Identification of Herpesvirus-Like DNA Sequences in AIDS-Associated Kaposi's Sarcoma

Representational difference analysis was used to isolate unique sequences present in more than 90 percent of Kaposi's sarcoma (KS) tissues obtained from patients with acquired immunodeficiency syndrome (AIDS). These sequences were not present in tissue DNA from non-AIDS patients, but were present in 15 percent of non-KS tissue DNA samples from AIDS patients. The sequences are homologous to, but distinct from, capsid and tegument protein genes of the Gammaherpesvirinae, herpesvirus saimiri and Epstein-Barr virus. These KS-associated herpesvirus-like (KSHV) sequences appear to define a new human herpesvirus.

[]y32P]ATP (6000 Ci/mmol), and 250 units of cAMP-dependent protein kinase catalytic subunit from bovine heart tissue for 1 hour at room temperature. The beads were then washed extensively and eluted with 10 mM glutathione. The specific activity of all preparations was typically >1 x 107 cpm/pg. SDS-PAGE analysis showed a single band at the predicted sizes for the GST-SHC fusion proteins with either Coomassie staining or autoradiography. Immunoprecipitates or portions of cell lysates containing equal amounts of total protein were separated by SDS-PAGE and transferred to nitrocellulose. The filters were blocked for 2 hours at 4°C in nonfat dry milk (5%) in hybridization buffer [20 mM Hepes (pH 7.7), 75 mM KCI, 0.1 mM EDTA, 2.5 mM MgCI2, 1 mM DTT, and 0.05% Triton X-100]. The filters were then incubated overnight at 40C in hybridization buffer containing milk (1 %) and 2.5 x 105 cpm/ml of 32P-GST-SHC fusion protein as a probe. The filters were then washed three times in hybridization buffer with milk (1 %), dried, and exposed to x-ray film with an intensifying screen for 6 to 36 hours at -700C. 14. Lysate was prepared in hybridization buffer from 2.5 x 107 BAL17 B cells stimulated by cross-linking the B cell antigen receptor as described [T. M. Saxton et al., J. Immunol. 153, 623 (1994)]. The lysate was incubated with approximately 250 ng of GST-SHCASH2 protein containing the 1HA epitope tag for 1 hour at 40C. The mixture was then subjected to immunoaffinity chromatography with the use of a monoclonal antibody to 1HA covalently linked to agarose beads. The column was washed with 50 column volumes of hybridization buffer and eluted with 2% SDS. Proteins in equal fractions of the starting mixture, column flowthrough, and SDS eluate were separated by SDS-PAGE, transferred to nitrocellulose, and blotted with 32P-labeled PTB domain protein probe. In B cells, ppl 45 was seen as a doublet. 15. Anti-SHC immunoprecipitates from PDGF-stimulated fibroblasts immobilized on nitrocellulose filters were incubated in 25 mM imidazole (pH 7.0), 50 mM NaCI, 2.5 mM EDTA, 5 mM DTT, acetylated bovine serum albumin (100 gg/mI), and 5 units each of LAR and T cell tyrosine-specific phosphatases for 60 min at 300C. An equivalent sample was treated identically except that 5 mM sodium orthovanadate was included. The filters were then washed extensively and blotted with 32P-GST-SHC as above, except that the hybridization buffer included 1 mM sodium orthovanadate. 16 analysis (RDA) to identify and characterize unique DNA sequences in KS tissue that are either absent or present in low copy number in nondiseased tissue obtained from the same patient (7). This method can detect adenovirus genome added in single copy to human DNA, but has not been used to identify previously uncultured infectious agents. RDA is performed by making simplified "representations" of genomes from diseased and normal tissues obtained from the same individual through polymerase chain reaction (PCR) amplification of short restriction fragments. The DNA representation from the diseased tissue is then ligated to a priming sequence and hybridized to an excess of unligated, normal-tissue DNA representation (8). Only unique sequences found in the diseased tissue that have priming sequences on both DNA strands are preferentially amplified during subsequent rounds of PCR amplification. This process can be repeated with different ligated priming sequences to enrich the sample for unique DNA sequences that are found only in the tissue of interest.
The initial round of amplification-hybridization from KS and excess normal-tissue DNA resulted in a diffuse banding pattern (Fig. 1, lane 2), but four bands at approximately 380, 450, 540, and 680 base pairs (bp) were identifiable after the second amplification-hybridization (Fig. 1, lane 3). These bands became discrete after a third round of amplification-hybridization (Fig. 1, lane 4). Control RDA, performed by hybridizing DNA extracted from AIDS-KS tissue against itself, produced a single band at -540 bp (Fig. 1 (1 1), a gammaherpesvirus that causes fulminant lymphoma in New World monkeys. This fragment is also 39% identical to the amino acid sequence encoded by the corresponding BDLF1 ORF of Epstein-Barr virus (EBV) (12) is a 330-bp sequence with a 51% G:C content ( Fig. 2A), and KS631Bam is a 631-bp sequence with a 63% G:C content (Fig. 2B).
The polypeptide encoded by the KSassociated DNA open reading frame shows extensive amino acid homology to the proteins encoded by herpesvirus saimiri ORF26 and EBV BDLF1 (Fig. 3). Although it is homologous to these herpesvirus regions, the polypeptide does not match any other known sequence and thus provides evidence for a viral genome related to but distinct from known members of the herpesvirus family. In addition, the 5' end of the 1853-bp sequence (bp 1 to 607) is 66% and 67% identical to corresponding regions of the major capsid protein (MCP) genes of herpesvirus saimiri (ORF25) and EBV (BcLF1), respectively. In both EBV and herpesvirus saimiri genomes, the MCP gene is found immediately adjacent to the BDLF1-ORF26 gene (11,12). This region also has lower degrees of similarity to MCP genes of other human herpesviruses, including HSV1, VZV, HHV6, CMV, and HHV7 (14).
To determine the specificity of KS330Bam and KS631 Bam for AIDS-KS, these sequences were random-primed, 32p_ labeled, and hybridized to Southern blots of DNA extracted from cryopreserved tissues obtained from patients with and without AIDS (15). Twenty of 27 (74%) AIDS-KS DNA specimens hybridized with variable intensity to both KS330Bam and KS631Bam, and one additional KS specimen hybridized only to KS63 1 Bam by Southern blotting (Fig. 4 and Table 1). In contrast to AIDS-KS lesions, only 6 of 39 (15%) non-KS tissues from patients with AIDS hybridized to KS330Bam and KS63 1 Bam. Specific hybridization did not occur with lymphoma or lymph node DNA from 36 persons without AIDS or with control DNA from 49 tissue biopsy specimens obtained from a consecutive series of patients. DNA specimens extracted from vascular tumors and tissues with opportunistic infections common in AIDS were also negative ( Table 1). In addition, DNA samples from EBV-infected peripheral blood lymphocytes and pure cultures of Mycobacterium avium-complex were negative as well. Overall, 20 of 27 (74%) AIDS-KS specimens hybridized to KS330Bam and 21 of 27 (78%) AIDS-KS specimens hybridized to KS63 1 Bam,as compared to only 6 of 142 (4%) non-KS human DNA control specimens (X2 = 85.02, P < 10-7 and x2 = 92.4, P <10 -7, respectively).
The sequence copy number in the AIDS-KS tissues was estimated by simultaneous and a hybridization with KS330Bam 440-bp probe for the single-copy constant region of the T cell receptor P gene (16). Fig. 4 showed similar intensities for the two probes, indicating an average copy number of approximately two KS330Bam sequences per cell, whereas remaining KS tissues had weaker hybridization signals for the KS330Bam probe.

Samples in lanes 5 and 6 of
These results were confirmed and extended by PCR amplification with primers designed from KS330Bam ( Fig. 2A) that amplify a 233-bp subfragment (17) designated KS330233. Although Southern blot hybridization detected the KS330Bam sequence in only 20 of 27 KS tissues, 25 of the 27 tissues were positive by PCR amplification for KS330233 (Fig. 5A), demonstrating that KS330Bam is present in some KS lesions at levels below the threshold for detection by Southern blot hybridization. The two AIDS-KS specimens that were negative for KS330233 ap-

Spleen
Liver Muscle peared to be so for technical reasons: One had no microscopically detectable KS tissue in the frozen sample (Fig. 5A, lane 3), and the other (Fig. 5A, lane 15) was negative in the control PCR amplification for the p53 gene (18), indicating either DNA degradation or the presence of PCR inhibitors in the sample. All KS330233 PCR products hybridized to a 32P end-labeled 25-bp internal oligomer, confirming the specificity of the PCR (Fig. 5B).
Except for the six non-KS control samples from AIDS patients that were positive by Southern blot hybridization, none of the other 136 non-KS control specimens were positive by PCR for KS330233. Overall, DNA samples from 25 (93%) of 27 AIDS-KS tissues were positive by PCR, as compared to 6 (15%) of 39 non-KS lymph nodes and lymphomas from AIDS patients (X2 = 38.2, P < 10-6), 0 of 36 lymph nodes and lymphomas from non-AIDS patients (X2 = 55.2, P < 10-7), and 0 of 49 consecutive biopsy specimens (X2 = 67.7, P < 10-7). All control specimens were amplifiable for p53, indicating that inad-  (19), were negative for KS330233. Several KS-tissue DNA samples tested with EBV-specific and mycoplasmata-specific consensus PCR primers were also negative (20).
Of the six control tissues from AIDS patients that were positive by both PCR and Southern hybridization, two patients had KS at other sites, two did not develop KS, and complete clinical histories for the remaining two patients were unobtainable. Three of these tissues were lymph nodes with follicular hyperplasia taken from patients with AIDS. Undetected microscopic KS foci may have been present in these lymph nodes, given the high lifetime occurrence of KS (>50%) in some 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 19), and (B) shows specific B hybridization of the PCR products to a 32P end-~m labeled 25-bp internal oligonucleotide (Fig. 2) after transfer of the gel to a nitrocellulose filter. Negative samples in lanes 3 and 15, respectively, lacked microscopically detectable KS in the sample or did not amplify the human p53 exon 6, suggesting that these samples were negative for technical reasons. An additional eight AIDS-KS samples were amplified and all were positive for KS330233. Lane (21). Alternatively, these lymph nodes may have been asymptomatically infected with, or may have been incubating, the putative agent. The other three positive tissue specimens were a form of B cell immunoblastic lymphoma from AIDS patients. Given the previously noted association between KS and lymphoproliferative disorders (22), it is possible that the putative KS agent is also a cofactor for a subset of AIDS-associated lymphomas. A comparison of AIDS-KS tissues to only lymph node and lymphoma tissues from AIDS patients demonstrates that KS330Bam and KS63 1 Bam remain significantly associated with the KS phenotype when controlling for concurrent AIDS, indicating that HIV disease is not a confounding factor in our analysis. Among only AIDS tissue samples from separate patients (Table 1), over 90% of KS specimens (100% of confirmed and amplifiable KS specimens) were positive for KS330233, as compared to 15% of lymph node and lymphoma tissues from AIDS patients. These sequences therefore appear to be specifically associated with KS in AIDS patients, although it is not clear whether the presence of these sequences is causal or is an epiphenomenon of KS.
To show that KS330Bam and KS631Bam are not heritable polymorphic DNA markers for KS, we tested multiple unaffected tissue DNA samples from four additional patients with AIDS-KS ( Table  2). Whereas KS lesion DNA samples were positive by Southern hybridization and PCR, unaffected tissues were generally negative for these sequences. All other tissues except muscle and unaffected skin from patient A, stomach adjacent to the KS lesion in patient C, and adjacent skin and hilar lymph nodes in patient D were negative. These results are consistent with an infectious process and may represent local and disseminated spread of the putative virus.
Although these sequences suggest the presence of a new human herpesvirus in KS lesions, a causal link between these sequences and AIDS-KS cannot be established by our retrospective case control study. It is possible that this agent is a common latent virus in humans that preferentially colonizes KS lesions in immunosuppressed patients. Unlike previous studies searching for agents associated with KS, the sequences found in our study were present in all intact KS DNA samples from a large number of patients and were preferentially found in diseased as compared to normal tissues from the same host. Our results have been independently confirmed with 100% concordance in a blinded PCR evaluation with extracted AIDS-KS lesion DNA and non-KS brain DNA from the same patients (23).