The In Situ PCR:Amplification and Detection in a Cellular Context

Introduction The diversity of pathogenesis presents a spectrum of challenges to the researcher, who is required to utilize a wide variety of tools and techniques, including those of a histological, immunological and molecular nature. Within the larger context of pathogenesis, the determination of the mechanism of viral latency and slowly progressing viral diseases is particularly provoking. Here, the molecules involved in the initiation and progression of disease may be present in vanishingly small quantities in a minor population of cells or tissues, and the dynamics of expression of functions within the cells will vary over time. In many of these slowly-evolving diseases, requiring months or years to manifest clinically, it has been shown that, with respect to viral genetic information, the majority of the infected or affected population is in a transcriptionally inactive state, and at a level of one genome [or gene] per host cell.

An example of one of the most difficult issues in viral pathogenesis today is that presented by the lentiviruses, members of the "complex" subgroup of the retroviruses, which includes HIV, the causative agent of AIDS in humans, and the prototype for the class, visna-maedi, which causes neurological and pulmonary disease in sheep. Upon infection with these agents, the provirus frequently integrates into the host genome and establishes a persistent infection, whereby the infected cells are characterized by, among other things, a transcriptionally-quiescent state with respect to the viral antigens, which allows the infected cells to escape host immune surveillance [the Trojan horse mechanism]. It is this state which is pathogenically and epidemiologically most interesting, and potentially the most lethal, since these cells provide a reservoir for the future release of active virus. It is these latently infected individual cells, then, which need to be discovered and enumerated in order to gain some insight into the essence and the extent of the infection, the progression to expression and, eminently, the control of this progression.

The techniques of nucleic acid hybridization and the polymerase chain reaction [PCR] have been used extensively to investigate these issues of pathogenesis. While powerful in their own right, these techniques are essentially population studies: nucleic acids are isolated from a population of cells which contains either a sufficient number of molecules to detect directly by standard hybridization techniques, or, when a subpopulation contains as little as a single copy of nucleic acid, that molecule amplified by the PCR, and detected after amplification. In situ hybridization applies and extrapolates the technology of nucleic acid hybridization to the single cell level, and, in combination with the artistry of cytochemistry and immunocytochemistry, permits the maintenance of morphology and the identification of cellular markers to be maintained and identified, allows the localization of sequences to specific cells within populations, such as tissues and blood samples. However, with in situ, the technology is limited primarily to the detection of non-genomic material [e.g., RNA], reiterated genes or multiple genomes, since the limits of detection in most situations are several copies of the target nucleic acid per cell. In part due to these copy number considerations, hybridizations for RNA are considerably more sensitive than for DNA detection. In addition, other factors which affect the sensitivity of the sensitivity of the technique toward RNA targets are the strandedness of the target molecule and the lack of a complementary sequence proximal to the target sequences, which kinetic theory would predict to be a preferential substrate for annealing. Other techniques, specifically the reverse transcriptase-catalyzed in situ transcription, have been used as well to detect RNAs which occur at relatively high copy number. The single-copy problem proffered by the slow disease agents referred to above, however, is beyond the realm of repeatable reality with conventional in situ techniques.

Since the viral nucleic acid within these infected cells is below the level of routine detection, the application of amplification technologies, of which the PCR is the most logical choice due to its exquisite sensitivity, is essential. Clearly, the solution-phase corollary of this problem, namely the amplification and detection of DNA purified from a single cell, either as an individual cell or within a population of dissimilar cells, has been resolved [see Chapter xy]. For the histological identification of that cell within a population, such as in a tissue section or within a blood sample, the problem becomes more complex. In addition to the standard reaction components, consisting of enzyme, target DNA, substrate and primers, which must be optimized in addition to varying cofactors [Mg, primer-template ratios, etc.], one must consider other variables and potential problems, all of which have direct consequences in the success of this technique. The importance of the preservation of morphology and antigens of cells and tissues limits the choice of fixation methods; these procedures are crucial to the detection of sequences generally in in situ hybridization, and are even more so when considering the PCR in situ . The modified cellular milieu, maintained as a consequence of the fixation procedure, dramatically affects the efficiency of amplification, as it introduces potential interference from protein contaminants and cross-linking of nucleic acids and proteins. This is compounded by the general reduction of signal that occurs in solid-phase systems. The diffusion of reaction components into the cells likely also proceeds at a lower rate than in solution. Finally, the amplified product DNA must be retained within the permeabilized cell for the detection by radiolabelled probes; this necessitates a relatively large size for the designed product DNA, which is contrary to the relatively low efficiency for the in situ amplification reaction. All of these considerations amplify the overall problem of detecting PCR-generated sequences within a morphologically-identifiable cell, in addition to amplifying the sequences.

Technical Aspects/Philosophy The techniques described in this chapter apply the PCR in the context of in situ identification, and have been developed using the ovine lentivirus, visna-maedi, as a model system. The details of in situ hybridization have been previously described in detail[], and are consistent with established standards. The amplification technique described here is applicable to cultured cells, both adherent and non-adherent, which have been released from the growth support and fixed in solution, as well as to cultured cells and tissues that have been fixed, embedded, sectioned, and attached to slides; differences in strategies for the two types of preparations will be discussed. The reaction conditions are detailed at the end of the discussion.

Primer Design In general, the PCR requires the critical design of a primer set. For in situ PCR, the special considerations alluded to in the previous section make this step even more crucial. The relatively small size of the amplification product frequently chosen for the PCR [200-500 bp] is within the size class of molecules frequently used as probes for in situ hybridization; that is, the diffusion characteristics of this size class of nucleic acid is such that it passes with facility through permeabilized cell membrane, and this diffusion is very likely enhanced in the PCR by the elevated temperatures developed during the course of the reaction. We have demonstrated that the retention of this size material within a cell is minimal, and the size of choice is counter-intuitive to the principles of in situ hybridizations. In addition, the very nature of the fixed cellular material appears to substantially hamper the efficiency of the PCR, making the choice of a pair of primers sufficiently spaced so that they will successfully give rise to a product DNA which will be retained within the cell [900-1300 bp] virtually impossible. To resolve these dilemma, we designed a set of primer pairs whose individual reaction products were 1] small enough [app 350 bases] for efficient amplification within the constraints of fixed cells, and 2] whose product DNA overlaps in a complementary nature to other primers and elongated product DNA. Thus, even the cellularly-impeded reactions should be able to give rise to a product or product network of such a size as that described above for cellular retention. In the visna-maedi model, nine primers, (five [+]- and four [-]-sense), each 20 nucleotides in length, are used to represent approximately 1300 bp of contiguous sequence information from the relatively conserved 5'-terminus of the viral genome. In addition, no two [+] and [-] primers share identical or complimentary sequence information, and in fact are slightly offset [approximately 20 bases], so that the primers must elongate in order to pair with adjacent primer sets in successive cycles. As with the design of sequencing primers, these primers are chosen from highly conserved sequences [when that information is available], and tested for similarity to both strands of sequence information represented by the complete GenBank Nucelotide Sequence Database, using both pattern-matching and similarity programs; any primer having significant similarity in the 3'-terminal 6-10 nucleotides to any gene that may potentially be present [i.e., any ovine gene or any potentially homologous structure in other species] in the sample is rejected. Inverted repeats, and the consequent potential to form hairpins, are also tested for, using stem and loop designations which would have an extremely low probability of existing under the conditions utilized in the amplification reaction [stems, four base pairs; loops, two nucleotides]; again, failed sequences are generally rejected. In addition, the primers are so designed as to incorporate, to the extent feasible, the criteria of Sninsky and Whoever.

In the overall reaction scheme, any [+] / [-] primer pair must be capable of producing an appropriate product in a standard liquid-phase PCR with purified template. While this test is not a general part of the protocol, it should be considered if problems are encountered in the amplification in situ . These primer sets, however, cannot be optimized individually, since they must exist in the context of the multiple-primer reaction; primer pairs which cannot be used in the standardized conditions must be re-designed

Reaction Conditions During the initial cycles of the reaction, the stoichiometry of the nucleic acid components are such that the reaction must be primer-driven. However, as the reaction proceeds, it is likely that the PCR product, spanning some fraction less than the complete target sequence, and which has not diffused from the cell, begins to function as a target for further amplification, as well as a primer for further elongation. Thus, it is this accumulation of sequences elongated in multiple reactions which is retained by the fixed cell and acts as the target for the radiolabeled probe in the detection phase.

Due to the complex physical architecture within the fixed cell, which may impede both the efficiency of the polymerase, as well as diffusion and annealing, elongation times are substantially longer than those for liquid conditions. In our hands, the optimum signal is produced with a 15 minute elongation.

"Mechanics" of the Reaction In situ PCR is carried out on Denhardt-treated microscope slides to which cultured cells or tissue sections have been attached. Five microliters of the PCR mixture is applied directly to cytospun or spotted cells which have been fixed with paraformaldehyde; for tissue sections, sufficient reaction mixture is applied to wet the section [generally 20-40 microliters]. A coverslip is carefully applied, and, in a fashion similar to that used for Southern or Northern hybridizations, the slides are placed in thin, heat-sealable plastic bags and mineral oil is carefully added to the bag in a quantity sufficient to surround the slides; the bag is sealed, and placed in a thermal-cycling oven. The mineral oil serves to stabilize heat transfer and prevent the evaporation of the aqueous reaction solution. A controlling thermistor is attached to a microscope slide in another plastic bag with a similar quantity of oil to provide a comparable environment for the measurement of the sample temperature during the reaction. After 30 cycles, the bag is opened and the slides stripped of oil by dipping in chloroform. Taking care not to allow the slides to dry, the coverslip is carefully lifted, and an additional aliquot of fresh PCR mixture is introduced onto the cells or section; the slides are re-sealed into the oil-filled bag, and an additional 30 cycles performed.

As with any PCR, the empirical optimization of primer concentration [and consequently primer-template ratios], magnesium concentration, annealing and elongation temperatures will result in a higher efficiency, which, in the case of the in situ PCR, ultimately results in a better signal to noise ratio.

Methodology Two visna-maedi model systems are described here: 1] a tissue culture system, in which infection is efficient and productive, and viral DNA is abundant; and 2] an animal model system, in which productively-infected cells are rare and viral DNA in most infected cells is at levels consistent with other natural-host retroviral infections. Both systems have been described previously and are well characterized.

A. Cell and Tissues. 1. Cultured Cells: Infection and Preparation. Sheep choroid plexus [SCP] cells, a permissive host, were infected at a multiplicity of 3 plaque forming units [PFU] of visna virus (equivalent to about 30 copies of viral RNA) per cell and collected three hours post-infection [p.i.]. Previous work has shown that one to two copies of viral DNA per cell have been reverse transcribed from incoming viral RNA genomes at this time. Additional time points were taken at 20-24 hours p.i. [20-40 copies of DNA] and at 48-66 hours p.i. [100 copies of DNA].

Infected cells were harvested by trypsinization, pelleted at low speed [600xg, 5 minutes], washed in calcium and magnesium-free phosphate-buffered saline [PBS-CMF], and pelleted again at low speed. The cells were fixed by resuspension in freshly prepared 4% [w/v] paraformaldehyde in PBS-CMF for 20 minutes at room temperature. The cells were then pelleted [1600xg, 5 min] from the fixation solution, re-suspended in PBS-CMF, and diluted 25-fold with 70% [v/v] ethanol for storage.

2. Tissues: Pulmonary infection and preparation. 1x109 PFU of visna virus, concentrated from tissue culture medium by ultracentrifugation, were introduced via bronchoscopy into the lower segment of the right upper lobe of a 5 month old male Dorset Hampshire sheep, as described previously (). On the eleventh day post infection the experiment was terminated by euthanizing the animal by intravenous injection of 10 ml of Beuthanasia-D (Schering Corp.). After sacrifice, the lungs were removed en bloc and the consolidated segment of the right upper lobe was re-inflated by instilling 10% buffered formalin. The lungs were fixed for 72 hours and then transferred to 70% ethanol. Tissue segments from infected and uninfected lobes were embedded in paraffin for thin sectioning, histopathological examination, and in situ hybridization (7, 8). Eight-micron sections were attached to Denhardt-coated slides, deparaffinized and then stained with hematoxylin and eosin (H&E).

B. In Situ Amplification 1. Cultured Cells in Suspension. Immediately prior to amplification, and for each reaction, a 2x106 cell aliquot was pelleted from the storage solution at low speed, the ethanol removed, and the cells hydrated in PBS-CMF for at least 20 minutes at ambient temperature. The cells were pelleted at low speed, the initial hydration media removed and replaced with 100-400 ul of PBS-CMF per aliquot. Each aliquot was placed in a 0.5 ml microcentrifuge tube, spun at 1600xg for five minutes and the supernate removed. The cells were resuspended in 100 ul of PCR reaction mixture [200 uM dNTP's, 0.1 uM primers, 10 mM Tris-HCl, pH 8.3, 50 mM KCl, 1.5 mM MgCl2 and 0.01% gelatin]. The PCR was performed in a Perkin-Elmer Cetus Thermal Cycler, with an initial denaturation at 94°C for 10 minutes, after which time 2.5 units Taq polymerase [Perkin-Elmer Cetus] was added, as was a 50-100 ul mineral oil overlay. 25 cycles of amplification were performed [denaturation, 2 minutes, 94°C; annealing, two minutes, 42 °C; extension, 15 minutes, 72°C]. After the addition of fresh polymerase, an additional 25 cycles were performed in identical fashion.

2. Tissues on Slides. Deparaffinized tissue sections were overlaid with standard PCR cocktail (20-40µl, depending on the size of the section) containing 1 µM of the multiple primer set. The reaction mixture also contained 10mM Tris-HCl, pH 8.3, 50mM KCl, 1.5 mM MgCl2 , 0.01% gelatin, 200mM deoxynucleoside triphosphates and 5% (v/v) taq polymerase at 5 U/ml. After covering the sections with a coverslip, the slides were placed in heat sealable plastic pouches (4x6" x 2 mils). 5-6 ml of mineral oil was layered over the slides, the bags were compressed to remove most of the air and the tops were heat sealed. A slide to which a thermal sensor had been attached by the manufacturer was similarly sealed in an oil-filled bag and the sensor and slides with tissue sections were placed in a rack in a Bios Oven Thermal Cycler. After 25 cycles (92°C denaturation for two minutes; 42°C annealing for two minutes; 72°C extension for 15 minutes), the slides were removed from the bags and washed twice [five minutes each] in CHCl3 (twice, 5 minutes each) to remove residual oil. Each slide was taken from the CHCl3 and, as soon as the CHCl3 had volatilized, the slips were removed with the tip of a scalpel and forceps. Fresh PCR mixture, taq polymerase and coverslip were added prior to placing the slides back in the oil-filled plastic bags for an additional 25 cycles. Oil was again removed in CHCl3 and the sections were washed twice in phosphate-buffered saline (5 minutes each) and dehydrated in graded alcohols.

C. Detection of Amplified DNA by in situ Hybridization 1. Solution amplification. After amplification, SCP cells were centrifuged at 16,500 x g for 5 minutes in a Sorvall HB-4 rotor, and the oil and reaction mix was removed by aspiration. The pellets were resuspended in 100 µl of PBS-CMF, transferred to 1.5 ml microfuge tubes, and centrifuged at 16,500 x g for 5 minutes. After aspirating the PBS-CMF, the pellets, containing approximately 2 x 106 cells, were resuspended in 0.6 ml of PBS-CMF and the cells were then deposited by cytocentrifugation (450 rpm x 5 minutes, Shandon cytocentrifuge) onto two glass slides. Viral RNA was digested with ribonucleases A (100 µg/ml) and T1 (10 µg/ml), and the cells were postfixed for two hours in 4% paraformaldehyde to improve retention of DNA during subsequent denaturation, hybridization and washing (4). After denaturation in 95% formamide, 0.1xSSC for 15 minutes at 65°C, the slides were cooled to 4°C in 0.1xSSC, dipped in water, and dehydrated in graded alcohols. Virus-specific probes (106 dpm) labelled with 125I by nick translation (see next section) in 5 µl of a hybridization mixture (50% [v/v] formamide, 0.6 M NaCl, 20 mM Hepes buffer, pH 7.2, 1 mM EDTA, 1X Denhardt's media [DM], 500 µg/ml yeast RNA, 5% [v/v] polyethylene glycol, 100 µM aurintricarboxylic acid) were annealed at 37°C for 12 hours. Unreacted probe was subsequently removed by washing. The slides were coated with nuclear track emulsion (Kodak NTB-2), developed after exposures at 4°C for appropriate intervals and stained with hematoxylin and eosin following established protocols ().

2. Tissue Sections. After pretreatments with ribonucleases, postfixation in paraformaldehyde, and denaturation in formamide as described, sections were hybridized for 12-16 hours to a gel-purified fragment of cloned visna virus DNA corresponding to the amplified segment, labeled with [125 I] dCTP by nick translation (17) to specific activities of 5-10 x10 8 dpm/µg (1-3 x10 6 dpm per section); or to an HIV-specific probe labeled similarly to comparable specific activities (10). After washing for 12-16 hours to remove unhybridized probe, the sections were dehydrated in graded alcohols containing 0.3 mM NH4 acetate, coated with nuclear track emulsion (Kodak NTB-2), exposed at 4°C, developed and stained with H & E. Detection and quantitation of visna virus RNA by in situ hybridization in tissue sections followed previous descriptions (7, 8).

125 I-labelled Probes. Virus and region-specific probes were prepared by nick translation of cloned viral DNA using 125 I dCTP as precursor; or by using the PCR, specific primer sets, at 1 µM, and 125I-dCTP as a precursor, as described by Schowalter and Sommer (8). The PCR probes were prepared as follows: To 19 µl of a reaction mixture described above but lacking dCTP was added 1 µl containing 2 ng of bacteriophage λDNA as carrier and 1 ng of cloned visna DNA (9) (10 kb genome) in pBR 322 which had been linearized with ClaI. The reaction mix was used to redissolve 0.5 mCi of lyophilized 125I-dCTP (2200 Ci/mmol) to give a final concentration of about 10 µM. The DNA was denatured, enzyme was added and the PCR was carried out as follows: 94°C x 1 minute; annealing at 50°C x 2 minutes; extension at 72°C x 3 minutes. After 50 cycles, the reaction mix was removed. The products were separated by gel filtration chromatography on Sephadex G50 and concentrated by precipitation with ethanol and carrier RNA. The final probe products had specific activities of 3.7 x 109 dpm per µg and were reduced to about 150-300 bp in length as a consequence of radiolysis.

References References pnas   1. Montagnier, L. & Gallo, R. (1987) Nature 326, 435-436.   2. Haase, A.T. (1986) Nature 322, 130-136.   3. Fauci, A.S. (1988) Science 239, 617-622.   4. Haase, A.T. (1986) in In Situ Hybridization--Applications to Neurobiology , eds. Valentino, K., Roberts, J. & Barchas, J. (Symposium Monograph, Oxford University Press, Fairlawn, New Jersey), pp. 197-219.   5. Saiki, R.K., Scharf, S., Faloona, F., Mullis, K.B., Horn, G.T., Erlich, H.A. & Arnheim, N. (1985) Science 230, 1350-1354.   6. H.A. Erlich, D.H. Gelfand, & R.K. Saiki, Nature 331, 461-462 (1988).   7. Haase, A.T., Stowring, L., Harris, J.D., Traynor, B., Ventura, P., Peluso, R. & Brahic, M. (1982) Virology 119, 399-410.