Search for DNA of exogenous mouse mammary tumor virus-related virus in human breast cancer samples

Mouse mammary tumor virus (MMTV), in exogenous or endogenous form, is a major cause of mammary tumours in mice (Bittner, 1936). The presence of DNA from an alleged closely related retrovirus, human exogenous retrovirus (HMTV), and a retrovirus almost identical to HMTV, has been reported in human breast cancer and in primary biliary cirrhosis (Pogo & Holland, 1997; Holland & Pogo, 2004; Mant & Cason, 2004; Xu et al., 2004). Several reports indicate that HMTV exists, is associated preferentially with breast cancer and is not present in normal breast tissue. For example, Ford et al. (2004a) detected highly MMTV-like sequences, probably not of endogenous origin, in cancerous epithelial cells of female and male breast tumours and not in adjacent, non-malignant cells. This finding should not be confused with that of Yin et al. (1997), where high RNA expression of endogenous human MMTV-like sequences (HMLs), related more distantly to MMTV (Andersson et al., 1996), in one of 60 breast cancers was reported. Needless to say, the identification of an exogenous retrovirus causative of human breast cancer would have great clinical implications. To verify the HMTV reports, a prospective study using Swedish breast cancer patients and controls was performed. A real-time PCR, based on env of MMTV and the non-endogenous highly MMTV-like sequences reported from humans (here collectively called HMTV), was designed. It can detect one to ten copies of MMTV DNA per PCR. As in many other real-time PCRs, the threshold cycle number is correlated linearly with the logarithm of template DNA concentration from 10¹ to 10⁶ copies per reaction. Patients and samples.
During 2003 and 2004, at Uppsala University Hospital, Sweden, 18 consecutive patients with a palpable breast cancer of 2 cm or larger were included. Eleven control women with a diagnosis, obtained before surgery, of a benign breast condition were included during the same period. Tumours were graded (Wärnberg et al., 2001) as shown in Table 1. After consent, a blood sample was collected and, during surgery, the specimen was sent on ice to the Department of Pathology. Within 30 min, a piece of the tumour and normal breast parenchyma was cut out and stored at 70 °C. Further patient and tumour data are shown in Table 1. Thus, samples from both the tumour and control normal glandular tissue of the same breast, as well as plasma and white blood cells (WBCs), were taken from each patient. The control persons provided samples of the same kind, with the exception of a tumour sample. The study was approved by the Ethical Committee of the Faculty of Medicine of the University of Uppsala (permit no. 01/381).

Table 1. Clinical data for breast cancer patients and controls

Nucleic acid extraction.
DNA was extracted from 20 mg breast cancer tissue and normal breast tissue from the cancer patients and controls by using a Qiagen RNADNA mini kit (catalogue no. 14123; Qiagen). DNA concentrations varied from 15 to 115 ng µl¹ in the extracts from patient samples. Eight millilitres of citrate-treated blood was collected in CPT tubes (catalogue no. 362761; Becton Dickinson). Plasma and WBCs were separated as described by the manufacturer. RNA was extracted from 140 µl plasma by using a QIAamp Viral RNA mini kit (catalogue no. 52904; Qiagen) from 18 breast cancer patients and 11 non-malignant control samples, and DNA was extracted from 140 µl WBC preparation from same samples by using QIAmp columns (catalogue no. 51304; Qiagen). cDNA was synthesized from RNA extracted from plasma and breast tissue of breast cancer patients and control persons as described below.

DNA from WBCs of 100 Swedish blood donors was also prepared. The DNA concentration of each sample was determined in a DyNA Quant fluorimeter (Amersham Biosciences).

cDNA synthesis.
cDNA was synthesized by using 5 µl 1x StrataScript buffer, 5 µl 0.1 M dithiothreitol, 0.8 µl 100 mM dNTP mix (Sigma), 2 µl random hexamers (265 ng µl¹; Pharmacia), 29.2 µl RNase-free H₂O, 2 µl RNasin (40 U µl¹; Promega), 1 µl StrataScript reverse transcriptase (50 U µl¹) and 5 µl RNA (approx. 1 µg µl¹), at 25 °C for 10 min, 42 °C for 60 min and 90 °C for 5 min. Each cDNA reaction was run with and without reverse transcriptase.

Primers and probe for the real-time PCR method.
The primers and the probe were derived from the env region of published MMTV and the highly MMTV-like HMTV sequences reported from humans (GenBank accession nos AF346816[GenBank] , AF239172[GenBank] and M15122[GenBank] ) (Fig. 1). The primers were as follows: ABL (5'-TAGTTCCCCATACAGAATTGTTTCGCT-3') and ABR (5'-TCATCACCAATATCTACAGGTAGCAGCAC-3'), and ABL1 (5'-TAGTCCCCCATACAGAATTGTTTCGCT-3') and ABR1 (5'-TCATCACCAATATCTACAGGTAGCAGTGAC-3'); and the probe, ABP [5'-6-FAM*ACTATGATCGCT*(TAMRA)GCATAGTCGTAGGCAGAAGAATCT-phosphate-3']. Positions of mismatch between MMTV and HMTV are underlined. Primers ABL and ABL1, and ABR and ABR1, respectively, were tested singly (i.e. ABL versus ABR and ABL1 versus ABR1, in separate tubes) and as a 1+1 mixture (ABL+ABL1 versus ABR+ABR1 in the same tube). All primers and probe were purchased from Thermo Hybaid. They were used at 125, 125 and 300 nM final concentrations, respectively, in a TaqMan universal master mix with uracil N-glycosylase (UNG; Applied Biosystems). The reaction volume was 50 µl, with 2 µl DNA sample. Real-time PCR was performed in a RotorGene thermocycler (Corbett Research) with a temperature profile of 50 °C for 600 s and 95 °C for 600 s, followed by 60 repeats of 95 °C for 25 s and 60 °C for 60 s. Fluorescence was measured during the latter period at 60 °C. The ability of the PCR to amplify MMTV env was tested by using DNA from the mouse strain C3H/HeJ, which contains one proviral MMTV copy per genome, as positive template (information from Jackson Laboratories). A dilution series of this DNA in six 10-fold steps, covering from 1000 to 0.01 proviral copies per PCR, was included in every PCR run. Negative-control samples without templates were also included. Synthetic control DNA templates of 140 bp were ordered from Thermo Hybaid (see Supplementary Table S1, available in JGV Online). Mismatches in the primers were covered by the two primer variants ABL and ABL1, and ABR and ABR1, respectively. The TaqMan real-time PCR has several built-in features, which minimize the risk of false-positive results due to amplicon carry-over: UNG was used to degrade any amplicon, which may contaminate the reaction mixture, and the measurements took place in a closed tube during the PCR.

(46K): Fig. 1. Nucleotide alignment of a segment of three published HMTV env gene sequences. The positions of primers (grey background) and probes (black background) are indicated. An asterisk indicates identity in all three sequences.

Housekeeping gene control PCR.
The presence and amplifiability of DNA was assayed by using a Histone 3.3 TaqMan real-time PCR (Medstrand et al., 1992; Medstrand & Blomberg, 1993; Yin et al., 1997; Andersson et al., 2005). Amplification mixtures (25 µl) contained approximately 100 ng template DNA, 12.5 µl 2x TaqMan Master Mix with UNG (Applied Biosystems), 200 nM forward and reverse primers and 200 nM Histone TaqMan probe. The thermal conditions comprised 2 min UNG treatment followed by 10 min polymerase activation at 95 °C, followed by 55 cycles at 95 °C for 15 s and 54 °C for 60 s. The primers were: Histone forward (HisFw), 5'-CTCTACTGGAGGGGTGAAGAA-3'; Histone reverse (HisRev), 5'-TGCCTCCTGCAAAGCACCGATA-3'; and Histone probe (HisProbe), 5'-CTCTGGAAGCGCAGATCTGTTTTAAAGTCCTG-3'. Reactions were run on a Rotor-Gene 2000 Real-time thermocycler (Corbett Research). Serial dilutions of Histone 3.3 plasmids containing 10⁶10⁰ copies per PCR were used in the experiment as quantitative standards.

A control for PCR inhibition was to run all DNA samples, and two samples containing water only, in the presence of (spiked with) 10 proviral copies of C3H/HeJ DNA per reaction.

The PCR detected murine genomic MMTV DNA in 10-fold dilutions down to five, seven and nine copies per reaction, in three determinations. In contrast, all human DNA samples, as well as the negative water controls, gave negative results. Thus, tumour and normal control breast samples, as well as blood mononuclear cells from the 18 breast cancer patients, were all negative for HMTV DNA. Likewise, normal control breast samples and blood mononuclear cells from 11 control persons were negative for HMTV DNA. All cDNAs prepared from the RNA samples also became negative. The PCR volume contained around 100 ng tissue or leukocyte DNA. This corresponds to approximately 3.4x10⁴ cells. Thus, the limit of detection was better than one HMTV copy per 10³ cells. Synthetic HMTV and MMTV oligonucleotides with a size of 140 bp were used as positive controls to verify the sensitivity. The result showed that the PCR detected down to one copy in 10-fold serial dilutions. Forward and reverse primers with (degenerated primers ABL1 and ABR1) and without the single-nucleotide mismatches mentioned above were tested against positive controls and patient samples. The presence or absence of any of the two mismatches did not influence the result. Limit cycle (C_t) values were identical for PCRs with any of the four possible combinations of primers. The Histone 3.3 genes were detected in all samples, showing that the DNA was amplifiable. Analysis by gel electrophoresis resulted in homogeneous band intensities for all samples. The spiking of 34 cancer and control patient samples with a low number of MMTV DNA copies showed only a slight inhibition, less than one C_t step, in the patient samples compared with water controls. MMTV causes tumours in mice by insertional mutagenesis. MMTV, the prototype betaretrovirus, was discovered by Bittner more than 70 years ago (Bittner, 1936). Recently, several research groups claim to have detected HMTV/MMTV-specific (distinctive from human endogenous retrovirus) env sequences in humans (Wang et al., 1995; Etkind et al., 2000; Liu et al., 2001; Ford et al., 2004b). The human genome contains a large variety of sequences related to MMTV, i.e. HMLs (Andersson et al., 1999). These are much more evolutionarily distant from MMTV than is the alleged HMTV (Andersson et al., 1999). HMTV/MMTV has also been reported recently in primary biliary cirrhosis (Xu et al., 2004). In the present study, the endogenous betaretroviral sequences were distinguished from HMTV by PCR primers located in the env region of the MMTV provirus, as this region is clearly distinct in nucleotide sequence from that of the HMLs (Wang et al., 1995).

The aim was to search for DNA of an exogenous MMTV-related virus (here referred to as HMTV). Despite having a very sensitive technique, we did not detect HMTV DNA in human breast cancer tissue. The PCR could detect not only MMTV, but also minor HMTV sequence variants in synthetic DNA form, with good efficiency. This diminishes the likelihood of a false-negative outcome. Our inability to detect the virus indicates that the concentration of a putative HMTV DNA is nil or very low in human malignant and non-malignant breast tissue. The negative outcome of this study deviates from some, but is concordant with other (reviewed by Mant & Cason, 2004), reports on HMTV sequences in human breast cancer. We conclude that HMTV DNA is not present in Swedish breast cancer samples at concentrations higher than one copy per 10³ cells. If there is an HMTV at very low concentration in human breast cancers, it is unlikely to cause tumours in the same fashion as MMTV, i.e. by cis activation of onc genes by enhancement from adjacent proviruses, which requires a highly replicating virus and generates high viral DNA concentrations. The possibility that previous reports on HMTV in human breast cancers may have been due to unspecific (e.g. due to plasmid or amplimer contamination) PCR amplification should be considered.

Conclusion
A highly sensitive, env-based HMTV and MMTV detection technique, independent of previous ones, was developed. HMTV DNA was not detected in breast cancer or controls, due either to low concentration or to lack of it.

We thank Patric Jern for critical reading of the manuscript and useful suggestions. We thank associate professors Ewa Lundgren and Anders Liss for their valuable help with collecting samples. The study was supported by funds at the Academic Hospital, Uppsala, Sweden.

Footnotes

^,†, S. Muradrasoli¹ †These authors contributed equally to this work.

A supplementary table showing synthetic control DNA templates is available with the online version of this paper.