desired restriction endonuclease (Fig. 1) using 100 µL reaction volume.This amount was enough to enable labeling at least 10 times.Linearized templates were extracted with TRIS-HCl saturated phenol, pH = 8 (RNase free) and chloroform isoamyl alcohol (49:1), at a ratio of 1:1, twice, and once with chloroform only.The aqueous phase was precipitated with 3 M sodium acetate, pH = 5.5 (fi nal molarity of 0.3 M sodium acetate) and 3 volumes of 100% ethanol overnight at -20 ˚C.The DNA pellet obtained by centrifugation was washed with ice cold 70% ethanol (diluted with RNase free water from 100% ethanol).The air dried pellet was dissolved in 20 µL 10 mM TRIS (pH = 7.5) (RNase free) and stored at -20 ˚C until later use.The measured concentrations of templates for probes were between 0.5 and 1.5 µg/µL.

FIGURE 1.
This fi gure is an example how to generate antisense and sense RNA probes transcribing cDNA template from a plasmid construct.The portion of the plasmid construct colored light grey (S-labeled is sense strand, sequence same as mRNA) and black (AS-labeled is antisense strand, sequence complementary to mRNA) represents cDNA probe insert cloned in polyclonal site of plasmid (dark gray).Bluescript plasmid has T3 and T7 polymerases (dots), which utilization can transcribe sense or antisense RNA probes using cDNA insert as probe template.To localize mRNA in tissue, we would generate an antisense RNA probe which is complementary to mRNA sequence (sense).Sense probes can be used for control.In order to generate an antisense RNA probe, the plasmid should be cut with an appropriate restriction enzyme close to the 5' sense strand (light grey) of cDNA insert prior to transcription.Cutting the plasmid will limit cDNA insert transcription only, excluding vector sequences.The new antisense RNA probe can be synthesized from the 3' end of the sense cDNA strand (light grey) as template using corresponding polymerase (T3, in this example).To generate a sense RNA probe, plasmid DNA should be cut with another appropriate restriction enzyme close to 3' end of sense cDNA strand (light grey) and can be synthesized from 3' end of the antisense cDNA strand (black) as template, using corresponding polymerase (T7, in this example).
Cut here to make Antisense; use T3 promoter Cut here to make Sense; use T7 promoter
Using the procedure described above, most of our probes were of the same concentration as the control probe after labeling.The control probe was 0.1 µg/mL.We made aliquots of 10 µL (1 µg) in each tube for all probes.The aliquoted probes were stored at -70 ˚C for up to one year for future experiments.For the experiment, we added 1 mL of hybridization solution on each aliquot.

Otkrivanje signala in situ hibridizacije
Metoda 1: otkrivanje alkalnom fosfatazom Nakon ispiranja rezovi su se kroz 1 sat na sobnoj temperaturi namakali u smjesi za blokiranje: 1% reagens za blokiranje (Roche), svježe razrijeđen u puferu maleične kiseline (100 mM maleične kiseline, 150 mM NaCl, kruti NaOH hand until all water droplets were removed between the section and tissue.Slides were left in a rack perpendicular to surface of heater set at 55 ºC overnight.The sections were stored in boxes at +4 ºC with desiccant and will retain an in situ hybridization and immunocytochemistry signal for years.For better attachment prior to in situ hybridization, the night before the slide sections are laid horizontally on a headed plate set at 55 ºC until the next morning (for immunocytochemistry this is omitted).Sections treated with the extra heated incubation preserve mRNAs and the sections remained better attached on slides throughout the in situ hybridization procedure.

Methods 1, 2 and 3: non-radioactive in situ hybridization; procedures using digoxigenin labeled probes
The non-radioactive in situ hybridization was performed using procedures developed in our laboratory plus methods from others using digoxigenin probes as described (7,20).Prior to hybridization, sections were de-paraffi nized with xylene and 100% ethanol following re-hydration with 50%, 75% and 95% Ethanol and PBS (diluted with RNase free water).After treatment with proteinase K solution (Proteinase K, 5 mg/mL in 50 mM TRIS, 5 mM EDTA pH = 7.6) for 10 minutes at 37 °C, sections were refi xed in 4% formaldehyde in PBS (0.2 M phosphate buffer, 0.3 M NaCl), acetylated (100 mM triethanolamine, 0.25% acetic anhydride) and pre-hybridized in 2XSSC.Hybridization was performed at 55 °C overnight, in the hybridization solution containing 50% formamide, 20 mM TRIS-HCl (pH = 8.0), 1 mM EDTA, 0.3 M NaCl, 10% Dextran sulfate, 1X Denhardt solution, 100 µg/mL denatured sheared Salmon Sperm-DNA, 500 µg/mL tRNA) and 1 µg/mL DMP1 UTP-digoxigenin labeled RNA probe.After hybridization, the parafi lm cover slips were removed in 5XSSC at room temperature and sections were washed in 50% formamide, 5XSSC and 1% SDS solution at 55 °C two times for 30 minutes.Next, the sections were incubated with RNase (40 mg/mL RNase A 1 and 10 U/mL RNase T 1 ) in RNase buff er solution (0.3 M NaCl, 10 mM TRIS, 5 mM EDTA) at 37 °C for one hour, followed by incubation in the same buff er without RNase for 30 minutes.Consecutive washes at 55 °C were done twice with 50% formamide and 2XSSC for 30 minutes, and three times for 5 minutes each at room temperature in TBST (100 mM TRIS pH = 7.5, 150 mM NaCl, Tween-20 0.1%).At this point, slides were ready for detection of hybridization signal with three different methods:

Method 2: Tyramide signal amplifi cation and alkaline phosphatase detection
In situ hybridization was done as for method 1, only quenching of endogenous peroxidase step (3% H 2 O 2 in PBS for 60 minutes) was added before proteinase K digestion followed by three 3x5 minutes washes with PBS.This step was necessary to minimize background, since activation and covalent binding of DNP Amplifi cation reagent is catalyzed by horseradish peroxidase (HRP).
For this detection we used in situ TSA plusDNP-AP system from PerkinElmer.After post hybridization washing, the sections were soaked for 30 minutes at room temperature in TNB blocking reagent: 0.5% blocking reagent-PerkinElmer (or provided in kit), fresh made or stored at -20 °C in TNT buff er (100 mM TRIS pH = 7.5; 150 mM NaCl; 0.05% Tween-20 Method 3: Tyramide signal amplifi cation and cyanine 5 fl uorescence detection In situ hybridization was done the same as for method 1, only quenching of endogenous peroxidase step was added before proteinase K digestion followed by three 3 minute washes with PBS, as described in method 2. For this detection, we used in situ TSA PLUS Fluorescence system from PerkinElmer.After post hybridization washing, the same as in method 2, the sections were blocked for 30 minutes at room temperature in TNB blocking reagent.Anti Digoxigenin-POD antibody was diluted 1:50 in blocking reagent, and incubated overnight at 4 °C following 3 X 5 minute washes in TNT.Incubation for 3-10 minutes in Tyramide Signal Amplifi cation reagent (supplied in kit) was followed by 3 x 5 minute washes in TNT.After these washes, the sections were permounted in Vector shield fl uorescent mounting media and viewed with a confocal microscope (Excitation 648nm, Emission 570 nm).

Method 4: Radioactive in situ hybridization and signal detection
The radioactive in situ hybridization was extensively performed using procedures developed in our laboratory and described by Gluhak-Heinrich et al. (23).Hybridization was performed with 32 P rUTP labeled collagen 1a1 (Col1) and Osteocalcin (OC) RNA probes.

Results
The aim of this study was to develop and determine the power of our non-radioactive in situ techniques in comparison to our well established and widely used radioactive method (23) using paraffi n sections.As a novel approach to prove usefulness of our newly developed methods in a variety of applications, especially paraffi n sections, we also showed the possibility of quantifi cation of non-radioactive hybridization signals.To test this, we used de-mineralized tissue sections with bone, tooth and soft tissues.First we developed and optimized three different procedures of non-radioactive in situ hybridization methods.Second, we analyzed the expression of Col1 and OC mRNA in sections of mouse alveolar bone adjacent to fi rst molars using our routine radioactive in situ analysis (Figure 3).Third, we determined the expression of Col1, OC and Osteopontin (OP) mRNA using a non-radioactive approach (Figure 4).We used exactly the same templates of our cDNAs to transcribe 32 P or digoxigenin labeled RNA probes.The Col1 and OC mRNA expression in alveolar bone and tooth were compared and quantifi ed.To prove that our non-radioactive in situ hybridization method 2 is as sensitive as our radioactive method in soft and bone tissue, we tested three diff erent non-radioactive methods and radioactive method with three different probes: BMP2 (very low mRNA copy number), Col1 (very high mRNA copy number) and OP (medium copy number and specifi c for osteocytes).In developing our in situ procedures, we fi rst tested different parameters for optimal performance.The results of this testing are described below, but in situ pictures of these numerous testing are not shown because a large space would be needed.We started our testing with optimum preparation of bone tissue.This is crucial since in situ hybridization always begins with fi xation of the tissue or cells, which should preserve mRNA within the cells and morphology of tissue.The paraformaldehyde fi xation with EDTA treatment should be RNase free from the beginning.We also found that 8-micron thick sections were optimal for our methods: thicker sections can be used and thickness may give se rabiti deblji rezovi i debljina bi nam mogla dati više hibridizacijskog signala, ali će se rezolucija smanjiti ako se snimke rade uz veliko uvećanje i možda ne ostanu prionuti uza stakalca kroz čitavo vrijeme hibridizacijskog postupka.Najbolji rezultati za linearizaciju DNA probi restrikcijskim endonukleazama dobiveni su enzimskim cijepanjem kojim su proizvedeni 5' ljepljivi ili tupi završetci.Rezultati su također bili bolji ako su probe prije obilježavanja digoksigeninom i prije alikvotiranja radi čuvanja podvrgnute vorteksiranju i zagrijane uz 10 mM TRIS, pH = 7,6 bez RNAza na 65 °C kroz 20 minuta i/ili uz dodatak SDS u 0,1% kako bi se pripomogla topljivost probi.Dobili smo optimalni signal i nisku pozadinu s 5 µL probe na 1 cm 2 stakalca za hibridizaciju, sa svim digoksigeninom obilježenim probama u koncentraciji od 1 µg na 1 mL hibridizacijske otopine.Uporaba veće količine otopine s više probe na mm 2 na stakalcu ili više koncentrirane probe dala je jaču pozadinu.Utvrdili smo kako se koncentracija probe može točno procijeniti gel elektroforezom i bojanjem etidijevim bromidom uspoređujući alikvote RNA DIG obilježene probe sa standardnom RNA poznate koncentracije (jedna takva se isporučuje sa setom za obilježavanje DIG) ili usporedbom DOT blota serije razrjeđenja obilježene RNA probe i obilježenih RNA standarda (Slika 2.).Za radioaktivnu hibridizaciju procijenili smo količinu probe upotrebljenu u gel elektroforezi i kod bojanja etidijevim bromidom s standardom RNA poznate koncentracije.Koncentracija probi koju smo mi rabili kretala se od 100 do 500 ng/mL hibridizacijske otopine, a na stakalca smo  Purple color is counterstained using hematoxylin.The signal above the background is the highest in osteoblasts for both Collagen 1a1 and Osteocalcin probes.Note that the hybridization signal is more abundant for collagen than for osteocalcin.PD, periodontium; OB, osteoblasts; OC, osteocytes.us more hybridization signal, but resolution will decrease if pictures are taken on high magnifi cation and they may not stay adhered to the slides throughout the hybridization procedure.Best results for linearizing DNA probes with restriction endonucleases were obtained with enzymes cleaving producing 5' overhangs or blunt ends.The results were also better if probes, after digoxigenin labeling and before aliquoted for storage, were vortexed and heated in RNase free 10 mM TRIS pH = 7.6 at 65 ˚C for 20 minutes and/or the addition of SDS to 0.1% to aid solubility of probes.We obtained optimal signal and low background with 5 µl of probe per 1 cm 2 of slide used for hybridization, with all digoxigenin labeled probes at a concentration of 1 µg in one mL of hybridization solution.Use of more solution with more probe per mm 2 on the slide or more concentrated probe gave us higher background.We found that the accurate probe concentration estimation can be obtained by gel electrophoresis and ethidium bromide staining comparing aliquots of the DIG labeled probe RNA with a standard RNA of known concentration (one is supplied with the kit for DIG labeling) or by comparing dot blots of a dilution series of labeled RNA probe and labeled RNA standards (Fig. 2).For radioactive hybridization, we estimated the amount of probe used by gel electrophoresis and ethidium bromide staining with an RNA standard of known concentration.The concentration of probes we used was 100-500 ng/mL of hybridization solution and on slides we added 80 µl (10 µl per 1 cm 2 ).Regarding probe length, our experiments showed that probes of up to 1 kb were not a dodali 80 µL (10 µL na 1 cm 2 ).Glede duljine probe, naši pokusi su pokazali kako sa probama do 1 kb nema problema kod prodiranja u tkivo kod neradioaktivnog in situ postupka, ali su za radioaktivni in situ postupak sve probe hidrolizirane do prosječne duljine od 150-250 bp kako bi se dobio optimalan signal.Pri ispitivanju tkivne propustljivosti naša neradioaktivna metoda pokazala je dobre rezultate s 5 µg/mL proteinaze K. Ova je koncentracija prethodno ispitana za našu dobro utvrđenu radioaktivnu metodu (23).Nakon pokusa s različitim koncentracijama proteinaze K najbolje rezultate smo polučili primjenom iste koncentracije kakvu smo rabili za radioaktvni in situ postupak, jer su koncentracije veće od 10 µg/mL proteinaze K uništavale morfologiju i povećavale pozadinu.U našim prijašnjim i usporednim studijama radeći s radioaktivnom in situ hibridizacijom analizirali smo digestiju s RNazom tijekom faza ispiranja i njezin učinak na pozadinu.Našli smo da je optimalna koncentracija RNaze 40 mg/mL.Ova nam je količina dala najnižu pozadinu, a veće količine su imale manji utjecaj na pozadinu.Velika razlika u smanjenju pozadine kod in situ hibridizacije nastupila je dodavanjem RNaze T 1 u količini do 10 U/mL.Ovo ispitivanje je obavljeno uz radioaktivni in situ postupak te je preuzeto za neradioaktivni postupak, pokazujući vrlo čistu, nisku pozadinu.Ispitivali smo različita razrjeđenja anti-digoksigeninskih antitijela.Najbolje rezultate za visoko zastupljene gene dobili smo s razrjeđenjem 1:500, no za probe koje smo rabili za otkrivanje malobrojnih mRNA ciljeva optimalno razrjeđenje je bilo 1:50.Temeljito smo ispitivali duljinu vremena za razvijanje obojene otopine enzimatskom reakcijom alkalne fosfataze.Utvrdili smo kako je vrijeme razvijanja za probu Col1 jedan sat, probama OC i OP trebalo je 6 sati, dok je za probu BMP2 kao ciljnu mRNA, koja je prisutna u vrlo niskoj količini, trebalo prekonoćno razvijanje pomoću 1. metode.Ovi su uvjeti vrijedili za razrjeđenje anti-DIG antitijela 1:50.Što su antitijela više razrijeđena, to je više vremena potrebno za razvijanje.Pomoću 2. metode sve su se probe razvile u vremenu od 10 minuta do 1 sata, jer je signal bio pojačan (ovi rezultati nisu prikazani).Naši su rezultati pokazali kako razvijanje duže od 20 sati obično dovodi do jače pozadine nego što je prihvatljivo.

Kvantifi kacija rezultata in situ hibridizacije dobivenih pomoću Col1 i OC
Kako je prikazano na slikama 4. i 5., vidljivost signala neradioaktivne hibridizacije može se otkriti na različitim razinama.Najviši intenzitet hibridizacijskog signala otkriven je u odontoblastima (OD) pomoću probe Col1 (slika 5A.).Najniži signal je otkriven u osteoblastima (OB) pomoću probe Col1 (slika 5B.i 6B.).Nakon kvantifi kacije je hibridizacijski signal najvišeg intenziteta izmjeren pomoću problem for penetration in the tissue with non-radioactive in situ, but for radioactive in situ all probes were hydrolyzed to average length of 150-250 bp to obtain optimal signal.On testing for tissue permeability, our non-radioactive method showed good results with: 5 µg/mL of Proteinase K.This concentration was previously tested for our well established radioactive method (23).After experiments with diff erent concentrations of Proteinase K, we obtained best results using the same concentration as we used for radioactive in situ, since concentrations higher than 10 µg/mL of Proteinase K destroyed morphology and increased background.In our previous and parallel studies working with radioactive in situ hybridization, we analyzed digestion with RNase during washing steps and its eff ect on background.We found that 40 mg/mL RNase A to be optimum.This amount gave us lowest background and higher amounts had little infl uence on background.The great diff erence in decreasing background in situ hybridization was the addition of RNase T 1 up to 10 U/mL.This testing was done with radioactive in situ and it was adopted for non-radioactive showing very clear low background.We tested diff erent dilutions of anti-digoxigenin antibodies.The best results for high abundant genes were at 1:500 dilution but for probes used for detection of few mRNA targets the optimal dilution was 1:50.The length of time with developing color solution with alkaline phosphatase enzymatic reaction was tested extensively.We found developing time for Col1 probe to be one hour, OC and OP probes needed 6 hours, and for BMP2 probe as a very low abundant mRNA target we needed overnight development using method 1.These conditions are for dilution of anti-DIG-Antibodies 1:50.The more diluted antibodies, the longer the time of development needed.Using method 2, all probes developed between 10 minutes to 1 hour, since the signal was amplifi ed (results not shown).Our results showed that development longer than 20 hours usually resulted in higher background than acceptable.

Quantifi cation of in situ hybridization results obtained with Col1 and OC
As shown in Figures 4 and 5, visibility of non-radioactive hybridization signal is detectable at diff erent levels.The highest intensity of hybridization signal was detected in odontoblasts (OD) with Col1 probe (Fig. 5A).The lowest signal was detected in osteoblasts (OB) with Col1 probe (Figs.5B and 6B).After quantifi cation, the hybridization signal of highest intensity was measured with ImageJ as 95% (100% is black) and the lowest detected in this experiment was 23% (0% is white).The other two middle expression levels are: signal in odontoblasts (OD) obtained with OC (80%) and signal in osteoblasts (OB) from Col1  The signal is highest in osteoblasts for both Collagen 1a1 and Osteocalcin probes but there is also some signal in PDL and bone lining cells.Osteocytes are not expressing collagen or osteocalcin in these sections.Note also that the hybridization signal is more abundant for collagen than for osteocalcin.PD, periodontium; OB, osteoblasts; OC, osteocytes, BLC, bone lining cells.SLIKA 5. Različita jačina hibridizacijskog signala dobivenog neradioaktivnom digoksigeninom obilježenom in situ hibridizacijom Collage-n1a1 (A) i Osteocalcin (B) u reprezentativnim rezovima mišje maksilarne alveolarne kosti i dijela prvog kutnjaka.Plavo obojenje predstavlja pozitivan hibridizacijski signal, a zeleno predstavlja kontraboju.Signal je najviši u odontoblastima obilježenima Collagen1a1, nakon čega slijedi signal nižega intenziteta ali još uvijek visok u osteokalcinom obilježenim odontoblastima.Srednja jačina hibridizacijskog signala nalazi se u osteoblastima hibridiziranima pomoću Collagen1a1.Dobro vidljiv ali najniži intenzitet hibridizacijskog signala otkriven je u osteoblastima hibridiziranim osteokalcinskom probom.Stanice koštane obloge također su visoko obilježene.Osteociti ne izražavaju niti kolagen niti osteokalcin u ovim rezovima.Vidi se također kako je hibridizacijski signal obilniji za kolagen nego za osteokalcin i u odontoblastima i u osteoblastima.PD, periodont; OB, osteoblasti; OC, osteociti; BLC, stanice koštane obloge; OD, odontoblasti.The signal is highest in Collagen 1a1 labeled odontoblasts followed by lower intensity but still very high signal in osteocalcin labeled odontoblasts.The middle intensity of hybridization signal is in osteoblasts hybridized with Collagen 1a1.Well visible but the lowest hybridization signal intensity was detected in osteoblasts hybridized with the osteocalcin probe.The bone lining cells are also highly labeled.Osteocytes are not expressing collagen or osteocalcin in these sections.Note also that the hybridization signal is more abundant for collagen than for osteocalcin in both odontoblasts and osteoblasts.PD, periodontium; OB, osteoblasts; OC, osteocytes, BLC, bone lining cells, OD, odontoblasts.

Sensitivity of non-radioactive in situ hybridization methods
To test sensitivity of our non-radioactive in situ hybridization methods with radioactive in situ using tooth and alveolar bone tissue, we performed hybridization with three diff erent non-radioactive methods and the radioactive method using three diff erent probes: BMP2, Col1 and OPN.The BMP 2 probe was used as a representative for localizing one of the lowest abundant mRNAs in the bone.The best hybridization signal, with non-radioactive in situ, was obtained with method 2 using BMP2 digoxigenin labeled probe with Tyramide Signal Amplifi cation and alkaline phosphatase detection (Fig. 7B).The intensity of this BMP2 signal was comparable with method 4-radioactive in situ hybridization (Fig. 7D and 7E).Localization of BMP2 mRNA transcripts with Tyramide Signal Amplifi cation and Cyanine 5 fl uorescence detection, method 3, showed minimal BMP2 hybridization signal (Fig. 7C).Non-radioactive in situ using alkaline phosphatase detection without amplifi cation, method 1, did not show any hybridization signal for BMP2 (Fig. 7A).This method 1 was not sensitive enough to label rare BMP2 mRNA transcripts.The Col1 probe was used as a positive control since it represents one of the most abundant mRNAs in bone.The hybridization signal of Col1 was localized with all tested in situ methods with great intensity (Fig. 7F to 7O).OP probe was used as a marker for osteocytes, bone cells embedded in the matrix.This gene is much less abundant in bone than Col1, but much more than BMP2.The hybridization signal for OP in osteocytes was also found with all the in situ methods tested (Fig 7F to 7O).

Discussion
We have developed a high sensitive in situ hybridization method at a level with radioactivity, with identifying steps which are critical for localization of low copy number mRNAs in de-mineralized paraffi n embedded tissue with almost nonexistent background and possibility of quantifi cation.Up to date, quantifi cation for non-radioactive in situ hybridization using image analyzer has been published, yet only by Japanese authors (24), while the one for radioactive in situ was published 22 years ago (25).To test the power of this technique, we chose probes that label bone cells in teeth and soft tissue.Such probes include Col1 known to be abundant, whereas OC and OP are found at a lower copy number, but BMP2 are found in a very low copy number compared to Col1 in the same tissue.To achieve high sensitivity required for low-abundance mRNAs, we optimized crucial steps of hybridization and detection to achieve the best possible signal-to-noise ratio, comparable to radioactive in situ hybridization.All of these experiments were conducted on similar alveolar  F-J) and OP-osteopontin (K-O) mRNA expression performed with four diff erent in situ hybridization methods: 1) non-radioactive in situ using digoxigenin labeled probes with alkaline phosphatase detection (A, F, K); 2) non-radioactive in situ using digoxigenin labeled probes with Tyramide Signal Amplifi cation and alkaline phosphatase detection (B, G, L); 3) non-radioactive in situ using digoxigenin labeled probes with Tyramide Signal Amplifi cation and Cyanine 5 fl uorescence (C, H and M); and 4) radioactive in situ using 32 P labeled probes and autoradiography detection (D, I and N, darkfi eld images; E, J and O, bright fi eld images).In non-radioactive in situ hybridization with alkaline phosphatase detection (methods 1 and 2), positive hybridization signal is represented with blue stain and red color is counterstain from nuclear red (A, B, F, G, K and L).In non-radioactive in situ hybridization using fl uorescence of Cyanine 5 as detection (method 4), positive signal is represented with bright red color (C, H and M).The white silver grains represent positive hybridization signal in autoradiographs of radioactive in situ hybridization (method 3) using dark fi eld images (D, I and N), and black silver grains using bright fi eld images (E, J and O).The best hybridization signal for BMP 2 probe was obtained with (method 2) non-radioactive in situ using digoxigenin labeled probes with Tyramide Signal Amplifi cation and alkaline phosphatase detection (B) and (method 4) radioactive in situ hybridization (D, E).Note advantage of non-radioactive localization of BMP2 mRNA at the level of cells (B), compared to the radioactive one (D, E).Localization of BMP2 mRNA transcripts with Tyramide Signal Amplifi cation and Cyanine 5 fl uorescence (method 3) showed minimal BMP2 hybridization signal (C).Non-radioactive in situ using alkaline phosphatase detection without amplifi cation (method 1) did not show any hybridization signal for BMP2 (A).Col1 probe hybridization signal was localized with all four in situ methods (F-O).The hybridization signal for OP in osteocytes was also found with all four in situ methods (F-O).Labeling description: d, dentin; PD, periodontium; b, bone; OC, osteocytes.
enhances the color signal.The second step is the use of TSA (Tyramide Signal Amplifi cation) that signifi cantly enhances chromogenic and fl uorescence signals.For in situ signal detection, TSA technology uses horse radish peroxidase (HRP), attached to anti digoxigenin antibodies bound to digoxigenin labeled probes localizing mRNA, to catalyze the deposition of dinitrophenyl (DNP) labeled amplifi cation reagent adjacent to the immobilized HRP enzyme.These DNP labels can then be indirectly detected using anti-DNP-Alkaline Phosphatase antibodies and fi nally AP enzyme color detection.Fluorescent method 3 also uses TSA technology.With this method, we obtained an in situ hybridization signal using the BMP2 probe, relatively low abundant message in bone.Signal intensity using method 3 was lower than using method 2. Reason: method 3 uses fl uorescent detection rather than AP.The fl uorescent method 3 also uses TSA reagent but directly labeled with fl uorophore simplifying the detection for immediate visualization via fl uorescence microscope.AP enzyme detection for method 2 is more sensitive than fl uorescence since after TSA reagent fi nal detection is indirect, using additional anti-DNP-Alkaline Phosphatase antibodies and PVA development AP color enhancement lasting from 30 minutes to overnight, both additionally increasing sensitivity.With method 1 we did not localize BMP2; this method uses sensitive color AP enzyme detection but does not use TSA amplifi cation which enhances in situ signal signifi cantly.Our estimation is that non-radioactive method 2 is at least 10 times more sensitive than non-radioactive method 1 and 3-5 times more sensitive than method 3. The advantage of non-radioactive localization of BMP2 mRNA with method 2 is signal at the level of the cells (Fig. 7B), compared to radioactive one (Fig. 7D and 7E).The reason for non-precise localization of radioactive signal at the cell level is spreading high 32 P energy to surrounding areas, eff ecting exposure of photograph emulsion used for hybridization signal detection.Result is a lot of scattered silver grains, at and around the source of radioactive hybridization signal, which enables detection at the single cell level.Our fi nal in situ hybridization technique testing used osteopontin (OP) probe as a relatively abundant mRNA marker for osteocytes, i.e. bone cells embedded in matrix.Using this probe, we localized signals in osteocytes with all non-radioactive methods (Fig. 7K, 7L and 7M).Obtained signal intensity was comparable to the radioactive method (Fig. 7N and 7O) and provided fi nal proof of usefulness of our non-radioactive in situ hybridization methods in a variety of cells and tissues.We feel that the use of these non-radioactive in situ hybridization analyses with our new quantifi cation protocol off ers greater potential for novel biological insights in many fi elds including research and clinical chemistry.

FIGURE 2 .
FIGURE 2. Quantifi cation of digoxigenin labeled probes using dot blot and control labeled RNA from Roche.Dots labeled as 0 represent undiluted probes.A, represents 1/10; B, 1/100; C, 1/1000 and D, 1/10000 dilution of probes.The concentration of control probe is 0.1 µg/mL.In this example our labeled probes (1-8) have a similar concentration as the control probe (CTR).

FIGURE 3 .
FIGURE 3. Radioactive in situ hybridization of Collagen 1a1 (A) and Osteocalcin (B) in representative sections of mouse maxilla.Positive hybridization signal is represented with silver grains, black dots over tissue.Purple color is counterstained using hematoxylin.The signal above the background is the highest in osteoblasts for both Collagen 1a1 and Osteocalcin probes.Note that the hybridization signal is more abundant for collagen than for osteocalcin.PD, periodontium; OB, osteoblasts; OC, osteocytes.

FIGURE 4 .
FIGURE 4. Non-radioactive in situ hybridization performed with digoxigenin labeled probes, Collagen 1a1 (A) and Osteocalcin (B) in representative sections of mouse maxilla.Positive hybridization signal is represented with blue stain.Green color represents methyl green staining as counterstain.The signal is highest in osteoblasts for both Collagen 1a1 and Osteocalcin probes but there is also some signal in PDL and bone lining cells.Osteocytes are not expressing collagen or osteocalcin in these sections.Note also that the hybridization signal is more abundant for collagen than for osteocalcin.PD, periodontium; OB, osteoblasts; OC, osteocytes, BLC, bone lining cells.

FIGURE 5 .
FIGURE 5. Diff erent intensity of hybridization signal obtained with non-radioactive digoxigenin labeled in situ hybridization of Collagen 1a1 (A) and Osteocalcin (B) in representative sections of mouse maxillary alveolar bone and part of the fi rst molar.The blue stain represents positive hybridization signal and green represents counterstain.The signal is highest in Collagen 1a1 labeled odontoblasts followed by lower intensity but still very high signal in osteocalcin labeled odontoblasts.The middle intensity of hybridization signal is in osteoblasts hybridized with Collagen 1a1.Well visible but the lowest hybridization signal intensity was detected in osteoblasts hybridized with the osteocalcin probe.The bone lining cells are also highly labeled.Osteocytes are not expressing collagen or osteocalcin in these sections.Note also that the hybridization signal is more abundant for collagen than for osteocalcin in both odontoblasts and osteoblasts.PD, periodontium; OB, osteoblasts; OC, osteocytes, BLC, bone lining cells, OD, odontoblasts.

FIGURE 6 .
FIGURE 6. Quantifi cation using ImageJ of diff erent levels of hybridization signal intensities was obtained with non-radioactive in situ hybridization.The relative expression of Collagen 1a1 (A) and Osteocalcin (B) mRNA from Figure 5 was quantifi ed in odontoblasts and osteoblasts.The black bar in the graph represents almost the highest possible (95%) hybridization signal which was obtained with Collagen 1a1 mRNA in odontoblasts and marked with the dark blue arrow as in Figure 5.The osteocalcin mRNA expressed in odontoblasts labeled with a dark gray bar in the graph and a red arrow in Figure 5 represents a still very high hybridization signal (80% of maximal signal), although lower than previously in odontoblasts labeled with Collagen 1a1.The next light gray bar in the graph showing 45% of maximum hybridization signal is represented with Collagen 1a1 expression in osteoblasts and labeled with a light blue arrow in Figure 5.The still highly visible (23% of maximum signal) hybridization signal is shown with a white bar graph representing osteoblasts hybridized with a osteocalcin probe and marked with light a red arrow in Figure 5.Note the good correlation of relative hybridization signal, blue color in Figure 5 with the possible range of Im-ageJ quantifi cation in Figure 6.

FIGURE 7 .
FIGURE 7.The representative sections of mouse maxilla showing BMP2-Bone Morphogenetic Protein 2 (A-E), Col1-Collagen 1a1 (F-J) and OP-osteopontin (K-O) mRNA expression performed with four diff erent in situ hybridization methods: 1) non-radioactive in situ using digoxigenin labeled probes with alkaline phosphatase detection (A, F, K); 2) non-radioactive in situ using digoxigenin labeled probes with Tyramide Signal Amplifi cation and alkaline phosphatase detection (B, G, L); 3) non-radioactive in situ using digoxigenin labeled probes with Tyramide Signal Amplifi cation and Cyanine 5 fl uorescence (C, H and M); and 4) radioactive in situ using 32 P labeled probes and autoradiography detection (D, I and N, darkfi eld images; E, J and O, bright fi eld images).In non-radioactive in situ hybridization with alkaline phosphatase detection (methods 1 and 2), positive hybridization signal is represented with blue stain and red color is counterstain from nuclear red (A, B, F, G, K and L).In non-radioactive in situ hybridization using fl uorescence of Cyanine 5 as detection (method 4), positive signal is represented with bright red color (C, H and M).The white silver grains represent positive hybridization signal in autoradiographs of radioactive in situ hybridization (method 3) using dark fi eld images (D, I and N), and black silver grains using bright fi eld images (E, J and O).The best hybridization signal for BMP 2 probe was obtained with (method 2) non-radioactive in situ using digoxigenin labeled probes with Tyramide Signal Amplifi cation and alkaline phosphatase detection (B) and (method 4) radioactive in situ hybridization (D, E).Note advantage of non-radioactive localization of BMP2 mRNA at the level of cells (B), compared to the radioactive one (D, E).Localization of BMP2 mRNA transcripts with Tyramide Signal Amplifi cation and Cyanine 5 fl uorescence (method 3) showed minimal BMP2 hybridization signal (C).Non-radioactive in situ using alkaline phosphatase detection without amplifi cation (method 1) did not show any hybridization signal for BMP2 (A).Col1 probe hybridization signal was localized with all four in situ methods (F-O).The hybridization signal for OP in osteocytes was also found with all four in situ methods (F-O).Labeling description: d, dentin; PD, periodontium; b, bone; OC, osteocytes.
ImageJ is a public domain, a Java based image processing program developed at the National Institutes of Health (NIH), USA.The ImageJ software is freely available for Microsoft Windows, Mac OS, Mac OS X, Linux, Sharp Zaurus PDA and can be downloaded if logged on: http://rsb.info.nih.gov/ij/download.html.This software has been designed with an open architecture that provides extensibility via Java 1.1 version or later versions, using plugins and recordable macros.Custom acquisition, analysis and processing plugins can be developed using ImageJ's built-in editor and a Java compiler.mentsunderAnalyze.We used the tool "free hand selections" surrounding areas for measurements.All measurements were read clicking on Measure under Analyze.All areas measured for a hybridization signal were corrected for background reading.To defi ne background, 5 areas vrijednost sivog.RezultatiCilj ove studije bio je razviti i odrediti snagu naših neradioaktivnih in situ tehnika u usporedbi s našom dobro utvrđenom i široko primjenjivanom radioaktivnom metodom (23) uz uporabu parafi nskih rezova.Kao nov pristup u dokazivanju korisnosti naših novo razvijenih metoda u različitim primjenama, poglavito s parafi nskim rezovima, pokazali smo i mogućnost kvantifi ciranja signala neradioaktivne hibridizacije.Za ispitivanje ove mogućnosti rabili smo demineralizirane tkivne rezove s koštanim, zubnim i mekim tkivom.Najprije smo razradili i optimirali tri različita postupka za metode neradioaktivne in situ hibridizacije.Zatim smo analizirali izraženost Col1 i OC mRNA u rezovima mišje alveolarne kosti blizu prvih kutnjaka pomoću naše rutinske radioaktivne in situ analize (slika 3.).