Molecular Immunology, Vol. 29, No. 10, pp. 1229-1236, 1992 Printed in Great Britain.
0161-5890/92$5.00 + 0.00 Pergamon Press Ltd
USE OF QUANTITATIVE POLYMERASE CHAIN REACTION TO QUANTITATE CYTOKINE MESSENGER RNA MOLECULES WADASANKANANGAT,*ALAN SOLOMON~ and BARRYT. ROUSERS *Department of Microbiology, College of Veterinary Medicine, University of Tennessee, Knoxville, TN 37996-0845,U.S.A. and TDepartment of Medicine, Human Immunology and Cancer Programme, University of Tennessee, Medical Centre/Graduate School of Medicine, Knoxville, TN 37996-0845, U.S.A. (First received 31 October 1991; accepted in revisedform
12 March 1992)
quantitative polymerase chain reaction (PCR) using an internal control (Standard) RNA was developed for precise quantitation of human cytokine mRNA. The target mRNA and internal control were simultaneously reverse transcribed and co-amplified using the same set of primers. The amount of specific target mRNA is then quantitated by extrapolating against the standard curve generated with the internal standard. The internal control RNA consisted of linearly connected sequences of the 5’ primers of multiple cytokine genes followed by the complementary sequences to their 3’ primers in the same order. This structure of the internal standard enables one to use the same internal standard for quantitating multiple cytokine mRNAs. Using this approach, we estimated the induced levels of cytokine mRNA, IL2, IL6, and TNF-a to be 2.8 x 106,2.4 x lo5 and 3.4 x 1O*molecules per 1.Op g total cellular RNA of Jurkat, THP- 1 and HL 60 cells, respectively.
Abstract-A
Comparable values were obtained when quantitation experiments were done on another batch of THP-1 and HL-60 total cellular RNA. The excellent sensitivity and reproducibility makes this approach a valuable one in following changes in cytokine gene expression in wide variety of conditions, both in vivo and in vitro.
INTRODUCTION
The polymerase chain reaction (PCR) has found extensive use and has proven to be the technique of choice for the detection and analysis of minute amounts of DNA (reviewed by Erlich et al., 1991). Moreover, by employing a reverse transcription step, PCR is also a valuable means of detecting and characterizing low abundance transcripts (Chelly et al., 1988). As usually employed, the PCR approach is a qualitative procedure. However, there is clearly a need to develop a PCR strategy that will quantitate RNA transcripts since this approach should be more sensitive and easier to use than Northern blotting and solution hybridization. Various quantitative procedures have been reported, but most rely on the use of external standards and compare products generated in two separate reactions (Simmonds et al., 1990). Since both RT and DNA amplification may be nonlinear and subject to great variability, such approaches are extremely unreliable. Others have reported quantitative procedures based on a competitive reaction between an added competitor DNA fragment that differs from the target cDNA by having either a small intron or a mutated restriction enzyme site (Gilliland et al., 1990). However, this approach does not take into account variations in RT efficiency. A better approach for mRNA quantitation by PCR would be to employ an internal standard so that reverse transcription and coamplification of the RT products could be performed simultaneously. Detection of the SAuthor to whom correspondence should be addressed.
amount of unknown product generated using the same set of primers can then be achieved by designing the system so that test and standard products are of disparate size. Such an approach was described by Wang et al. (1989), and we have modified their system to quantitate multiple human cytokines mRNA in small quantities of sample RNA. The synthetic gene constructed could be used to quantitate 13 different human cytokines and the construct still has provisions to include more primer sequences of interest. In this report, we document the value of the approach and document induced levels of the three cytokines IL-2, IL-6, and TNF-o! in three cell lines. The approach we have described should prove valuable to follow changes in cytokine gene expression in a variety of disease situations and experimental manipulations in vitro. MATERIALS AND METHODS Oligonucleotide
preparation
Oligonucleotides were synthesized on an Applied Biosystem DNA synthesizer (Model 391). The 5’ and 3’ primer sequences (Table 1) were derived from published reports (Brenner et al., 1989). Construction of synthetic internal control RNA
gene and the preparation
of
A synthetic gene comprising the 5’ primers and 3’ primers was constructed by the technique of oligonucleotide overlap extension and subsequent amplification by PCR (Ho et al., 1989; Dillon and Rosen, 1990).
1229
S. KANANC;ATet al.
1230 Table 1. Human the construction
cytokine primers used for of the internal standard
Primers
Expected size of the target Number of base pairs
IL-l-a IL-l+ IL-2 IL-3 IL-4 IL-5 IL-6 IL-7 IL-8 IFN-ITNF-a TNF-fi TGF-fi PActin
816 810 462 459 462 414 639 707 302 501 702 608 200 548
The internal control was designed in such a way that -350 bp bands were obtained with each of the above mentioned primer pairs.
purification system. The purified RNA product was quantitated by spectrophotometric analysis at 260 nm.
CELL LINES
THP-1
THP-1, a human monocytic leukemia cell line was obtained from ATCC (No. TlB 202), and was maintained in RPM1 1640 containing 15% fetal bovine serum and 2 x lop5 2ME. These cells have Fc and C3b receptors and lack surface and cytoplasmic immunoglobulins. These cells stain positively for a-naphthyl-butyrate esterase, produce lysozyme and are phagocytic. HL -60 HL-60 is a human promyelocytic cell line obtained fromATCC (CCL 240) and was maintained in DMEM containing 15% FCS. These cells lack specific markers for lymphoid cells but express surface receptors for Fc fragments and complement. JURKA T
The strategy for the construction of the synthetic gene is shown in Fig. 1 and the structure in Fig. 2. The synthetic gene was cloned into an in vitro transcription vector/ pCR 2000 TA and was designated as SYNTA-2000 (Fig. 3). The linearized SYNTA-2000 was then used as a template to transcribe the internal control RNA using T, RNA polymerase employing a standard in vitro transcription protocol (Promega). The resulting RNA preparation was treated extensively with RNAse free DNAse (RQl DNAse), extracted with phenol: chloroform, ethanol precipitated or purified by a RNAID
PCR
Jurkat cell line is a human lymphoma derived T cell line. The culture was obtained from ATCC (CRL 8163) and maintained in RPM1 1640 with 15% FCS. Mitogen
stimulation
Jurkat cells were stimulated at lo6 cells ml--’ with phorbol myristate acetate (PMA) at a concentration of 10 ng ml-‘. THP-1 and HL-60 cells were activated at lo6 cells ml-’ with E. coli lipopolysaccharide (LPS) at a concentration of 5 p g ml ‘. All cells were harvested after 8 hr.
- 1
4 PRIMER - 2
* PRIMER
- 1
Fig. 1. Synthesis of internal standard by overlap extension PCR. 0.5 pg each of the long oligonucleotides (_ 100 bases) with overlapping ends (indicated by the broken lines) were mixed together in standard PCR (100 ~1 volume). The reaction was denatured at 94°C for 7 min, annealed at 55°C for 2 min and extended at 72°C for 3 min. The reactions went through seven additional cycles with the denaturation step reduced to 1.5 min. One microlitre of this reaction mix was used as template for second PCR. In the second PCR, 2 primers (21 bases each) spanning the 5’ and 3’ ends of the template formed in PCR I, was added to the standard PCR mix and subjected to 25 cycles of amplification (94°C 1.5 min; annealing at 55°C for 2 min; extension at 72°C for 3 min).
Quantitation
1231
of cytokine mRNA by PCR T 7 Eco RI
Not I Fig. 3. The synthetic gene produced by the overlap extension PCR was purified by gene clean and cloned into TA cloning vector which contains T7 promoter sequence.
45min. At the end of RT, the mixture was heated at 99°C for 5 min to inactivate the RT enzyme and denature the cDNA. PCR IL- I IL- 2 IL-
3
ILILILIL-
4 5 6 7
IL- x IIJN- r ‘rNI’_ a TNF- b TGV- b b- ACTIN I’OLY IllNIl
A III
Fig. 2. Structure of SYNTA-2000. The plasmid contains 5’ primers of 14 target genes connected in sequence followed by the complementary sequence of the 3’ primers in the same order. The multiple primer region is flanked by the T7 polymerase promoter upstream and by poly A downstream.
One-fourth of the RT mix was used as template for the amplification reaction. All PCRs were done in a 25 ~1 volume. The mixture contained 1X PCR buffer (10 mM Tris-HCl pH 8.3; 50 mM KCl, 1.5 mM MgCI,, 0.01% Triton X-100), 1.25 mM dNTP, 1.OU Taq DNA polymerase (Promega), 5 pmoles each of 5’ and 3’ primers, the 5’ primer being end labelled with y3’P using T, polynucleotide kinase. The unincorporated radionucleotides were removed from the probe by passing it through a Sephadex G-25 spun column (Pharmacia). The PCR mixture was overlaid with 50 ~1 mineral oil (Sigma), and incubated in a Perkin-Elmer/Cetus thermal cycler. The amplification steps involved denaturation at 94°C for l-2 min, annealing at 55°C for 2 min and extension at 72°C for 3 min. Appropriate positive (internal control and relevant primers) and negative (PCR mix with relevant primers but no template) controls were included with each batch of reactions.
Total cellular RNA
Total cellular RNA from induced and uninduced cells was isolated by the method of guanidium thiocyanate, phenol: chloroform extraction (Chirgwin et al., 1979). Reverse transcription
The internal control RNA and the target RNA (total cellular RNA) were reverse transcribed simultaneously in the same tube according to a standard protocol (Promega). Varying amounts of total cellular RNA (4.0 ,ug-+O.125 pg) and the internal control RNA, SYNTA-PCR 2000 transcribed in vitro from (8.8 x 10*+0.53 x 10’ molecules),* were reverse transcribed in a 20 ~1 reaction mix containing 1X RT buffer (10 mM Tris-HCl pH 8.8, 50 mM KCl; 0.1% Triton X-100), 5 mM MgCl,, 2 mM dNTP, 40 U RNAs inhibitor, oligo dT primer (0.5-1.5 ,ug) and AMV RT (10-20 units). The RT reaction was performed at room temperature for 30 min followed by incubation at 42°C for *No. of molecules were calculated using Avaqadro’s hypothesis (McQuarrie and Rock, 1984).
Fig. 4. Specificity of the internal control RNA and the Primers. The internal control RNA was reverse transcribed and amplified [25 cycles] using IL-2, IL-6, TNF-a primers in separate reactions in duplicate tubes. The products were resolved on a 2% agarose gel. Lanes b, c: IL-2; d, e: IL-6; f, g: TNFcr; h: Positive control. SYNTA-200.IL-6. Lane a is the 1 Kb DNA Ladder (Bethesda Research Laboratories).
S. KANANGAT et al.
1232
A -5
50000
'2
40000
$ cd
30000 20000 10000
0.0
0.2
0.4
0.6
n
Target
0
ht. Control
0.8
1.0
12
Templatedilutions Fig. 7. Quantitation of TNF-a-message in LPS-induced THP-1 cells. Four microgrammes of total THP-1 RNA and 8.8 x lOa molecules of internal control RNA were mixed and serially diluted as shown. RT was done on all dilutions, one-fourth of the RT mix was then amplified using TNF-cr primers, the 5’ primer being end-labelled with y”P. After 25 cycles, the PCR products were resolved on a 2% agarose gel and band radioactivity was measured.
No. of Cycles
@I
120000
-
.2: +. 100000
-
., z8
80000
-
._ 'D 2
60000
-
40000-
20000 1
0
! 10
I No.
Fig. 5. Relation amplification. 4.3 control RNA was mix was amplified
30
20
40
of Cycles
between no. of PCR cycles and rate of x lo* (a)/2.15 x 10’ (b) molecules of internal reverse transcribed and one-fourth of the RT using IL-6 primers, the 5’ primer being end labelled with y3’P.
Quantitative analysis Ten microlitres of each PCR mixture was electrophoresed on an agarose gel (2%). Gels were stained with ethidium bromide and photographed. The specific bands were excised from the agarose gel and the radioactivity was measured. The amount of radioactivity obtained
0
0
from excised gel bands were plotted against the template concentrations or the number of PCR cycles. The straight line graphs were drawn using the radioactivity counts of internal standard bands and target bands against their respective template concentrations and using the internal control graph, the specific quantity of target mRNA was calculated as shown in Figs 6-9. The best fit lines were drawn using the computer program C Cricket Graph 1-3.
Semiquantitation of /3-actin mRNA levels in unstimulated and stimulated cells Total RNA (2.0 pg each) obtained from unstimulated and stimulated THP-1 Jurkat and HL-60 cells were reverse transcribed in a 20 ~1 reaction mix each. Onefourth of each reaction mix was subjected to 25 cycles of PCR using b-actin primers, 5’ Primer being end-labelled with y32P. Ten microliters of the PCR product was analysed on a 2% agarose gel and stained with ethidium bromide. The /3-actin specific 548 base pair bands were excised and the radioactivity measured using a scintillation counter.
Target Ini.control -I
0:o
0:2
0.4
0.6
0.8
1.0
1.2
Templatedilutions
Fig. 6. Quantitation of IL-6 message in LPS-induced THP-1 cells. 4.0 pg of total THP-1 (LPS-induced) RNA, 8.6 x 10’ molecules of internal control RNA were mixed and serial twofold dilutions of the RNA mix were made. RT was done on all of them and one-fourth of each RT mix was amplified using IL-6 primers, the 5’ primer being end-labelled with y32P. The PCR products were resolved on a 2% agarose gel and the specific band radioactivity measured.
0.0
0.2
0.4
0.6
0.8
1.0
1.2
Templatedilutions
Fig. 8. Quantitation of TNF-c( message in LPS induced HL-60 cells. 3.41 pg of total HL-60 RNA and 8.8 x lo8 molecules of internal control RNA was mixed and serially diluted up to 1: 8. RT was done on all dilutions, one-fourth of the RT mix was then amplified using TNFcl primers, the 5’ primer being end-labelled with y ‘*P. After 25 cycles, the PCR products were resolved on a 2% agarose gel, the specific bands were cut out and radioactivity estimated.
Quantitation of cytokine mRNA by PCR
30000 r
1233
Rate of amplification
Template dilutions
Fig. 9. Quantitation of IL-2 message in PMA induced Jurkat cells. 3.42 pg of total Jurkat cell RNA and 8.8 x lo6 molecules of internal control RNA were mixed, and serially diluted up to 1: 64. RT was done on all dilutions, one-fourth of the RT mix was then amplified using IL-2 primers, the 5’ primer being end-labelled with y32P.After 25 cycles, the PCR products were resolved on a 2% agarose gel, the specific bands were cut out and radioactivity determined using a liquid scintillation counter.
RESULTS
Internal control The PCR is an exponential amplification and if the efficiency of amplification method is ideal, the total amount of target DNA formed following amplification could be calculated using the equation, Y: S x 2” where ‘Y’ is the amount of DNA after amplification, ‘S’ the starting number of target molecules, and the ‘n’ the number of cycles performed. However, in reality, the efficiency of amplification can deviate from ideal and hence the equation relating input and output amount of DNA must take into account the efficiency of the reaction. The methods originally described by Wang et al. (1989) and reported here, take into account this fact and have included an internal control RNA to overcome the problem of efficiency in quantitating various mRNA levels. The internal control was made by overlap extension PCR, the principle of which is shown in Fig. 1. The structure of the internal control and the in vitro transcription vector containing the internal control are shown in Figs 2 and 3, respectively. Since the quantitation was based on comparing the signal from an internal control with that from test RNA, initial experiments were done to verify the specificity of the internal control RNA and the various cytokine primers. Using T, RNA polymerase, RNA was transcribed from the synthetic gene vector SYNTA-2000. The RNA thus obtained was shown to be of free of any template DNA by doing PCR on the RNA preparation prior to reverse transcription. The specificity of the cytokine primers was tested by performing PCR using all the cytokine primers used to construct the internal control and the reverse transcribed product of the internal control RNA. As designed and expected, all primers gave -350 BP products and none gave any nonspecific products. An example of results obtained for IL-2, IL-6 and TNFcr is shown in Fig. 4. The results for the other cytokines were essentially identical (data not shown).
Having established the specificity of the internal control RNA and the cytokine primers, the next step was to determine the optimum number of PCR cycles required to keep the amplification in the exponential phase and also to amplify only the specific product. The results of experiments done using IL-6 primers are shown in Fig. 5a and b. A known amount of internal control RNA was reverse transcribed and amplified using IL-6 primers, the 5’ primers being end labelled with y3’P. The PCR products after IS, 20, 25, 30 and 35 cycles were analysed on a 2% agarose gel and the specific bands were excised to estimate their radioactivity. As shown in Fig. 5a, the optimum number of cycles fell between 25-35 cycles. The experiment was repeated using a different amount of template DNA in order to see if the starting template concentration could influence the exponential nature of the PCR. This experiment also showed that the optimum number of cycles fell between 25-35 cycles (Fig. 5b). Since it is advisable to keep the PCR cycles to the minimum number required in order to avoid nonspecific amplification, 25 cycles was chosen for all quantitation experiments. Quantitative analysis of IL-6, TNF-cr and IL-2 mRNA levels Using the internal control RNA described above, we have quantitated IL-6 and TNFa mRNA levels in LPS-stimulated THP-1 cells, TNF-cr mRNA level in LPS-stimulated HL-60 cells and IL-2 mRNA levels in PMA stimulated Jurkat cells. The strategy involved simultaneous reverse transcription of known amounts of internal control RNA, unknown amounts of target mRNA followed by simultaneous PCR on specified amounts of cDNA. The PCR products after 25 cycles were resolved on an agarose gel and the radioactivity of excised specific bands measured. The radioactivity of the target specific and the internal control specific bands were plotted against their respective template concentration. In the case of target RNA, the total RNA concentration values were used to construct the graph. The amount of target mRNA was then calculated from a standard line drawn using the internal control as shown in Figs 69. Using this quantitative approach, l.Opg of LPS stimulated THP-I cell RNA contained 2.4 x 10’ molecules of IL-6 mRNA (Fig. 6) and 1.7 x IO* molecules of TNF mRNA (Fig. 7). Similarly, 1.0 pg of LPS-induced HL-60 cell RNA was found to contain 3.4 x 10’ molecules of TNF mRNA (Fig. 8). Comparable values were obtained when the experiments were repeated with another batch of stimulated THP.1 and HL.60 cellular RNA (Table 2). PMA-stimulated Jurkat cells contained 2.8 x lo6 molecules of IL-2 mRNA per 1.Opg of total cellular RNA (Fig. 9). None of the three cell lines tested had the respective cytokine mRNA expressed in their unstimulated condition. The deduced number of molecules of various cytokine mRNA per stimulated single cell is shown in Table 2.
1234
S.KANANGAT
Table 2. No. of specific cytokine
Cell line
mRNA
THP. 1 THP. 1 HL.60 Jurkat
IL-6 TNF-u TNF-a IL-2
et al.
mRNA molecules and Jurkat cells
No. of molecules per 1.0 pg total RNA Batch I Batch II 2.4 1.7 3.3 2.8
x x x x
10’ 108 108 lo6
2.46 x 10’ 3.3 x 108 2.69 x 10’
in stimulated
THP-1,
HL-60
No. of molecules per single cell Batch 1 Batch II 2.4 1700 3300 28.00
2.46 3300 2690
All the cytokine messages tested were undetectable in the unstimulated condition of the respective cell line. For calculation/single cell, it is assumed that a single mammalian cell yield 10 pg of total cellular RNA under ideal conditions. The assay is theoretically capable of detecting even 1.O molecule of cytokine mRNA per single cell which would translate into 1 x lo5 molecules per 1.0 pg of total RNA assuming the recovery of RNA was ideal.
/?-Actin Based on a semiquantitative analysis, the levels of p-actin mRNA in stimulated and unstimulated cells were found to remain almost unchanged (Table 3). Sensitivity of the assay The sensitivity of the quantitative RT-PCR method described was shown by the minute amount of starting material required to detect the expression of cytokine genes. IL-6 mRNA could be detected in 62.5 ng of total THP-1 cellular RNA. This would approximately correspond to 6250 cells assuming that a normal mammalian cell would yield 10 pg total RNA per cell. TNF-a mRNA could be detected in 32 ng total THP-1 RNA (- 3200 cells), and 108 ng of total HL-60 RNA (- 10,800 cells). In the case of Jurkat cell lines, IL-2 mRNA could be detected in 13.5 ng of total cellular RNA which would approximately correspond to 1350 cells. Overall, it can be said that the method described here would allow one to work with a limited number of cells and also to quantitatively compare the cytokine gene expression at the mRNA level precisely. The amount of starting material required is considerably less than the microgram quantities required for other classical quantitative estimations such as Northern blotting, solution hybridization and dot blot hybridizations. Reproducibility of the assay Cytokine mRNA levels were estimated in two batches of stimulated THP.1 (IL-6 and TNF-a) and HL-60 (TNF-cr) cell lines. As shown in Table 4, the values obtained were comparable. Considering the sensitivity and possible variation in the assay, all the quantitative Table
3. Semiquantitation of p-actin stimulated and unstimulated Radioactivity
Source THP-1 HL-60 Jarkat
of RNA
Unstimulated 4653 4006 3251
mRNA cells
levels
(cpm) of /I-actin Stimulated 4705 4263 3321
band
experiments were done 2-3 times. There was no significant differences between the values obtained, for any of the cytokines tested, on the same day (Table 3). However, the values obtained on the first day of quantitation was always slightly higher than the subsequent days experiment. Since the same batch of RNA was used for the experiments, this difference possibly reflects instability of the mRNA. The inherent assay variability also should be considered for this minor differences observed. The fact that similar values were obtained on the same days’ experiment makes this method a reliable one to compare precisely the cytokine gene expression at the mRNA levels in various cell population under various conditions. DISCUSSION
The RT-PCR system has been proven to be an extremely powerful means of detecting and characterizing transcripts that are present in a wide array of sample material. As usually employed, the approach is only qualitative. In this report, we describe a method that permits precise quantitation of multiple specific mRNA species present in a small amount of total RNA. The approach which relies on simultaneous amplification of DNA from an internal standard and a test template, was designed to quantitate multiple cytokine mRNA species. We report the result of quantifying levels of mRNA of three cytokines, IL-2, IL-6 and TNFcr, following stimulation of cells with known inducers. PMA induced Jurkat cells contained 28 molecules of IL-2 mRNA per cell while LPS stimulated THP.1 cells had 2.4 molecules Table 4. Reproducibility
of the assay
No. of molecules
in Cell line
Cytokine
THP. 1
IL-6 TNF-a TNF-a IL-2
HL-60 Jurkat
Expt 2.4 1700 3300 28.00
1
of mRNA
single cell
Expt 2”
Expt 3“
2.52 1159 3370 11.00
2.4 1347 ND 11.6
“Experiments 2 and 3 were done on the same day. ND = Not Done.
Quantitation
of cytokine mRNA by PCR
of IL-6 mRNA and 1700 molecules of TNF-a mRNA molecules per single cells. And HL-60 cells activated with LPS contained 3300 molecules of TNF-amRNA/cell. In resting cells, insufficient mRNA was present to permit quantitation. Our approach promises to be a valuable means to measure more precisely, changes in expression of cytokine genes during various disease states and in response to a variety of modulating factors such as exposure of producer cells to viruses, Some of these kinds of investigations are underway in our laboratory. Others have adapted the PCR approach so that it can be used quantitatively, but most rely on comparing the amount of product generated by two separate reactions, one of which is the external standard. Since the PCR is usually non-linear and subject to many variables of the reaction mix, the external standard approach is not ideal. The approach we have used, which is a modification of a method reported by Wang et al. (1989) relies on an internal standard and deals with reverse transcription and amplification simultaneously from the templates using the same set of primers. This rules out the possibilities of any variations attributable to primer differences, tube differences, and variations in reaction conditions. However, our approach is still subject to some variables such as the possible differences in RT efficiency between target and internal control and also the difference in length of the target sequences and internal control sequences. [In our present system, however, the differences in length of targets and internal control would not affect the amplification efficiency considering the nucleotide incorporation capacity (> 60 nucleotides per second) of Taq DNA polymerase.] In spite of this possible drawback, the method described here could be considered as the most precise way of quantitating mRNA on a comparative level especially when one is dealing with very low abundance transcripts. Moreover, we found the assay to be very reproducible when used to quantitate mRNA on the same sample on different occasions. Among the various parameters that could affect the PCR amplification efficiency, primer efficiency is the most important. It is well illustrated by the following example. In the present study, during standardization experiments, it was found that TNFa primers were less efficient than IL-6 primers (data not shown). Similarly in the study by Wang et al. (1989) it was found that IL-/I primers were lo5 fold more efficient than the apo E primers. These observations indicate very clearly that it is critical to use the same primers for amplification of the target mRNA and the internal standard in any attempts to quantitate mRNA expression by PCR. In this study, we have also taken into account the fact that the amount of DNA amplified in a given sample could have a major influence on the amplification efficiency. When a high template concentration was used, phenomena such as substrate saturation of enzyme, inhibition of the amplification by the enzyme, incomplete strand separation and other effects could occur which would result in decreased amplification efficiency
1235
(Larzul et al., 1988). Bearing this in mind, we used varying amounts of template concentrations both for RT as well as PCR. The amount of target mRNA levels were calculated from various dilutions. In this preliminary study, we have shown that cytokine messages IL-2, IL-6 and TNF-a could be precisely quantitated in a small amount of total cellular RNA. Though theoretically it is possible to quantitate cytokine mRNA levels in a very small number of cells, the techniques available at present in the laboratory are not efficient enough to work with less than 10,000 cells. Currently we are attempting to improve our RNA isolation method to use smaller numbers of cells. It is to be emphasized that the internal control described here could be used to measure 13 different human cytokines quite precisely. An added advantage is that, the construct has provision to accommodate other primers of interest. In other words, the system could be adapted to many other situations, such as quantitating viral mRNA species in states of latency, autoimmunity and possibly neoplasia. Acknowledgements-This work was supported by the NIH grants EY05093, AI24762 and also by an Ina M. Berger Memorial grant for Cancer Research (IM-430) from the American Cancer Society and a grant from the Physician’s Medical Education and Research Foundation (University of Tennessee Medical Centre). We would like to thank Dr Carol Brenner for valuable scientific discussions and Audrey Williams for preparing this manuscript.
REFERENCES Brenner C. A., Tam A. W., Nelson P. A., Engleman E. E., Suzuki N., Fryke and Larrick (1989) Message Amplification Phenotyping (MAPPING): A technique to simultaneously measure multiple mRNAs from small numbers of cells. Biotechniques 7, 1096-1099. Chelly J., Kaplan J. C., Maire P., Gautron S. and Kahn A. (1988) Transcription of the dystrophin gene in human muscle and non-muscle tissues. Nature 333, 858-860. Chirgwin J. M., Przybyla A. E., McDonald R. J. and Rutter W. (1979) Isolation of biologically active ribonucleic acid from sources enriched in ribonuclease. Biochemistry 18, 5294-5299.
Dillon P. J. and Rosen C. A. (1990) A rapid method for the construction of synthetic genes using the polymerase chain reaction. Biotechniques 9, 298-300. Erlich H. A., Gelfand D. and Sninsky J. J. (1991) Recent advances in the Polymerase Chain Reaction. Nature 252, 1643-1650.
Gilliland G., Perrin S., Blanchard K. and Bunn F. (1990) Analysis of cytokine mRNA and DNA: Detection and quantitation by competitive polymerase chain reaction. Proc. natn. Acad. Sci. 87, 2725-2729.
Ho S. N., Hund H. D., Horton R. M., Pullen J. K. and Pease L. R. (1989) Site-directed mutagenesis by overlap extension using the polymerase chain reaction. Gene 77, 51-59. Horn G. T., Bugwan T. L., Long C. M. and Erlich H. A. (1988) Allelic sequence variation of the HLA. DQ locus: Relationship to serology and to insulin dependent diabetes susceptibility. Proc. mtn. Acad. Sci. 85, 6012-6016.
1236
S.
KANANGAT
Larzul D., Guigue F., Sninsky J. J., Mack D. H., Brechot C. and Guesdon J. L. (1988) Detection of hepatitis B virus sequences in serum by using in vitro enzymatic amplification. J. Virol. Meth. 20, 221-237. McQuairre D. A. and Rock P. A., Eds (1984) In General Chemistry, Ch. 3, Chemical Calculations, pp. 85-113. W. H. Freeman and Company, New York. Simmonds P., Balfe P., Peulhevev J. F., Ludlam C. A., Bishop J. 0. and Brown A. J. (1990) Human immunodeficiency
et al. virus-infected individuals contain provirus in small numbers of peripheral mononuclear cells and at low copy numbers. J. Virol. 64, 864-872. Todd J. A., Bell J. I. and McDevitt H. (1987) HLA. DQ13 gene contributes to susceptibility and resistance to insulin dependent diabetes mellitus. Nature 329, 599-604. Wang A. M., Doyle M. V. and Mark D. F. (1989) Quantitation of mRNA by the polymerase chain reaction. Proc. natn. Acad. Sci. 86, 9717-9721.
0161-5890/92$5.00 + 0.00 Pergamon Press Ltd
USE OF QUANTITATIVE POLYMERASE CHAIN REACTION TO QUANTITATE CYTOKINE MESSENGER RNA MOLECULES WADASANKANANGAT,*ALAN SOLOMON~ and BARRYT. ROUSERS *Department of Microbiology, College of Veterinary Medicine, University of Tennessee, Knoxville, TN 37996-0845,U.S.A. and TDepartment of Medicine, Human Immunology and Cancer Programme, University of Tennessee, Medical Centre/Graduate School of Medicine, Knoxville, TN 37996-0845, U.S.A. (First received 31 October 1991; accepted in revisedform
12 March 1992)
quantitative polymerase chain reaction (PCR) using an internal control (Standard) RNA was developed for precise quantitation of human cytokine mRNA. The target mRNA and internal control were simultaneously reverse transcribed and co-amplified using the same set of primers. The amount of specific target mRNA is then quantitated by extrapolating against the standard curve generated with the internal standard. The internal control RNA consisted of linearly connected sequences of the 5’ primers of multiple cytokine genes followed by the complementary sequences to their 3’ primers in the same order. This structure of the internal standard enables one to use the same internal standard for quantitating multiple cytokine mRNAs. Using this approach, we estimated the induced levels of cytokine mRNA, IL2, IL6, and TNF-a to be 2.8 x 106,2.4 x lo5 and 3.4 x 1O*molecules per 1.Op g total cellular RNA of Jurkat, THP- 1 and HL 60 cells, respectively.
Abstract-A
Comparable values were obtained when quantitation experiments were done on another batch of THP-1 and HL-60 total cellular RNA. The excellent sensitivity and reproducibility makes this approach a valuable one in following changes in cytokine gene expression in wide variety of conditions, both in vivo and in vitro.
INTRODUCTION
The polymerase chain reaction (PCR) has found extensive use and has proven to be the technique of choice for the detection and analysis of minute amounts of DNA (reviewed by Erlich et al., 1991). Moreover, by employing a reverse transcription step, PCR is also a valuable means of detecting and characterizing low abundance transcripts (Chelly et al., 1988). As usually employed, the PCR approach is a qualitative procedure. However, there is clearly a need to develop a PCR strategy that will quantitate RNA transcripts since this approach should be more sensitive and easier to use than Northern blotting and solution hybridization. Various quantitative procedures have been reported, but most rely on the use of external standards and compare products generated in two separate reactions (Simmonds et al., 1990). Since both RT and DNA amplification may be nonlinear and subject to great variability, such approaches are extremely unreliable. Others have reported quantitative procedures based on a competitive reaction between an added competitor DNA fragment that differs from the target cDNA by having either a small intron or a mutated restriction enzyme site (Gilliland et al., 1990). However, this approach does not take into account variations in RT efficiency. A better approach for mRNA quantitation by PCR would be to employ an internal standard so that reverse transcription and coamplification of the RT products could be performed simultaneously. Detection of the SAuthor to whom correspondence should be addressed.
amount of unknown product generated using the same set of primers can then be achieved by designing the system so that test and standard products are of disparate size. Such an approach was described by Wang et al. (1989), and we have modified their system to quantitate multiple human cytokines mRNA in small quantities of sample RNA. The synthetic gene constructed could be used to quantitate 13 different human cytokines and the construct still has provisions to include more primer sequences of interest. In this report, we document the value of the approach and document induced levels of the three cytokines IL-2, IL-6, and TNF-o! in three cell lines. The approach we have described should prove valuable to follow changes in cytokine gene expression in a variety of disease situations and experimental manipulations in vitro. MATERIALS AND METHODS Oligonucleotide
preparation
Oligonucleotides were synthesized on an Applied Biosystem DNA synthesizer (Model 391). The 5’ and 3’ primer sequences (Table 1) were derived from published reports (Brenner et al., 1989). Construction of synthetic internal control RNA
gene and the preparation
of
A synthetic gene comprising the 5’ primers and 3’ primers was constructed by the technique of oligonucleotide overlap extension and subsequent amplification by PCR (Ho et al., 1989; Dillon and Rosen, 1990).
1229
S. KANANC;ATet al.
1230 Table 1. Human the construction
cytokine primers used for of the internal standard
Primers
Expected size of the target Number of base pairs
IL-l-a IL-l+ IL-2 IL-3 IL-4 IL-5 IL-6 IL-7 IL-8 IFN-ITNF-a TNF-fi TGF-fi PActin
816 810 462 459 462 414 639 707 302 501 702 608 200 548
The internal control was designed in such a way that -350 bp bands were obtained with each of the above mentioned primer pairs.
purification system. The purified RNA product was quantitated by spectrophotometric analysis at 260 nm.
CELL LINES
THP-1
THP-1, a human monocytic leukemia cell line was obtained from ATCC (No. TlB 202), and was maintained in RPM1 1640 containing 15% fetal bovine serum and 2 x lop5 2ME. These cells have Fc and C3b receptors and lack surface and cytoplasmic immunoglobulins. These cells stain positively for a-naphthyl-butyrate esterase, produce lysozyme and are phagocytic. HL -60 HL-60 is a human promyelocytic cell line obtained fromATCC (CCL 240) and was maintained in DMEM containing 15% FCS. These cells lack specific markers for lymphoid cells but express surface receptors for Fc fragments and complement. JURKA T
The strategy for the construction of the synthetic gene is shown in Fig. 1 and the structure in Fig. 2. The synthetic gene was cloned into an in vitro transcription vector/ pCR 2000 TA and was designated as SYNTA-2000 (Fig. 3). The linearized SYNTA-2000 was then used as a template to transcribe the internal control RNA using T, RNA polymerase employing a standard in vitro transcription protocol (Promega). The resulting RNA preparation was treated extensively with RNAse free DNAse (RQl DNAse), extracted with phenol: chloroform, ethanol precipitated or purified by a RNAID
PCR
Jurkat cell line is a human lymphoma derived T cell line. The culture was obtained from ATCC (CRL 8163) and maintained in RPM1 1640 with 15% FCS. Mitogen
stimulation
Jurkat cells were stimulated at lo6 cells ml--’ with phorbol myristate acetate (PMA) at a concentration of 10 ng ml-‘. THP-1 and HL-60 cells were activated at lo6 cells ml-’ with E. coli lipopolysaccharide (LPS) at a concentration of 5 p g ml ‘. All cells were harvested after 8 hr.
- 1
4 PRIMER - 2
* PRIMER
- 1
Fig. 1. Synthesis of internal standard by overlap extension PCR. 0.5 pg each of the long oligonucleotides (_ 100 bases) with overlapping ends (indicated by the broken lines) were mixed together in standard PCR (100 ~1 volume). The reaction was denatured at 94°C for 7 min, annealed at 55°C for 2 min and extended at 72°C for 3 min. The reactions went through seven additional cycles with the denaturation step reduced to 1.5 min. One microlitre of this reaction mix was used as template for second PCR. In the second PCR, 2 primers (21 bases each) spanning the 5’ and 3’ ends of the template formed in PCR I, was added to the standard PCR mix and subjected to 25 cycles of amplification (94°C 1.5 min; annealing at 55°C for 2 min; extension at 72°C for 3 min).
Quantitation
1231
of cytokine mRNA by PCR T 7 Eco RI
Not I Fig. 3. The synthetic gene produced by the overlap extension PCR was purified by gene clean and cloned into TA cloning vector which contains T7 promoter sequence.
45min. At the end of RT, the mixture was heated at 99°C for 5 min to inactivate the RT enzyme and denature the cDNA. PCR IL- I IL- 2 IL-
3
ILILILIL-
4 5 6 7
IL- x IIJN- r ‘rNI’_ a TNF- b TGV- b b- ACTIN I’OLY IllNIl
A III
Fig. 2. Structure of SYNTA-2000. The plasmid contains 5’ primers of 14 target genes connected in sequence followed by the complementary sequence of the 3’ primers in the same order. The multiple primer region is flanked by the T7 polymerase promoter upstream and by poly A downstream.
One-fourth of the RT mix was used as template for the amplification reaction. All PCRs were done in a 25 ~1 volume. The mixture contained 1X PCR buffer (10 mM Tris-HCl pH 8.3; 50 mM KCl, 1.5 mM MgCI,, 0.01% Triton X-100), 1.25 mM dNTP, 1.OU Taq DNA polymerase (Promega), 5 pmoles each of 5’ and 3’ primers, the 5’ primer being end labelled with y3’P using T, polynucleotide kinase. The unincorporated radionucleotides were removed from the probe by passing it through a Sephadex G-25 spun column (Pharmacia). The PCR mixture was overlaid with 50 ~1 mineral oil (Sigma), and incubated in a Perkin-Elmer/Cetus thermal cycler. The amplification steps involved denaturation at 94°C for l-2 min, annealing at 55°C for 2 min and extension at 72°C for 3 min. Appropriate positive (internal control and relevant primers) and negative (PCR mix with relevant primers but no template) controls were included with each batch of reactions.
Total cellular RNA
Total cellular RNA from induced and uninduced cells was isolated by the method of guanidium thiocyanate, phenol: chloroform extraction (Chirgwin et al., 1979). Reverse transcription
The internal control RNA and the target RNA (total cellular RNA) were reverse transcribed simultaneously in the same tube according to a standard protocol (Promega). Varying amounts of total cellular RNA (4.0 ,ug-+O.125 pg) and the internal control RNA, SYNTA-PCR 2000 transcribed in vitro from (8.8 x 10*+0.53 x 10’ molecules),* were reverse transcribed in a 20 ~1 reaction mix containing 1X RT buffer (10 mM Tris-HCl pH 8.8, 50 mM KCl; 0.1% Triton X-100), 5 mM MgCl,, 2 mM dNTP, 40 U RNAs inhibitor, oligo dT primer (0.5-1.5 ,ug) and AMV RT (10-20 units). The RT reaction was performed at room temperature for 30 min followed by incubation at 42°C for *No. of molecules were calculated using Avaqadro’s hypothesis (McQuarrie and Rock, 1984).
Fig. 4. Specificity of the internal control RNA and the Primers. The internal control RNA was reverse transcribed and amplified [25 cycles] using IL-2, IL-6, TNF-a primers in separate reactions in duplicate tubes. The products were resolved on a 2% agarose gel. Lanes b, c: IL-2; d, e: IL-6; f, g: TNFcr; h: Positive control. SYNTA-200.IL-6. Lane a is the 1 Kb DNA Ladder (Bethesda Research Laboratories).
S. KANANGAT et al.
1232
A -5
50000
'2
40000
$ cd
30000 20000 10000
0.0
0.2
0.4
0.6
n
Target
0
ht. Control
0.8
1.0
12
Templatedilutions Fig. 7. Quantitation of TNF-a-message in LPS-induced THP-1 cells. Four microgrammes of total THP-1 RNA and 8.8 x lOa molecules of internal control RNA were mixed and serially diluted as shown. RT was done on all dilutions, one-fourth of the RT mix was then amplified using TNF-cr primers, the 5’ primer being end-labelled with y”P. After 25 cycles, the PCR products were resolved on a 2% agarose gel and band radioactivity was measured.
No. of Cycles
@I
120000
-
.2: +. 100000
-
., z8
80000
-
._ 'D 2
60000
-
40000-
20000 1
0
! 10
I No.
Fig. 5. Relation amplification. 4.3 control RNA was mix was amplified
30
20
40
of Cycles
between no. of PCR cycles and rate of x lo* (a)/2.15 x 10’ (b) molecules of internal reverse transcribed and one-fourth of the RT using IL-6 primers, the 5’ primer being end labelled with y3’P.
Quantitative analysis Ten microlitres of each PCR mixture was electrophoresed on an agarose gel (2%). Gels were stained with ethidium bromide and photographed. The specific bands were excised from the agarose gel and the radioactivity was measured. The amount of radioactivity obtained
0
0
from excised gel bands were plotted against the template concentrations or the number of PCR cycles. The straight line graphs were drawn using the radioactivity counts of internal standard bands and target bands against their respective template concentrations and using the internal control graph, the specific quantity of target mRNA was calculated as shown in Figs 6-9. The best fit lines were drawn using the computer program C Cricket Graph 1-3.
Semiquantitation of /3-actin mRNA levels in unstimulated and stimulated cells Total RNA (2.0 pg each) obtained from unstimulated and stimulated THP-1 Jurkat and HL-60 cells were reverse transcribed in a 20 ~1 reaction mix each. Onefourth of each reaction mix was subjected to 25 cycles of PCR using b-actin primers, 5’ Primer being end-labelled with y32P. Ten microliters of the PCR product was analysed on a 2% agarose gel and stained with ethidium bromide. The /3-actin specific 548 base pair bands were excised and the radioactivity measured using a scintillation counter.
Target Ini.control -I
0:o
0:2
0.4
0.6
0.8
1.0
1.2
Templatedilutions
Fig. 6. Quantitation of IL-6 message in LPS-induced THP-1 cells. 4.0 pg of total THP-1 (LPS-induced) RNA, 8.6 x 10’ molecules of internal control RNA were mixed and serial twofold dilutions of the RNA mix were made. RT was done on all of them and one-fourth of each RT mix was amplified using IL-6 primers, the 5’ primer being end-labelled with y32P. The PCR products were resolved on a 2% agarose gel and the specific band radioactivity measured.
0.0
0.2
0.4
0.6
0.8
1.0
1.2
Templatedilutions
Fig. 8. Quantitation of TNF-c( message in LPS induced HL-60 cells. 3.41 pg of total HL-60 RNA and 8.8 x lo8 molecules of internal control RNA was mixed and serially diluted up to 1: 8. RT was done on all dilutions, one-fourth of the RT mix was then amplified using TNFcl primers, the 5’ primer being end-labelled with y ‘*P. After 25 cycles, the PCR products were resolved on a 2% agarose gel, the specific bands were cut out and radioactivity estimated.
Quantitation of cytokine mRNA by PCR
30000 r
1233
Rate of amplification
Template dilutions
Fig. 9. Quantitation of IL-2 message in PMA induced Jurkat cells. 3.42 pg of total Jurkat cell RNA and 8.8 x lo6 molecules of internal control RNA were mixed, and serially diluted up to 1: 64. RT was done on all dilutions, one-fourth of the RT mix was then amplified using IL-2 primers, the 5’ primer being end-labelled with y32P.After 25 cycles, the PCR products were resolved on a 2% agarose gel, the specific bands were cut out and radioactivity determined using a liquid scintillation counter.
RESULTS
Internal control The PCR is an exponential amplification and if the efficiency of amplification method is ideal, the total amount of target DNA formed following amplification could be calculated using the equation, Y: S x 2” where ‘Y’ is the amount of DNA after amplification, ‘S’ the starting number of target molecules, and the ‘n’ the number of cycles performed. However, in reality, the efficiency of amplification can deviate from ideal and hence the equation relating input and output amount of DNA must take into account the efficiency of the reaction. The methods originally described by Wang et al. (1989) and reported here, take into account this fact and have included an internal control RNA to overcome the problem of efficiency in quantitating various mRNA levels. The internal control was made by overlap extension PCR, the principle of which is shown in Fig. 1. The structure of the internal control and the in vitro transcription vector containing the internal control are shown in Figs 2 and 3, respectively. Since the quantitation was based on comparing the signal from an internal control with that from test RNA, initial experiments were done to verify the specificity of the internal control RNA and the various cytokine primers. Using T, RNA polymerase, RNA was transcribed from the synthetic gene vector SYNTA-2000. The RNA thus obtained was shown to be of free of any template DNA by doing PCR on the RNA preparation prior to reverse transcription. The specificity of the cytokine primers was tested by performing PCR using all the cytokine primers used to construct the internal control and the reverse transcribed product of the internal control RNA. As designed and expected, all primers gave -350 BP products and none gave any nonspecific products. An example of results obtained for IL-2, IL-6 and TNFcr is shown in Fig. 4. The results for the other cytokines were essentially identical (data not shown).
Having established the specificity of the internal control RNA and the cytokine primers, the next step was to determine the optimum number of PCR cycles required to keep the amplification in the exponential phase and also to amplify only the specific product. The results of experiments done using IL-6 primers are shown in Fig. 5a and b. A known amount of internal control RNA was reverse transcribed and amplified using IL-6 primers, the 5’ primers being end labelled with y3’P. The PCR products after IS, 20, 25, 30 and 35 cycles were analysed on a 2% agarose gel and the specific bands were excised to estimate their radioactivity. As shown in Fig. 5a, the optimum number of cycles fell between 25-35 cycles. The experiment was repeated using a different amount of template DNA in order to see if the starting template concentration could influence the exponential nature of the PCR. This experiment also showed that the optimum number of cycles fell between 25-35 cycles (Fig. 5b). Since it is advisable to keep the PCR cycles to the minimum number required in order to avoid nonspecific amplification, 25 cycles was chosen for all quantitation experiments. Quantitative analysis of IL-6, TNF-cr and IL-2 mRNA levels Using the internal control RNA described above, we have quantitated IL-6 and TNFa mRNA levels in LPS-stimulated THP-1 cells, TNF-cr mRNA level in LPS-stimulated HL-60 cells and IL-2 mRNA levels in PMA stimulated Jurkat cells. The strategy involved simultaneous reverse transcription of known amounts of internal control RNA, unknown amounts of target mRNA followed by simultaneous PCR on specified amounts of cDNA. The PCR products after 25 cycles were resolved on an agarose gel and the radioactivity of excised specific bands measured. The radioactivity of the target specific and the internal control specific bands were plotted against their respective template concentration. In the case of target RNA, the total RNA concentration values were used to construct the graph. The amount of target mRNA was then calculated from a standard line drawn using the internal control as shown in Figs 69. Using this quantitative approach, l.Opg of LPS stimulated THP-I cell RNA contained 2.4 x 10’ molecules of IL-6 mRNA (Fig. 6) and 1.7 x IO* molecules of TNF mRNA (Fig. 7). Similarly, 1.0 pg of LPS-induced HL-60 cell RNA was found to contain 3.4 x 10’ molecules of TNF mRNA (Fig. 8). Comparable values were obtained when the experiments were repeated with another batch of stimulated THP.1 and HL.60 cellular RNA (Table 2). PMA-stimulated Jurkat cells contained 2.8 x lo6 molecules of IL-2 mRNA per 1.Opg of total cellular RNA (Fig. 9). None of the three cell lines tested had the respective cytokine mRNA expressed in their unstimulated condition. The deduced number of molecules of various cytokine mRNA per stimulated single cell is shown in Table 2.
1234
S.KANANGAT
Table 2. No. of specific cytokine
Cell line
mRNA
THP. 1 THP. 1 HL.60 Jurkat
IL-6 TNF-u TNF-a IL-2
et al.
mRNA molecules and Jurkat cells
No. of molecules per 1.0 pg total RNA Batch I Batch II 2.4 1.7 3.3 2.8
x x x x
10’ 108 108 lo6
2.46 x 10’ 3.3 x 108 2.69 x 10’
in stimulated
THP-1,
HL-60
No. of molecules per single cell Batch 1 Batch II 2.4 1700 3300 28.00
2.46 3300 2690
All the cytokine messages tested were undetectable in the unstimulated condition of the respective cell line. For calculation/single cell, it is assumed that a single mammalian cell yield 10 pg of total cellular RNA under ideal conditions. The assay is theoretically capable of detecting even 1.O molecule of cytokine mRNA per single cell which would translate into 1 x lo5 molecules per 1.0 pg of total RNA assuming the recovery of RNA was ideal.
/?-Actin Based on a semiquantitative analysis, the levels of p-actin mRNA in stimulated and unstimulated cells were found to remain almost unchanged (Table 3). Sensitivity of the assay The sensitivity of the quantitative RT-PCR method described was shown by the minute amount of starting material required to detect the expression of cytokine genes. IL-6 mRNA could be detected in 62.5 ng of total THP-1 cellular RNA. This would approximately correspond to 6250 cells assuming that a normal mammalian cell would yield 10 pg total RNA per cell. TNF-a mRNA could be detected in 32 ng total THP-1 RNA (- 3200 cells), and 108 ng of total HL-60 RNA (- 10,800 cells). In the case of Jurkat cell lines, IL-2 mRNA could be detected in 13.5 ng of total cellular RNA which would approximately correspond to 1350 cells. Overall, it can be said that the method described here would allow one to work with a limited number of cells and also to quantitatively compare the cytokine gene expression at the mRNA level precisely. The amount of starting material required is considerably less than the microgram quantities required for other classical quantitative estimations such as Northern blotting, solution hybridization and dot blot hybridizations. Reproducibility of the assay Cytokine mRNA levels were estimated in two batches of stimulated THP.1 (IL-6 and TNF-a) and HL-60 (TNF-cr) cell lines. As shown in Table 4, the values obtained were comparable. Considering the sensitivity and possible variation in the assay, all the quantitative Table
3. Semiquantitation of p-actin stimulated and unstimulated Radioactivity
Source THP-1 HL-60 Jarkat
of RNA
Unstimulated 4653 4006 3251
mRNA cells
levels
(cpm) of /I-actin Stimulated 4705 4263 3321
band
experiments were done 2-3 times. There was no significant differences between the values obtained, for any of the cytokines tested, on the same day (Table 3). However, the values obtained on the first day of quantitation was always slightly higher than the subsequent days experiment. Since the same batch of RNA was used for the experiments, this difference possibly reflects instability of the mRNA. The inherent assay variability also should be considered for this minor differences observed. The fact that similar values were obtained on the same days’ experiment makes this method a reliable one to compare precisely the cytokine gene expression at the mRNA levels in various cell population under various conditions. DISCUSSION
The RT-PCR system has been proven to be an extremely powerful means of detecting and characterizing transcripts that are present in a wide array of sample material. As usually employed, the approach is only qualitative. In this report, we describe a method that permits precise quantitation of multiple specific mRNA species present in a small amount of total RNA. The approach which relies on simultaneous amplification of DNA from an internal standard and a test template, was designed to quantitate multiple cytokine mRNA species. We report the result of quantifying levels of mRNA of three cytokines, IL-2, IL-6 and TNFcr, following stimulation of cells with known inducers. PMA induced Jurkat cells contained 28 molecules of IL-2 mRNA per cell while LPS stimulated THP.1 cells had 2.4 molecules Table 4. Reproducibility
of the assay
No. of molecules
in Cell line
Cytokine
THP. 1
IL-6 TNF-a TNF-a IL-2
HL-60 Jurkat
Expt 2.4 1700 3300 28.00
1
of mRNA
single cell
Expt 2”
Expt 3“
2.52 1159 3370 11.00
2.4 1347 ND 11.6
“Experiments 2 and 3 were done on the same day. ND = Not Done.
Quantitation
of cytokine mRNA by PCR
of IL-6 mRNA and 1700 molecules of TNF-a mRNA molecules per single cells. And HL-60 cells activated with LPS contained 3300 molecules of TNF-amRNA/cell. In resting cells, insufficient mRNA was present to permit quantitation. Our approach promises to be a valuable means to measure more precisely, changes in expression of cytokine genes during various disease states and in response to a variety of modulating factors such as exposure of producer cells to viruses, Some of these kinds of investigations are underway in our laboratory. Others have adapted the PCR approach so that it can be used quantitatively, but most rely on comparing the amount of product generated by two separate reactions, one of which is the external standard. Since the PCR is usually non-linear and subject to many variables of the reaction mix, the external standard approach is not ideal. The approach we have used, which is a modification of a method reported by Wang et al. (1989) relies on an internal standard and deals with reverse transcription and amplification simultaneously from the templates using the same set of primers. This rules out the possibilities of any variations attributable to primer differences, tube differences, and variations in reaction conditions. However, our approach is still subject to some variables such as the possible differences in RT efficiency between target and internal control and also the difference in length of the target sequences and internal control sequences. [In our present system, however, the differences in length of targets and internal control would not affect the amplification efficiency considering the nucleotide incorporation capacity (> 60 nucleotides per second) of Taq DNA polymerase.] In spite of this possible drawback, the method described here could be considered as the most precise way of quantitating mRNA on a comparative level especially when one is dealing with very low abundance transcripts. Moreover, we found the assay to be very reproducible when used to quantitate mRNA on the same sample on different occasions. Among the various parameters that could affect the PCR amplification efficiency, primer efficiency is the most important. It is well illustrated by the following example. In the present study, during standardization experiments, it was found that TNFa primers were less efficient than IL-6 primers (data not shown). Similarly in the study by Wang et al. (1989) it was found that IL-/I primers were lo5 fold more efficient than the apo E primers. These observations indicate very clearly that it is critical to use the same primers for amplification of the target mRNA and the internal standard in any attempts to quantitate mRNA expression by PCR. In this study, we have also taken into account the fact that the amount of DNA amplified in a given sample could have a major influence on the amplification efficiency. When a high template concentration was used, phenomena such as substrate saturation of enzyme, inhibition of the amplification by the enzyme, incomplete strand separation and other effects could occur which would result in decreased amplification efficiency
1235
(Larzul et al., 1988). Bearing this in mind, we used varying amounts of template concentrations both for RT as well as PCR. The amount of target mRNA levels were calculated from various dilutions. In this preliminary study, we have shown that cytokine messages IL-2, IL-6 and TNF-a could be precisely quantitated in a small amount of total cellular RNA. Though theoretically it is possible to quantitate cytokine mRNA levels in a very small number of cells, the techniques available at present in the laboratory are not efficient enough to work with less than 10,000 cells. Currently we are attempting to improve our RNA isolation method to use smaller numbers of cells. It is to be emphasized that the internal control described here could be used to measure 13 different human cytokines quite precisely. An added advantage is that, the construct has provision to accommodate other primers of interest. In other words, the system could be adapted to many other situations, such as quantitating viral mRNA species in states of latency, autoimmunity and possibly neoplasia. Acknowledgements-This work was supported by the NIH grants EY05093, AI24762 and also by an Ina M. Berger Memorial grant for Cancer Research (IM-430) from the American Cancer Society and a grant from the Physician’s Medical Education and Research Foundation (University of Tennessee Medical Centre). We would like to thank Dr Carol Brenner for valuable scientific discussions and Audrey Williams for preparing this manuscript.
REFERENCES Brenner C. A., Tam A. W., Nelson P. A., Engleman E. E., Suzuki N., Fryke and Larrick (1989) Message Amplification Phenotyping (MAPPING): A technique to simultaneously measure multiple mRNAs from small numbers of cells. Biotechniques 7, 1096-1099. Chelly J., Kaplan J. C., Maire P., Gautron S. and Kahn A. (1988) Transcription of the dystrophin gene in human muscle and non-muscle tissues. Nature 333, 858-860. Chirgwin J. M., Przybyla A. E., McDonald R. J. and Rutter W. (1979) Isolation of biologically active ribonucleic acid from sources enriched in ribonuclease. Biochemistry 18, 5294-5299.
Dillon P. J. and Rosen C. A. (1990) A rapid method for the construction of synthetic genes using the polymerase chain reaction. Biotechniques 9, 298-300. Erlich H. A., Gelfand D. and Sninsky J. J. (1991) Recent advances in the Polymerase Chain Reaction. Nature 252, 1643-1650.
Gilliland G., Perrin S., Blanchard K. and Bunn F. (1990) Analysis of cytokine mRNA and DNA: Detection and quantitation by competitive polymerase chain reaction. Proc. natn. Acad. Sci. 87, 2725-2729.
Ho S. N., Hund H. D., Horton R. M., Pullen J. K. and Pease L. R. (1989) Site-directed mutagenesis by overlap extension using the polymerase chain reaction. Gene 77, 51-59. Horn G. T., Bugwan T. L., Long C. M. and Erlich H. A. (1988) Allelic sequence variation of the HLA. DQ locus: Relationship to serology and to insulin dependent diabetes susceptibility. Proc. mtn. Acad. Sci. 85, 6012-6016.
1236
S.
KANANGAT
Larzul D., Guigue F., Sninsky J. J., Mack D. H., Brechot C. and Guesdon J. L. (1988) Detection of hepatitis B virus sequences in serum by using in vitro enzymatic amplification. J. Virol. Meth. 20, 221-237. McQuairre D. A. and Rock P. A., Eds (1984) In General Chemistry, Ch. 3, Chemical Calculations, pp. 85-113. W. H. Freeman and Company, New York. Simmonds P., Balfe P., Peulhevev J. F., Ludlam C. A., Bishop J. 0. and Brown A. J. (1990) Human immunodeficiency
et al. virus-infected individuals contain provirus in small numbers of peripheral mononuclear cells and at low copy numbers. J. Virol. 64, 864-872. Todd J. A., Bell J. I. and McDevitt H. (1987) HLA. DQ13 gene contributes to susceptibility and resistance to insulin dependent diabetes mellitus. Nature 329, 599-604. Wang A. M., Doyle M. V. and Mark D. F. (1989) Quantitation of mRNA by the polymerase chain reaction. Proc. natn. Acad. Sci. 86, 9717-9721.
