WO1995020676A1 - In situ nucleic acid amplification - Google Patents

Abstract

A method of amplification of nucleic acid which comprises treating a liquid suspension of cells or other cell-like structures containing nucleic acid within a natural or artificial membrane or coat protein, under conditions which maintain cell morphology, with an amplification reaction mixture designed to amplify a target nucleic acid sequence therein and continuing the amplification reaction for at least two rounds of amplification of said target sequence in situ.

Description

IN SITU NUCLEIC ACID AMPLIFICATION This invention relates to methods for the amplification of nucleic acids. Among the various methods known for amplifying nucleic acids present in very small quantities in biological material one based on the polymerisation chain reaction (PCR) is of outstanding utility for many practical purposes. One or more sequences may be amplified simultaneously. The present invention is especially applicable to the improvement of the PCR method within cells in suspension and will for convenience be more particularly described hereinafter in relation to PCR. It will be clear however that the invention is also applicable to other methods of amplification of nucleic acids within cells in suspension including the ligase chain reaction (LCR), the primed in situ extension reaction (PRINS), the self-sustained sequence reaction (3SR) or the nucleic acid sequence based amplification (NASBA) and other amplification reactions.

In order to amplify nucleic acids by PCR it is customary practice to liberate the nucleic acid from cells and to carry out many cycles of the PCR reaction extracellularly using the reaction substrates and reagents, notably the oligonucleotide primers, DNA polymerase, and the nucleotide units from which new strands of the desired target sequence are synthesised. About 30 cycles of the reaction are referred to as one "round" of PCR. A well-known development of the basic PCR process is known as "nested" PCR, a process in which one round of PCR to amplify a first selected target sequence is followed by a second round in which a subsegment of the first sequence is amplified, this sub-segment being the target sequence of ultimate interest.

The extracellular method of PCR has certain disadvantages. It is difficult to avoid contamination from inhibitors and other materials, nucleic acids carried over from previous experiments or introduced from the external environment. These contaminants may either act as templates for non-specifically primed reactions initiating production of false products or may be themselves amplified products from previous reactions. In addition, reaction inhibitors within body fluids and cells are hard to separate from extracellular nucleic acids. Destruction of the cells also inevitably makes it impossible to discriminate between the various cell types present and so identify the cell type from which the target DNA originates. Furthermore, relatively large amounts of clinical material may be required for the extraction of the DNA.

Attention has therefore been directed to the possibility of carrying out PCR intracellularly (in situ PCR). To achieve in situ PCR it is necessary to render the cell membranes sufficiently permeable to allow entry of the PCR reagents whilst avoiding transgression of cell membranes by amplified products. However, the methods which have been previously used for this purpose have involved fixation, followed by pretreatment reagents which may compromise cell morphology. Thus after fixation of the cells the pretreatment reagents used to permeabilise the cells have included proteases and detergents which are harsh to the cells, reverse the effects of fixation, and encourage leakage of products. Consequentfy a further post- fixation treatment of the cells after reaction has generally been necessary. The net effect of these procedures has been to make it impossible to carry out more than one round of in situ PCR by the methods proposed.

In accordance with this invention, we have now found that successful in situ amplification can be accomplished by avoidance of the harsh fixation and pretreatment conditions referred to above. It has been been found that the conditions under which the amplification reagents are usually added to cells probably cause minor perturbations of the cell membrane which are sufficient to achieve the limited degree of permeability required for entry of the amplification reagents.

The present invention therefore makes possible a method of amplification of nucleic acid which comprises treating a liquid suspension of cells, or other cell-like structures containing nucleic acid within a natural or artificial membrane e.g. a bilipid layer as in liposomes, or within a coat protein e.g. as in viruses, under conditions which maintain morphology, with an amplification reaction mixture designed to amplify a target nucleic acid sequence therein and continuing the amplification reaction for at least two rounds of amplification of said target sequence in situ.

In performing the method of this invention the cells may therefore be treated direcdy after recovery from a sample of biological material without intervening pretreatment to permeabilise the cell membrane. Indeed, the method may also be performed without prior fixation of the cells. However we do not exclude the use of fixation reagents provided these are sufficiendy mild to avoid disruption of cell morphology. Thus precipitation fixatives may be used eg ethanol whereas cross- linking fixation reagents should be avoided since they cross-link nucleic acid histones and ceil membrane proteins. It will also be appreciated that no post- treatments are contemplated that would prevent entry of reagents that are necessary for further rounds of amplification by PCR eg nested PCR or by any other method. The effect of heat of reaction during the first round of amplification does not alter membrane permeability in a way that would require further fixation or digestion step prior to the second round. Thus the conditions under which the method of this invention is performed are generally mild conditions in which the integrity of the cells is substantially preserved. With this criterion in mind we recommend the use of isotonic solutions throughout. Under such conditions morphology can be preserved for 80 cycles or more.

The present invention is applicable to all types of prokaryotic, eukaryotic and plant cells and to liposomes. The method may also be applied to nucleic acids separated from the external environment by a surface.

As indicated above the invention is of particular benefit in relation to PCR technology where two or more consecutive rounds of amplification are desirable. This is especially so for nested PCR in which the first target DNA sequence is amplified in the first round and the sub-sequence of interest is amplified in the second round. It is also contemplated in accordance with the invention to make use of a combination of different amplification methods in the two or more rounds over which amplification is continued eg LCR followed by PCR. Another method of amplification known as PRINS (primed in situ extension reaction) may also be used with advantage in which a single primer is used in the first round of amplification in order to generate a more than usually long product which is least susceptible to leakage through the cell membrane. The use of PRINS in situ may permit reduction in the number of cycles for a successive in situ PCR reaction.

For the detection of the amplified product the usual labeling methods can be adopted eg using labelled probes and in situ hybridisation. For increased sensitivity we prefer to use labelled nucleotides (dUTP) and especially labelled primers.

The foregoing description of the present invention has been presented in the context of DNA amplification wherein the target DNA may exist as such in the cells under investigation. It will be appreciated however that the DNA to be amplified can be produced in situ from cellular RNA by the technique of intracellular reverse transcription (in situ RT). Alternatively, RNA sequences may be amplified intracellularly (e.g. in situ NASBA). Thus we present a new method for combined intracellular reverse transcription (in situ RT) and intracellular two round nested polymerase chain reaction (in situ single round PCR, or in situ nested PCR respectively) within cells in suspension. This method has been established using measles virus infection of the human vero cell line as a model. In the nested PCR strategy, a second pair of primers with specificity for sequences internal to an amplified product after a first PCR reaction, is used to perform a second series of amplification cycles. More than two rounds of nested reaction are also possible. Another amplification or replication technique may be used.

These novel methods of in situ nucleic acid reactions (ISNAR) are suitable for cellular diagnostic material and for research. A further embodiment is the formation of a long amplified product during the first round of amplification, with the use of only one primer instead of a primer pair; second round amplification cycles may be reduced. Alternatively, successive rounds of reaction forming said long products may be performed, without the use of other rounds of amplification reactions. Visualisation of amplificants after reaction is performed either by detection of Digoxigenin - labelled dUTP (DIG-11-dUTP) or digoxigenin -labelled primers (DIG-primer) incorporation directiy into amplificants during the last round of single round or nested in situ PCR. Detection may also be performed by in situ hybridisation. Combined in situ RT/in situ nested PCR is suitable for cell suspensions, which are subsequentiy cytospun onto a surface for example a slide after reaction. Alcohol-fixation and unfixed cells gives optimal results, though formalin-fixed cells may be used. Suitable controls are used in all experiments and no significant back-diffusion of amplificants from the extracellular solution into cells is found. As a second confirmatory technique for cells in suspension, liberated amplificants from cells lysed after in situ PCR (two or more round) is examined by gel electrophoresis for detection of a band of expected size. The identity of this amplificant is confirmed by Southern blot analysis and by sequencing studies. Isotonic solutions are used throughout to obtain optimal morphology, enhance specificity, and lower cost of the procedure, thereby permitting 40 or more cycles per round of PCR to be performed with ease. Cells can survive 80 PCR cycles with good morphology and without excessive product leakage using our reaction solution and protocol. Complicated long denaturation, "hot start" techniques, enzymatic or detergent pretreatment, postfixation steps, proved unnecessary with this technique for dispersed cells. By minimising damage to membranes, small reaction substrates may enter the cell but longer amplified products for example (252 base pairs) may be less likely to diffuse out of cells or backdiffuse into cells, especially with incorporation of label into the amplified product. Similarly, large inhibitors or degrading enzymes (e.g. RNAses) are prevented from entering cells. The technology is cost-effective, may be used with a majority irrelevant DNA background, and is a highly efficient method for nucleic acid (RNA or DNA) evaluation in routine cytological practice in addition to research. By comparing relative amounts of RNA and DNA it might be possible to ascertain the degree of expression of a specific gene at a cellular level.

The field of molecular pathology has recently made immense advances, drawing routine diagnostics ever closer to precision technology previously reserved for specialised research. In situ hybridisation (ISH) preserves cellular morphology, but not infrequently has a relatively high detection threshold. By contrast, the polymerase chain reaction (PCR) is very sensitive, but until recentiy, cell lysis was required to liberate DNA. For mRNA, detection using ISH is notoriously difficult due to the almost universal presence of RNases, using principles different from those described for this invention, and required extracellular confirmation. By using various pretreatments for cellular preparations, small perturbations may be created permitted relatively small components for reaction to enter the cell, but preventing relatively large amplificants perhaps due to their size and/or charge from leaving or entering the cell, and protecting the cell from the effects of lysosomal or extracellular RNase and DNase. This must be combined with excellent morphological preservation and resolution of cellular features for utility.

The purpose of this invention is to optimise conditions for a new technique for amplifying targets by nested reaction within cells, which by its simplicity and effect is exquisitely sensitive, specific and reproducible, and therefore possesses immense potential both as a routine cytological diagnostic tool and for research. Measles virus was chosen as a model to test the technique due to its characteristic cytopathic effect, and predominandy cytoplasmic signal distribution by in situ hybridisation (ISH) widiin infected cell cultures. The presence of exclusively DNA nuclear signal would be indicative of a false result. We attempted to exclude pretreatments including long denaturation or "hot start" methods, enzymatic or detergent treatments or postfixation steps so that morphology would be maintained with the minimum interference. Simple air-drying rather than long dehydration procedures was also used, and storage at -20°C instead of -70°C to preserve cellular detail. Isotonic solutions were also used to optimise morphology, dramatically reduce costs of expensive materials, increase specificity over more concentrated reaction solutions, and permit up to 40 cycles of in situ PCR per round to be performed with ease. We employed identical reaction mixtures without alteration of magnesium or Taq polymerase enzyme concentrations for both intracellular and for free (cell lysates or extracted) nucleic acids, so that suitable controls could be used at every stage of our protocol. Although much previous work on intracellular RT/PCR has concentrated on ISH detection, we have confirmed our findings, not only by ISH, and by visualisation of digoxigenin label incorporated into die second round of nested in situ PCR, by simultaneous gel electrophoresis for bands of expected size, by Southern blot analysis and by sequencing studies. Without gel electrophoresis and sequencing studies, it may be very difficult to suggest that the in situ PCR reaction has occured. This is of great interest, since nested PCR reactions with detection of incorporated label into amplificants are highly specific when compared to single round PCR, and generally much simpler to perform that ISH.

For the purposes of this invention, the term amplicon or amplificant shall be identical to amplified product produced by a nucleic acid amplification or successive replication reaction.

The use of amplification technique within cells or extracted from cells or cell lysates are generally suitable in two situations:

(i) Amplification of an isolated nucleic acid target, such as in sperm analysis and preimplementation embryo diagnosis

(ii) The DNA target is surrounded by an excess of background irrelevant DNA molecules (majority DNA background). This situation is particularly common in clinical situations, for example prenatal diagnosis, infection, neoplasia and forensics. It may also be pertinent to samples of poorly preserved clinical material, or small samples of cellular materials which are insufficient for pathological diagnosis. The technology may be performed within prokaryotic cells, eukaryotic cells (e.g. humans, animals) or in plants. In addition, the technology may be applied to nucleic acids within artificial membranes (e.g. viruses), and to viruses.

With a decreasing number of DNA molecules, the probability of having a target sequence in a test sample decreases according to the poisson distribution P(n) = e "τn^m n! where P(n) is the probability of having n molecules per test aliquot, m is the mean number of molecules in the sample, and e is a constant (2.718...).

For example when there is on average one target molecule per test aliquot, then the probability of having no target in a particular test is 0.37. Sampling error is likely to contribute both to discrepancies in the literature and in clinical practice. Clearly it may be useful to increase target for detection.

The amount of extracted DNA used in a PCR assay limits the sensitivity for a particular assay. For example, one microgram of DNA is derived from as many as 150,000 cells. Detection methods may alter the sensitivity of detection, for example the use of fluorescent labels in amplified products rather than by detection using ultraviolet-mediated detection by gel electrophoresis. Alternatively, a detection probe itself may be amplified, or the use of secondary antibodies with antibody mediated detection of label.

Even though some nucleic acid target is lost during extraction of nucleic acid from cells, amplification is generally considered to be much less efficient widiin intact cells than amplification performed on extracted nucleic acid in the extracellular environment. This largely because reaction reagents have greater access to nucleic acid targets when no membrane barriers exist. Clearly the effect of a nested reaction, which improves both reaction sensitivity and specificity over single round amplification reaction, is particularly advantageous when reaction efficiency is relatively poor, for example amplification reactions performed within intact cells.

A number of strategies may be employed for enhancing amplified product in cases of a majority DNA background. These include the use of "hot-start" technologies, amplification refractory mutation system (ARMS), the use of nucleic acid detection probes by in situ hybridisation of amplified products (ISH); or by the use of nested reactions with incorporation of label in the second round or after multiple rounds of reaction.

In "hot-start", one or more reagents necessary for amplification reaction, for example PCR, is witiiheld from the reaction mixture until the temperature is raised above, for example, 50°C. This is performed in order to reduce the annealing of specific primers to incorrect nucleic acid targets (misprinting). However for samples in solution, Hot Start alone is incapable of preventing n ispriming. It is also of litde value where amplification reactions are performed largely at temperatures below, 50 °C, such as the nucleic acid sequence based amplification (NASBA) or self-sustained sequence reaction (3SR). Hot start also suffers from problems of starting all sample reactions simultaneously, mixing of missing reagent and contamination.

The ARMS system is a useful method of amplification of a minority sequence in a majority background. In this method since the specificity of PCR is conferred by the 3 '-end of the primers; the difference is amplification between a specific primer or a primer mismatched at the 3'-end, and a common primer, should be indicative of a specific product. The ARMS system however only confers relative selectivity since the mismatched primer will permit some amplification, albeit at much reduced efficiency compared to the specific primer. Hence the ARMS system may be used with difficulty when small numbers of specific target are present, and may require the use of a large number of PCR cycles - which would itself enhance the formation of nonspecific product. Similarly, methods to enhance specificity such as reduction of dNTP concentration or enzyme may reduce the amplification of small amounts of specific target. For certain primer - target combinations, ARMS only allows the detection of a target molecule amongst at most a 40 fold excess of other background molecules.

ISH is another method of detecting specific amplificants over a majority DNA background including nonspecific amplified products. ISH itself however suffers from problems of nonspecific hybridisation of probe to majority background signal, necessity in some cases of altering membrane permability for probe access or prevent leakage of hybridised target/probe complex, harsh hybridisation pretreatments and conditions which may severely compromise cellular morphology and detail, and relative insensitivity of ISH in cases where amplification of target is very poor. In addition, ISH protocols vary according to the target nucleic acid, so there is no uniform ISH protocol or procedure, which is of concern where reproducibility is required. ISH protocols are not infrequently complex and require some expertise.

Nested (two round) PCR in solution on liberated DNA offers very high specificity compared to single round PCR in solution, and die same may be pertinent for intracellular reactions. Clearly, highly specific direct incorporation of amplified product in the second round of in situ nested PCR would be of immense value. Another advantage of direct incorporation of label into the second round of in situ nested PCR is that increased leakage of amplificants from cells with time during in situ PCR may be reduced by the presence of bulky label (KOMMINOTH et al, DIAGN MOL PATHOL 1:85-97 (1992). It may be the case that up to 40 cycles of PCR for some genetic sequences in solution are required to give a detectable band on gel electrophoresis (INNIS et al; PCR Protocols, A Guide to Methods and Applications, Academic Press, P.8-9). By contrast, 80 cycles (i.e. 2 rounds nested) of in situ PCT might compromise morphology. Simultaneous use of a bulky label into the amplified product which is less likely to transgress the cell membrane, and isotonic solution for optional morphological preservation would clearly be beneficial. Prior conversion of RNA to DNA, for example intracellular reverse transcription, may be performed prior to in sitii nested amplification.

With nested PCR reactions, the second round PCR product is used to select out specific amplified products from any nonspecific first round amplified products diereby dramatically enhancing specificity. Using this strategy, the sensitivity of in vitro reaction is enhanced 100-1000 fold, to the extent that a single target may be amplified from a background DNA equivalent to 300,000 cells. The term nested reaction includes heminesting whereby the second round PCR is carried out using one internal primer and one of the original first round PCR primers. The nested reaction is particularly of value if reaction is performed within cells where reaction efficiency is poor and a majority DNA background is present.

The performance of in situ amplification such as nested PCR reaction within cells is advantageous over extracted in vitro DNA for a number of reasons. Firstiy, extraction of nucleic acid is not necessary, thereby preserving cellular morphology and detail, permitting descrimination of cell subpopulations. Extraction generally requires a relatively large amount of clinical sample, preferably unfixed though fixed material may be used. In situ amplification does not require large samples, and detection methods within individual cells may be more sensitive tiian in vitro gel electrophoresis or Southern blot techniques, where a "dilution" effect of large numbers of negative cells is present. There is no need for target enrichment compared to some in vitro techniques. Secondly, contamination and reaction inhibitors are major considerations for in vitro amplification such as PCR. Inhibitors may prevent reaction and may be present even in material extracted from paraffin blocks. The presence of inhibitors is particularly pertinent where inhibitors are in high concentration such as body fluids. Our findings suggest that maintaining membrane integrity may limit access of inhibitors into cells. With regard to contamination, sources include "carryover" effect of previously amplified product from previous reaction or contamination from external sources. For in situ PCR, cells need only be washed before reaction. Cell compartmentalisation into cytoplasm and nucleus may render evaluation of contamination - if present - relatively straightforward. Back-diffusion of preformed amplificants into cells may produce a characteristic membrane "rim" effect. In vitro PCR generally requires 3 separate requirements to prevent contamination - an area where reagents are prepared - an area where patient sample is prepared and added to the reaction mixture and cycling is performed - and a detection area. Such a process clearly limits practical application.

A major problem of in situ amplification techniques generally, particularly those which have harsh pretreatment steps before reaction or during ISH detection, is the leakage of amplified product out of cells. Indeed, most workers advocate postfixation to prevent leakage (Komminoth & Long, Virchows Archiv. B, Cell Pathol. (1993) 64: 67-73); this would prevent a second round reaction since the second round reaction substrates could not penetrate cell membranes. Contrarily, excessive permeabilisation techniques proposed by most workers in order to permit reaction substrates to enter the cell would similarly cause leakage of product; this would be particularly evident as thermal cycling increased. Even workers who claimed that no postfixation was required needed a predigestion step. No postfixation of predigestion step is needed in our protocol. In this way, small perturbations may be maintained without the problem of excessive leakage; indeed both alcohol fixation or no fixation provided adequate results. The success of reaction on samples with no fixation may be due to perturbations in the membrane caused by heat, Tween present in the Taq polymerase dilutent or gelatin. In this way small reaction substrates may enter cells but excessive leakage of amplified product prevented. The use of overlapping primers to produce very long products which crossanneal would prevent the sample being used as a substrate for subsequent nested reactions.

We use isotonic reagents in our reaction which preserves cell morphology, does not exaccerbate problems of membrane permeability caused by predigestion (used in other methods) or heat damage by adding excessive osmotic pressure conditions which might enhance membrane damage. In addition, diluted solutions of PCR reactants - in proportion to each reagent - enhance specificity over more concentrated solutions and reduce cost.

A further problem of excessive predigestion, such as proteinase K, is the loss of histone binding properties. Although cross-linking fixatives generally require a predigestion step - unlike precipitating fixatives such as alcohol - loss of histone binding may increase the susceptibility of DNA to damage and repair processes. Damaged DNA is no longer a suitable template for amplification reaction. Finally, alteration in membrane permeability by excessive predigestion is often in a nonuniform manner, thereby causing alteration of reaction efficiency between cells within the same sample.

Therefore sufficient target must be present after die first round of in situ amplification reaction; providing tiiere is gentie or absent predigestion, even distribution of fixation and - if required - gentie predigestion, sufficient permeability to allow reaction substrates to enter whilst preventing reaction inhibitors from entering cells and reaction products from leaking out or back diffusing in from the extracellular medium, no postfixation step is used which might alter the penetration of second round PCR substrates from entering cells, no crossannealing of reaction products occurs after the first round reaction, the DNA is not largely damaged by excessive predigestion together with thermal cycling and the reaction in situ is sufficiendy efficient and specific to generate sufficient target for the second round reaction. Our method provides the first technique which can reprodicibly perform the above conditions to perform in situ nested reactions whilst preserving excellent morphology.

A further embodiment of the present invention is the performance of a few PCR cycles first round of amplification using a single primer, which would produce long products. These long products are more easily retained than short products when only a few cycles are performed. In tiiis way the appropriate sequence may be "screened" within a majority irrelevant DNA background and retained within cells. With the second round of PCR, the first few cycles of second round PCR "screen out" the amplified sequences of the first round, and then amplify them. Therefore targets may be screened out during the first few cycles and retained, without having to perform 30 or more cycles which might include amplification of nonspecific targets. With long reaction products after the first round reaction, the number of cycles in the second round in situ amplification reaction may be reduced, for example to 10 cycles.

It should also be noted that the amount of label (e.g. Digoxigenin -11-dUTP) used is significantiy lower than that recommended for PCR on liberated DNA in solution (LION T, et al, ANALYTICAL BIOCHEMISTRY 188-335-7; 1990) ) both for single round or nested in situ PCR using our protocols. No fixation or gentie precipitation fixation, no pretreatments and no postfixation steps are required by our protocol, contrary to many other published methods of single-round in situ PCR including those where over 50 cycles are performed.

Labels that may be used in the present invention include either nonisotopic (absorbance or fluorescence) or isotopic (radioactive) signal attached to a primer or dNTP or other reaction substrate. Alternatively detection may be by a labelled probe after reaction.

A form of intracellular (in situ) nested PCR reaction different from the present invention has been reported (EMBLETON et al, NUCLEIC ACIDS RESEARCH, Vol 20, No. 15, p.3831-3837, 1992) in which the strategy involved several steps:

(1) cDNA synthesis or different sequences of RNA template by in situ RT.

(2) in situ PCR using primers with "tagged" and complimentary tail so that the 2 sequences amplified may "assemble" together.

(3) assembly of the 2 amplified sequences into a long DNA template.

(4) second round in situ PCR witii primers amplifying a sequence of the assembled DNA template produced after first round in situ PCR. The primers for so-called second round "nested" in situ PCR have tags present to introduce primer sites for "universal" fluorescent primers.

(5) third round in situ PCR with fluorescent-labelled primers that can prime to the 5' or 3' ends of 2nd round primer tags.

(6) Detection by confocal microscopy and flow cytometry.

A major problem of assembled reaction is that the extent of linkage between two sequences in one cell type is likely to depend inter alia on whether both mRNAs are transcribed and amplified efficientiy and to high copy number, and on whether there is competition from amplified DNA leaking from other cells to take part in the assembly process.

Our invention differs fundamentally from the method of Embleton et al, and offer significant changes over that work of value especially for routine diagnostics. Firstiy, our invention concerns only one sequence region of a template rather than an assembled template, so that much smaller regions of a template may be investigated. In addition, in our invention the outer flanking primers (for example MV1 and MV2) amplify a specific sequence of a template during first round in situ PCR. This amplified sequence will then itself form a template for second round in situ PCR using inner flanking primers using primers internal to those used for previous rounds of amplification. Alternatively, a long product produced after a few cycles in the first round may be used as a template for a second round reaction limited to as litde as 10 cycles. A relatively large amount of highly specific amplificant may therefore be obtained for one sequence region. Finally primers used in the present invention do not themselves act as primer binding sites for subsequent rounds of amplification.

Our invention also uses alcohol fixation or no fixation rather than formalin, which is advantageous since many cytological fixatives are alcohol-based. No pretreatment with detergent is necessary or recommended, since excellent signal is obtained without detergent pretreatment and morphology is not compromised by this pretreatment. Our method also lends itself to evaluation and comparison of cellular signal intensity and distribution, both by ISH after single round in situ PCR and with direct incorporation in the second round of in situ nested PCR after respective reactions (single round in situ PCR for ISH detection or nested in situ PCR reactions with incorporation of label in the second round of in situ PCR) are performed on different samples drawn from the same cellular specimen. In addition, aliquots drawn from every stage of the intracellular reaction process may be used and the amplificants examined by gel electrophoresis, Southern blotting and/or sequencing studies. In this way, independent confirmation techniques may be used.

Although the present technique was used for the polymerase chain reaction (PCR), the technology may be used for any amplification mixture involving at least two rounds of amplification in which the product of the first round reaction is used as a direct template for a second round reaction. The reaction product of the final round reaction should preferably but not necessarily be small, for example 100-400 base pairs. Advantages of small reaction products, providing there is no excessive leakage, include resistance of small sequences to damage relative to long sequences, thereby improving product yield. Amplification or successive replication reactions which may be used for the present invention within cells include in situ PCR, the ligase chain reaction (in situ LCR), the primed in situ extension reaction (in situ PRINS), the nucleic acid sequence based amplification (in situ NASBA) or the self sustained sequence reaction (in situ 3SR), strand displacement amplification (in situ SDA), transcription-mediated amplification (in situ TMA) or any other successive replication or amplification. As described, either high temperature reactions for example in situ PCR or in situ LCR, or an isothermal reaction for example in situ NASBA may be used. It may also be used for the amplification refractory mutation system within cells (in situ ARMS).

The invention is potentially valuable in clinical medicine, for example for detecting low levels of infection, detecting specific genetic changes in tumours and lymphomas, discrimination between different alleles, and analysing congenital diseases or embryological mechanisms. The technique can be applied to the pathogenesis (mechanism), aetiology (cause), diagnosis and prognosis of disease. The technique is also important in cell biology, microbiology and virology, veterinary and plant biology or pathology, and pathology in medicolegal medicine including forensics and paternity testing. Finally, there is potential for its use in gene therapy, and within artificial membranes for example within liposomes.

METHODS AND MATERIALS

A flow sheet of the process is set out below. FIXATION or NO FIXATION

Wash Cells

IN SITU RT

Wash Cells

IN SITU "SINGLE" ROUND PCR or PRINS

Wash Cells

IN SITU "SECOND" ROUND PCR

Aliquot of CELLS LYSED

onto slides)

^• Gel Electrophoresis • Detection of incorporated label

• Southern Blot (highly specific if incorporated in second round of in situ PCR)

- Sequencing Studies ^• ISH CELLS

Vero cells were infected (or not infected) with EDMONSTON strain measles virus and cultured in RPMI 1640 with Eagle's minimal essential medium, 1 % glutamine, 2% fetal bovine serum, penicillin G (200U/ml), streptomycin (200 micrograms/ml), amphotericin B (0.5 micrograms/ml). When 50-100% of cells shows significant cytopathic effect, cells were scraped into suspension, fixed in 95 % alcohol for 5-30 minutes and centrifuged in a 1.5ml Eppendorf tube at 1000 rpm for 1-3 minutes, after which the supernatant was discarded. Measles infected or uninfected vero cells were either maintained separately, or mixed together to produce mixed measles infected and uninfected cell populations. 0.5-2mm height of the cell pellet kept in suspension for subsequent ISNAR reactions. PRIMERS AND PROBES

Previously used primers with specificity for the measles nucleoprotein 'coding' sequence were used (Table 1). These have been described by M.J. Taylor et al J. Gen. Virol. (1991) 72 : 83-88.

Table 1 : Primers with specificity for the measles nucleocapsid 'coding' sequence

TYPE SIZE Nucleotide Size of

(Bases) Sequence Amplified Fragment (bp)

Outer MVl 21 5'TTA GGG

Flanking (sense) CAA GAG ATG

Primers GTA AGG 3' 432

MV2 21 5'GTT CTT CCG (antisense) AGA TTC CTG CCA 3'

Inner MV3 21 5'AGC ATC

Flanking (sense) TGA ACT CGG

Primers TAT CAC 3' 252

MV4 21 5'AGC TCT (antisense) CGC ATC ACT TGC TCT 3' A previously described digoxigenin-labelled single stranded riboprobe was used. Cosby SL, McQuaid S, Taylor MJ, Bailey M, Rima BK, Martin SJ and Allen IV. 1989; J. Gen. Virol.: 70; 2027-2036. IMMUNOHISTOCHEMISTRY

Formalin-fixed paraffin-embedded measles infected or uninfected or mixed measles infected and uninfected vero cells were submitted for an immunohistochemical assay using a monoclonal primary antibody against the nucleocapsid protein of die measles virus (Serolab, Crawley, Sussex, UK). This was followed by detection using a peroxidase-linked secondary antimouse antibody, the method of which is well- known.

INTRACELLULAR REVERSE TRANSCRIPTION (IN SITU RT) FOR CELLS IN SUSPENSION

Centrifugation of the cell suspension in PBS was performed at 1000 rpm for 1-2 minutes and the cell pellet transferred to a 0.5ml Eppendorf tube and gentiy resuspended in different pre-treatments (95 % alcohol, or no fixation) for a period of 0.5-30 minutes. The effect of proteinase K (lOμg/ml. 10-25 minutes, 37°C) was also investigated. Subsequentiy, centrifugation was repeated at 1000 rpm for 1-3 minutes and the supernatant removed and cells washed. The cell pellet was resuspended in the reverse transcription mix (per sample) composed of MgCL, 20μl (25mM), 10 x PCR Buffer 10μl, distilled water 5μl, each dNTP 5μl (lOmM), Reverse Transcriptase 5μl (50U/μl), RNase inhibitor 5μl (20U/μl), Random Hexamers 5μl (50μM). Two drops of mineral oil were dropped on to the surface and each tube was placed in a thermal cycler (Techne PHC-2, Techne Ltd, Cambridge, UK). Reverse Transcription was performed with the following specifications: 28°C 10 minutes, 42°C 2 hours, 95°C 5 minutes, 12°C 5 minutes. After reaction, a lOμl aliquot was taken from each sample for gel electrophoresis and a second 20μl aliquot sample was removed for cytospin as a control for ISH; the remainder was submitted for in situ PCR amplification. For Digoxigenin incorporation at the in situ RT stage, the 10μl (lOmM) deoxyribonucleoside thymidine triphosphate (dTTP) was reduced to 5μl (lOmM) dTTP and 4μl (25nM) Digoxigenin - 11 -deoxyribonucleoside uridine triphosphate (DIG - 11 - dUTP). (Boehringer Mannheim UK). An osmometer was used to confirm that the reaction mix was isoosmotic with human serum.

After intracellular reverse transcription, overlying mineral oil was removed and after taking a 20μl aliquot for gel electrophoresis, the sample was centrifuged at 1000 rpm 1-3 minutes. The supernatant was discarded and the cell pellet resuspended in a relatively isotonic PCR reaction mixture as follows: MgCl₂ 10μl 24mM, 10 x PCR mixture 20μl, each dNTP 5μl lOmM, lμl 5U/μl Taq polymerase (Perkin Elmer RNA-PCR kit, #N808-0017), together with 164μl distilled water and optionally 200μl PBS. For single round in situ PCR with in situ hybridisation detection (ISH), l-2μl (480 micro g/ml) of each inner flanking primer pair MV3 and MV4 was added to the reaction mixture. By contrast, for nested in situ PCR reactions, l-2μl each outer flanking primer pair 480 micro g/ml MVl and MV2 (table 1) was added to the above reaction mixture for the first round in situ PCR. Cycling conditions were as follows:- Initial Denaturing 94 °C for 10 minutes; then Denaturing 94°C 1 minute, Annealing 58°C 2 minutes, Extension 72°C 1 V_ minutes, 30 to 40 cycles; 10 minute extension after 40 cycles PCR. After centrifugation and resuspension of the cell pellet in PBS, centrifugation was repeated (1000 rpm 1-3 minutes). For the second round of in situ nested PCR the cell pellet resuspended in an identical reaction mix as for first round in situ nested PCR with replacement of outer flanking primers (MVl, MV2) with inner fl_uιking primers (MV3, MV4) at identical concentration. After adding 2 drops of mineral oil, thermal cycling was performed as before. If DIG-11-dUTP incorporation into die second round amplificant of nested PCR was performed, the 5μl 10 mM dTTP was replaced by 2.5μl (lOmM) dTTP and 4μl (25nM) DIG-11-dUTP. In experiments where labelled primer (MV4) was used at identical concentrations to unlabelled primers, labelled DIG-11-dUTP was omitted, and 5μl (lOmM) dTTP was used. A further series of experiments was performed in which the first round of thermal cycling (5 cycles) was performed with only 1 primer (primer MVl or MV2 only) and compared with simultaneous experiments in which two primers were added (MVl and MV2). For the second round of in situ PCR, only 10 cycles of thermal cycling were completed; identical cycling parameters and reagents were used as for in situ nested PCR.

After thermal cycling, the mineral oil was removed and a 20μl aliquot was taken for gel electrophoresis, Southern blotting or sequencing studies. After centrifugation of the remaining cell suspension at 1000 rpm for 1-3 minutes, the cell pellet was resuspended in PBS and cytospun at 1000-3000 rpm for 1-3 minutes onto PLL- or APES-coated slides, and allowed to air dry before storage at -20°C. Subsequendy, visualisation of incorporated Digoxigenin into amplificants, or ISH detection was performed on these stored slides.

GEL ELECTROPHORESIS

Samples (20μl) of cell suspension for gel electrophoresis after respective ISNAR reactions were freeze/tiiawed three times and centrifuged at 15000 rpm 3 minutes.

12μl of the supernatant of each sample was then mixed with 4μl loading dye. Gels

(2% agarose in TAE 1 x) with 3μl (10/mg/ml) ethidium bromide incorporated were prepared and gel electrophoresis performed according to standard well-known techniques.

SOUTHERN BLOTTING

Southern blotting was performed using a radiolabelled riboprobe identical to that used for in situ hybridisation. The technique of Southern blot analysis is well known.

SEQUENCE ANALYSIS OF AMPLIFICANT

Sequence analysis was accomplished according to standard well known techniques. CONTROLS

Rigorous controls were used through the protocol. For each in situ RT or in situ PCR reaction, samples with all reaction components, but with cells replaced by either cell lysates or by no cellular material were included. Other controls included measles-infected ceil samples placed in identical in situ RT or in situ PCR reaction mixtures including appropriate primers as test material but without enzyme added (reverse transcriptase for in situ RT or Taq polymerase for in situ PCR), or no primers (but appropriate enzyme added) or irrelevant primers for in situ PCR. In addition, uninfected vero cells were subjected to identical in situ RT and in situ PCR reactions as their measles-infected counterparts. In addition for the ISH detection, dissimilar probes to the measles-specific riboprobe were employed. For DIG-11-dUTP incorporation experiments in the second round of in situ nested PCR controls included DIG-11-dUTP incorporation at the in situ RT or first round of in situ nested PCR steps, in addition to in situ nested PCR without DIG-11-dUTP incorporation into the second round of in situ nested PCR. With experiments using labelled primers, irrelevant labelled primers were used as controls. For all ISH and experiments involving incorporation of DIG-11-dUTP into amplificants, evaluation was performed by blinded studies.

The possibility of "back diffusion" of amplificants from the extracellular reaction mixture into cells was investigated.

A sample of measles-infected vero cells was submitted for combined in situ RT/in situ nested PCR with incorporation of DIG-11-dUTP or labelled primers into amplificants in the second round of in situ PCR. After reaction an aliquot of cells was removed and cytospun onto slides to confirm the presence of incorporated label and amplification by digoxigenin detection, and a second aliquot or cells was lysed to confirm the presence of amplificant band of expected size by gel electrophoresis. The remainder of the cells were lysed by freeze-thawing thrice, and after removal of the cell debris by centrifugation 3000 r.p.m. 3 minutes the supernatant was mixed with an equivalent number of measles uninfected vero cells. Two rounds of thermal cycling (30-40 cycles each) were performed followed by washing cells in PBS, before centrifugation (1000 rpm 3 minutes) on to PLL-coated or APES-coated glass slides and airdrying. The presence of "back diffusion" of DIG-11-dUTP labelled preformed amplificants into these uninfected vero cells was evaluated by detection of digoxigenin.

ISH detection of measles amplificants using digoxigenin-labelled riboprobe

Probe used as described previously (see COSBY et al reference above) with minor changes.

1. Wash in diethylpyrocarbonate (DEPC) distilled water (5 minutes).

2. Wash in PBS/2 mM Ethylene Diammine Tetracetic Acid (EDTA) (5 minutes).

3. Proteinase K treatment (100-1000 μg/ml) 37 °C for 15 minutes in above buffer (PBS/EDTA).

4. Rinse 0.2% gly cine/PBS mixture (5 minutes).

5. Wash in PBS (5 minutes).

6. Fix in 4% paraformaldehyde/PBS mixture (2 minutes).

7. PBS rinse.

8. Denature slides in DEPC distilled water for 10 minutes at 95 °C; then plunge into ice cold water, dehydrate in alcohol and air dry.

9. Hybridisation at 42 °C with digoxigenin-labelled antimeasles riboprobe

(1 hour). The riboprobe was obtained as follows: A transcription vector (pgem I) containing the 3' sequence of the measles virus N-gene inserted in the Bam HI site of pGem I was provided by Dr Louise Cosby, Queen's University, Belfast. The riboprobe was syntiiesised using T7 polymerase promotor by standard techniques. This biotin or digoxigenin-labelled riboprobe has been used extensively by Dr. Cosby and ourselves, and is specific for measles virus. It is known to hybridise with all measles virus nucleocapsid sequences contained within the Genebank data systems and does not hybridise with the closely related morbillivirus, canine distemper virus.

10. Rinse in 2x SSC for 20 minutes at 37°C.

11. RNase A 100 μg/ml 2x SSC at 37°C for 15 minutes.

12. Rinse in 2x SSC, then lx SSC for 20 minutes.

13. Rinse in Tris buffered saline (TBS).

14. Quench in 10% normal rabbit serum (30 minutes).

15. Incubate in a mouse digoxin (D8156 Sigma clone Dl-22) diluted 1/10,000 TBS (1-1 Vi hours).

16. 3 washes in TBS (5 minutes each).

17. R α M IgG 1/25 (Dako) and 1/25 normal human serum distilled in TBS.

18. 3 washes in TBS (5 minutes each).

19. Alkaline phosphatase Anti- Alkaline Phosphatase complexes (APAAP) diluted 1/50 TBS (for 30 minutes).

20. 3 washes in TBS (5 minutes each).

21. R M IgG (10 minutes).

22. 3 washes in TBS (5 minutes each).

23. APAAP diluted 1/50 TBS (for 10 minutes +).

24. 3 washes in TBS (5 minutes each).

25. Rinse in 100 mM Tris pH 9.2, 100 mM sodium chloride, 50 mM magnesium chloride.

26. Detect using Sigma fast red (5-30 minutes).

27. Rinse in TBS, tapwater, and counterstain with Crazzi' s Haematoxylin. Rinse and mount in glycergel (Dako).

Detection of Digoxigenin-labelled amplificants According to standard protocols.

1. The slides were rinsed in 1 x (TBS) or a suitable isotonic solution (PBS).

2. Digestion was performed 10 μl/ml proteinase K (Sigma P4914) in PBS/ Ediylene diamine tetracetic acid for 25 minutes at 37 °C (200μl/section).

3. Wash in PBS. 4. Quench in 1 in 25 normal horse serum diluted in 1 in 25 TBS.

5. Add antidigoxigenin-AP-Fab (Boehringer) fragment diluted 1/200 in TBS for 45 minutes.

6. Wash in TBS - 3 washes for 5 minutes each.

7. Equilibrate in AP2 buffer (50 ml Tris, 50 ml MgC 50 mM, 400 ml of water).

8. Fast red stain. 10 ml distilled water and Tris tablet - Tris (hydroxymethyl), Sigma F4523) - to 10 ml ddH₂0 add 1 Tris tablet dissolve then add 1 Fast red tablet, containing fast red substrate levamisole.

9. Wash in PBS. Counterstain with haematoxylin (Crazzi' s).

RESULT

HISTOLOGICAL STAINING TECHNIQUES AND

IMMUNOHISTOCHEMISTRY

Cellular morphology was remarkably well preserved when stained by haematoxylin and eosin stain or Carazzis haematoxylin or papanicolou stain with only slight loss of detail even after combined in situ RT/nested PCR, particularly with intracellular nucleic acid reactions within cells in suspension.

Measles infected vero cells gave positive predominantly cytoplasmic signal with the monoclonal measles-specific antibody. Conversely measles uninfected vero cells showed no detectable signal by this method. Mixed vero infected and uninfected vero cells revealed positive signal in areas of confluent cells but no significant signal in dispersed cell areas.

GEL ELECTROPHORESIS

Cells lysed after combined in situ RT/PCR using MV3 and MV4 primers shows a band of 252 bp which was the expected size. Cells lysed after combined in situ RT/PCR (first round only using MVl and MV2 primers) showed a band of appropriate size (432 base pairs); after 2nd round nested PCR using MV3 and MV4 primers die expected band (252 base pairs) was identified and die band present after 1st round PCR (432 base pairs) was markedly less visible. The small amounts of Digoxigenin- 11-dUTP incorporation into amplificants resulted in slightly slower band sizes as amplificants without Digoxigenin- 11-dUTP incorporation. No band was seen either before or after reverse transcription alone and negative control samples (no Taq polymerase enzyme or no primers added). In addition, uninfected vero cells and control reaction mixtures without any cellular material, showed no bands.

SOUTHERN BLOTTING

Measles-infected vero cells were lysed before or after combined in situ RT/in situ PCR and gel electrophoresis performed revealed bands of expected size only if cells were kept intact during reaction. Amplificants produces were submitted for Southern blotting with radiolabelled probes identical to those used for ISH experiments with specific measles virus. The results of Southern blotting revealed positive signal for the amplificants confirming specificity for measles virus. Controls gave expected results.

SEQUENCE OF AMPLIFICANT MATERIAL

Amplificants were sequenced and the findings confirmed that the amplificants were absolutely specific for the measles virus sequence amplified.

IN SITU HYBRIDISATION

A markedly increased signal within cells (alcohol-fixed, unfixed) was found after in situ RT, followed by single round (30-40 cycles) in situ PCR when compared with cells submitted for in situ RT only. The signal was predominantly cytoplasmic in distribution, though some nuclear signal was also present. Positive and negative control samples gave expected results. DIGOXIGENIN INCORPORATION INTO AMPLIFICANTS FOR CELLS IN SUSPENSION

No increase signal relative to background was identified in cells not submitted for ISNAR and in cells subjected to reverse transcription only, or for signal round and nested in situ PCR without Digoxigenin incorporation. By contrast, markedly increased cytoplasmic signal was present in cells subjected for combined in situ RT and in situ PCR with Digoxigenin- 11-dUTP added to the reaction mix. When Digoxigenin- 11-dUTP was introduced for one round of PCR (using MVl and MV2, or MV3 and MV4 primers) an impressive intracellular signal relative to controls was found. Very impressive signal was identified when Digoxigenin- 11-dUTP was incorporated into the second round of in situ nested PCR. Alcohol (with as litde as 3 minutes fixation) gave superior results to no fixation. Indeed with no fixation, no product was identified unless nested in situ PCR reaction was performed, though excellent signal was seen after in situ nested PCR. In all cases, staining mainly cytoplasmic in distribution, though some nuclear staining was present. Identical results were found when labelled primers replaced labelled dNTP.

When uninfected vero cells were soaked in preformed amplificants obtained from equivalent numbers of infected cells lysed following in situ RT/PCR reaction a peripheral faint rim-like staining, was found, but no significant cytoplasmic or nuclear signal was identified. Other controls gave expected results. After brief alcohol fixation (5 minutes) and in situ RT, a first round reaction (5 cycles) in which one primer (MVl or MV2) was used gave superior results to two primers (MVl and MV2), when followed by the second round nested with two primers (MV3 and MV4) reaction (10 cycles).

In summary, the present invention comprises a method of examining nucleic acids using the polymerase chain reaction (PCR) characterised in that the reaction is carried out in intact cells (intracellularly) using nested PCR. The cells are maintained in suspension and more than one round of nested PCR is employed. The method can be used to analyse or otherwise evaluate genomic DNA or cDNA, the latter having been obtained by intracellular reverse transcription from RNA. It has not previously been recognised diat multiple cycles intracellular PCR can be carried out without adverse effect on cell morphology. Gentle or absent fixation, no pretreatments and no posttreatments are used. Amplification of nucleic acids may be by a number of different means including in situ PCR and in situ PRINS. The method enables the use of in situ hybridisation to be avoided.

The method may be carried out with a kit for synthesising a target nucleic acid sequence within cells in suspension, comprising at least two primer sequences of which at least one has specificity internal to another, a dilutent, and at least one iso- osmotic reagent to wash cells, and one or more enzymes necessary for nucleic acid synthesis.

Claims

1. A method of amplification of nucleic acid which comprises treating a liquid suspension of cells or other cell-like structures containing nucleic acid within a natural or artificial membrane or coat protein, under conditions which maintain cell morphology, with an amplification reaction mixture designed to amplify a target nucleic acid sequence therein and continuing the amplification reaction for at least two rounds of amplification of said target sequence in situ.

2. A method according to claim 1 in which the amplified product of the first round is itself used as template for the second or subsequent round(s) of amplification using one or more primers internal to the primer or primers used for previous rounds of amplification.

3. A method according to claim 1 or 2, in which one or more sequences are amplified simultaneously.

4. A method according to claim 1, 2 or 3, in which the cells or cell-like structures are treated directly without intervening pretreatment to permeabilise the cell membrane.

5. A method according to any of claims 1 to 4, in which the cells are fixed non- disruptively prior to treatment with the amplification mixture.

6. A method according to claim 5, in which the cells are fixed with a precipitation fixative.

7. A method according to claim 6, in which the fixative is ethanol.

8. A method according to any of claims 1 to 4, in which no fixation is used.

9. A method according to any of claims 1 to 8, in which substantially isotonic solutions are used throughout.

10. A method according to any of the preceding claims, in which the same amplification method is used in the two or more rounds of amplification.

11. A method according to any of claims 1 to 10, in which different methods of amplification are used in die two or more rounds of amplification.

12. A method according to any of the preceding claims in which the PCR method is used.

13. A mediod according to claim 12, in which nested PCR is used.

14. A method according to any of claims 1 to 11, in which the PRINS method is used.

15. A method according to any of the preceding claims in which the amplification mixture contains labelled individual nucleotides or labelled primers for detection of the amplified product.

16. A method according to any of claims 1 to 14, in which detection of the amplified product is achieved by in situ hybridisation.

17. A method according to any of the preceding claims in which the target DNA to be amplified has been produced in the cells by in situ reverse transcription of RNA.

18. A method of amplifying DNA in cells which comprises treating the cells, without pretreatment designed to permeabilise the cell wall or membrane and without fixation of the cells except for optional fixation with a precipitation fixative, with a first amplification mixture to amplify intracellularly a first target DNA sequence in a first round of amplification, recovering the cells and treating the cells with a second amplification mixture to amplify intracellularly a second target DNA sequence in a second round of amplification, said first and second target DNA sequences being the same or nested sequences, optionally conducting a further round or rounds of intracellular amplification, and detecting or measuring the amplified DNA.

19. A kit for synthesising a target nucleic acid sequence widiin cells in suspension, comprising at least two primer sequences of which at least one has specificity internal to another, a diluent, and at least one iso-osmotic reagent to wash cells, and one or more enzymes necessary for nucleic acid synthesis.