The Type B Leukemogenic Virus Truncated Superantigen Is Dispensable for T-Cell Lymphomagenesis

Abstract

Type B leukemogenic virus (TBLV) is a variant of mouse mammary tumor virus (MMTV) that causes T-cell lymphomas in mice. We have constructed a TBLV-MMTV hybrid, pHYB-TBLV, in which 756 bp of the C3H MMTV long terminal repeat (LTR) was replaced with 438 bp of the TBLV LTR. Intraperitoneal injection of pHYB-TBLV transfectants consistently resulted in T-cell lymphomas in 50% of injected weanling BALB/c mice with an average latency period of 5.7 (± 1.5) months. Transfectants of pHYB-TBLV containing a double-frameshift mutation in the truncated superantigen gene (sag) induced T-cell lymphomas with similar incidences, latency periods, and phenotypes, suggesting that cis-acting elements in the TBLV LTR determine disease specificity.

Mouse mammary tumor virus (MMTV) is a slow-acting oncogenic retrovirus that causes mammary tumors in mice after a long period (6 to 9 months) (1, 29). An MMTV variant, type B leukemogenic virus (TBLV), has been demonstrated to cause thymic lymphomas with a short latency period (1.5 to 2 months) when injected intrathymically into neonatal mice (2). Compared to MMTV, TBLV has a deletion of 443 nucleotides and a substitution of 124 nucleotides in the U3 region of the long terminal repeat (LTR) (3). This deletion removes a negative regulatory element (NRE) present in the MMTV LTR (17-19, 23, 36, 37), which has been implicated in tissue-specific MMTV expression (6, 30). The substitution creates a triplication of a 62-bp element comprised of sequences flanking the deletion (3) (Fig. 1A).

FIG. 1.

Construction and characterization of hybrid MMTV-TBLV proviral clones. (A) Diagram of parental and hybrid MMTV clones. The pHYB-MTV plasmid contains a hybrid between Mtv1 and C3H MMTV. A white box indicates the region upstream of the EcoRI site that is derived from Mtv1. The region downstream of the EcoRI site is derived from exogenous C3H MMTV (31) and is shown as a stippled box, except for the C3H MMTV NRE (black box) that is missing in TBLV. TBLV also contains a triplication of sequences surrounding the MMTV NRE deletion, creating a T-cell enhancer (21). The approximate positions of the viral structural genes are shown, and the LTRs are represented by the larger boxes at the ends of the provirus. The sag genes are designated by boxes above the proviral constructs. The Sag ORF in the TBLV LTR is truncated due to deletion and substitution mutations, yielding a putative Sag protein of 165 amino acids. A mutant TBLV hybrid provirus, pHYB-TBLVsagDFS, that contained two frameshift mutations at the AvrII and ClaI sites in the sag coding sequence (indicated by arrows) was constructed. These mutations would allow translation of only 17 amino acid residues. A 1.5-kb XhoI fragment from pUHD15-1-Hygro(-) provided by Paul Bates (University of Pennsylvania School of Medicine) was inserted into the NheI site of all constructs to provide hygromycin resistance. Details of constructions are available on request. aa, amino acids. (B) Western blotting of Jurkat cells stably transfected with proviruses containing the wild-type TBLV LTR or the TBLV LTR containing two frameshift mutations in the sag gene. Jurkat cells were transfected with various constructs by using Superfect (Qiagen, Valencia, Calif.) according to the manufacturer's directions and were selected in 0.2 mg of hygromycin per ml. Each lane contained 100 μg of cellular lysate, and viral protein expression was detected by using antibody specific for MMTV CA protein (Anti-Gag) (upper panel). Arrows indicate processed and unprocessed forms of the Gag protein. Lysates from untransfected Jurkat cells are shown for comparison. Western blots were performed as previously described (26). The same amounts of cellular lysates (50 μg) were incubated with antibodies specific for actin (Anti-Actin) (lower panel) as a control for protein loading on the gel. (C) Extracellular virus production from transfected Jurkat cells. Various amounts of cell supernatants from pHYB-TBLV or pHYB-TBLVsagDFS transfectants were analyzed by Western blotting with antiserum specific for MMTV CA protein and compared to virus production from the TBLV-induced lymphoma, 485-10 (11).

The LTR deletion and substitution also truncate the open reading frame (ORF) encompassing the U3 region of the LTR (Fig. 1A). This ORF encodes a 37-kDa, type II transmembrane glycoprotein known as superantigen (Sag) that functions in the milk-borne transmission of MMTV (1, 7, 9, 10, 16) by the stimulation of T cells through the T-cell receptor (TCR) (33, 35). The truncated Sag protein encoded in the TBLV LTR lacks the entire C-terminal region that interacts with the TCR, yet it retains the transmembrane region, a part of the extracellular region that includes the five glycosylation sites, a single proteolytic processing site, and the major histocompatibility complex class II protein binding motif (12, 20). This study shows that the truncated TBLV Sag protein is dispensable for virally induced T-cell lymphomas.

A 438-bp region of the TBLV LTR alters MMTV disease specificity.

Earlier studies showed that the tropism of the MMTV hybrid provirus pHYB-MTV (31) can be altered by replacement of the 3′ LTR with that from thymotropic MMTVs (34). However, these studies did not distinguish between the activity of the truncated Sag protein and that of the cis-acting elements in the LTRs. Therefore, TBLV-specific sequences from the ClaI-to-SstI region in the LTR (438 bp), including the T-cell-specific enhancer (21), were obtained by PCR and used to replace 756 bp from the ClaI-to-SstI region of the pHYB-MTV 3′ LTR. Subsequently, Jurkat T cells stably transfected with pHYB-TBLV were derived with the assumption that these cells would express high levels of HYB-TBLV, thus allowing high-efficiency infection and tumorigenesis in mice. Jurkat cells were also transfected with pHYB-TBLV that had been modified by the insertion of two frameshift mutations in the truncated sag gene (pHYB-TBLVsagDFS). Transfected cells were then tested for expression of the Gag CA protein by Western blotting. Comparable amounts of viral proteins from the two constructs were detectable in cell lysates (Fig. 1B) and in culture supernatants (Fig. 1C). Since the frameshift mutations also lie within the transcriptional regulatory sequence, we inserted the 3′ LTR from pHYB-TBLVsagDFS and pHYB-TBLV into a luciferase expression vector (32) for transient-transfection experiments in Jurkat cells as described previously (21). No significant difference between the luciferase activities of the two LTRs was observed (data not shown).

Subsequently, 2 × 10⁷ Jurkat cells stably expressing HYB-TBLV or HYB-TBLVsagDFS were injected intraperitoneally into weanling BALB/cJ mice. Injection of HYB-TBLV transfectants resulted in tumors in 50% of the mice with a latency period of 5.7 (± 1.5) months (mean [± standard deviation]), whereas mice receiving HYB-TBLVsagDFS transfectants developed T-cell lymphomas with a 60% incidence and an average latency period of 4.8 (± 0.6) months (Table 1) (P > 0.05). Expression of HYB-TBLV or the sag frameshift mutant in these tumors was confirmed by reverse transcription-PCR and sequencing as described by Mustafa et al. (26) (data not shown). These results revealed that an ∼440-bp region from the TBLV LTR was sufficient to alter MMTV disease tropism and that the truncated TBLV Sag protein was not essential for the development of virally induced tumors.

TABLE 1.

Incidence and latency of tumors induced by clonal TBLV proviruses after injection of transfected Jurkat T cells

Hybrid provirus used for transfection	Tumor type induced	No. of mice injected (F/M)a	Tumor incidence (%)	Avg tumor latency (mo)	Time of observation (mo)
pHYB-TBLV	T-cell lymphoma	12/6	50	5.7 ± 1.5b	12
pHYB-TBLVsagDFS	T-cell lymphoma	22/8	60	4.8 ± 0.6b	12
	Mammary tumor		4.6c	9.5d

^aTransfectants were injected intraperitoneally into the indicated number of female (F) and male (M) weanling BALB/cJ mice.

^bThere was no statistically significant difference (P = 0.08) between the average latencies of tumors induced by the wild-type and sag double-frameshift mutant viruses. Differences between wild-type and mutant-injected animals were analyzed with SPSS (SPSS, Inc., Chicago, Ill.). Time to tumor appearance was evaluated with the Kaplan-Meier survival analysis technique with a log-rank test of significance. This analysis excluded one HYB-TBLV-induced tumor that appeared at 12 months.

^cThe mammary tumor incidence was based on the number of female mice injected.

^dOnly one mammary tumor was observed.

Similar clonalities and phenotypes of HYB-TBLV and HYB-TBLVsagDFS-induced T-cell lymphomas.

To determine whether the phenotypes of T-cell lymphomas induced by the wild-type (HYB-TBLV) and the mutated (HYB-TBLVsagDFS) viruses were similar, we analyzed the tumor cell populations by flow cytometry (Table 2). Most tumors had a high percentage of Thy1.2⁺ cells, thereby confirming the T-cell origin of the cells. The Thy1.2⁺ cell populations were heterogeneous with respect to the surface distribution of CD4 and CD8 markers, but this distribution was not significantly different for the two groups. The tumors were not outgrowths of the originally injected cells since the antibodies used for cell surface staining did not react with Jurkat cells.

TABLE 2.

Comparison of cell surface markers on lymphomas induced by HYB-TBLV and HYB-TBLVsagDFS in adult BALB/c mice^a

Virus or control	Tumorb	% Cell surface marker
		Thy1.2+ CD4+	Thy1.2+ CD8+	Thy1.2+ CD4− CD8−	Thy1.2+ CD4+ CD8+
HYB-TBLVc	11	9	2.1	2	87
	12	73	0.8	25	1.7
	13	27	0.6	72	0.1
	14	4.1	22	2.4	71
	15	26	0.6	3	70
HYB-TBLVsagDFSc	1	21	45	13	22
	2	3.4	8.3	2.6	86
	3	19	3.8	5.5	72
	4	5.7	11.4	4.4	79
	5	74	0.2	24	1.5
Normal BALB/c thymus		11	3.3	2.1	84

^aCells were analyzed with a FACS Calibur flow cytometer (Becton Dickinson, San Jose, Calif.) and monoclonal antibodies from Pharmingen (San Diego, Calif.). The data were analyzed with Cellquest software.

^bTumor designation.

^cStatistical analysis was performed with Student's t tests with and without logarithm transformation to stabilize the variance.

To analyze the clonality of thymic lymphomas induced by the wild-type and sag frameshift TBLV strains, TCR β and γ chain rearrangements were analyzed by Southern blotting (Fig. 2). Most of the tumors induced by either virus showed rearrangement of both TCR chains, and many tumors had significant clonal populations. Therefore, the tumors induced by the wild-type and mutated TBLV hybrid proviruses were very similar and comparable to tumors induced by intrathymic injection of TBLV virions into newborn mice (8, 24, 25). These results, combined with cell surface analysis, suggest that the tumors were oligoclonal.

FIG. 2.

Analysis of T-cell lymphomas induced by HYB-TBLV and HYB-TBLVsagDFS for TCR rearrangements. (A) Southern blotting for TCR β chain rearrangements. (B) Southern blotting for TCR γ chain rearrangements. Tumors 13 and 18 were induced by HYB-TBLV, whereas tumors 1, 2, 3, 6, 8, 17, and 21 were induced by HYB-TBLVsagDFS. Additional tumors induced by wild-type TBLV have been characterized previously (22, 24). Digestion of liver DNA from uninfected animals was used to show the DNA fragments that were derived from the unrearranged TCR genes (arrows). Genomic DNA (15 μg) was digested with HindIII, separated on 0.8% agarose gels, transferred to nitrocellulose, and hybridized as previously described (22). Probes p86T5β (for TCR β chain) (13) and Cγ1.2 (for TCR γ chain) (15) were used for hybridization in panels A and B, respectively.

Lack of requirement for the truncated TBLV Sag during tumorigenesis.

Intrathymic injection with tissue culture supernatant from a TBLV-induced T-cell lymphoma line, 485-10, produces T-cell lymphomas in 90% of injected animals (2, 4, 5, 28). The disadvantages of this method are that these virus preparations are uncloned and heterogeneous (our unpublished data). Our clonal TBLV preparations were not injected into newborn mice, but the lower incidence and longer latency period of TBLV-induced tumors in the present study might be explained in two ways. First, since the thymus starts to regress in adult mice, less thymic tissue is available for TBLV infection and random insertions that lead to tumors (8, 27). Second, the route of infection may influence the efficiency of tumorigenesis. In previous experiments, culture supernatants were used to introduce TBLV intrathymically, thus directly inoculating the target for oncogenesis, whereas TBLV-expressing Jurkat T cells were introduced into the mouse peritoneum and must have been transmitted to thymocytes by an unknown mechanism. Inoculations of Jurkat cell transfectants into newborn mice may provide more efficient cell-to-cell transmission than injections of virions, which have given erratic results (unpublished observations).

The truncated Sag protein was not required for virally induced T-cell lymphomas in adult BALB/c mice when transfected Jurkat cells were used for TBLV infection (Table 2). This result may be specific for this route of infection, or, alternatively, the Sag proteins expressed from the endogenous MMTVs in BALB/c mice may complement the sag-null virus. However, it appears likely that the truncated Sag protein has no role in TBLV-induced tumors and that the loss of the NRE and/or other cis-acting elements is responsible for altering the viral disease specificity (14). The development of a cloned TBLV provirus provides us with a valuable tool for molecular dissection of the elements required for changes in MMTV disease specificity.