Effects of Proviral Integration on Host Gene Expression

Publication Details

Cellular mechanisms that regulate retroviral gene expression have been described in detail in the previous sections. However, it also clear that retroviral infection can alter host-cell gene expression. As described above, retroviral trans-activators such as Tax may affect the transcription of host genes. The integration of retroviral DNA into the host-cell genome also can have important implications both for retroviral gene transcription and for the transcription of cellular genes. Many examples of these effects are discussed in Chapters 8 and 10; in this section, some of the alterations in host gene expression are briefly described.

Retroviral LTRs serve as compact, mobile promoter and enhancer sequences that may alter the expression of cellular genes adjacent to integrated proviruses. Therefore, retroviral integration may result in transcriptional activation of cellular genes. Integrated proviruses may activate cellular gene expression in cis either in somatic cells or following germ-line infection. Effects on cellular gene expression following retroviral integration into somatic cells will be detected only if there is a resultant phenotype that offers a selective advantage to the cell. Most commonly, this has been detected as increased cellular growth associated with oncogenesis. Germ line integration can result in more subtle changes in gene expression, such as the development of new mechanisms of tissue-specific gene regulation.

One of the most dramatic effects of retroviral infection is oncogenic transformation. This phenotype may be a direct result of retroviral integration, with perturbation of host-cell gene expression adjacent to the integrated provirus. Interaction of cellular transcription factors with the integrated retroviral LTRs can result in activation of transcription of adjacent cellular proto-oncogenes. During the past 15 years, many examples of LTR-mediated activation of oncogenes have been described (for review, see Kung et al. 1991; Tsichlis and Lazo 1991) and are discussed in detail in Chapter 10 In its originally described form, this mechanism of transcriptional activation was referred to as “promoter insertion.” Integrated ALV LTRs were identified 5′ to c-myc proto-oncogene sequences in avian bursal lymphomas caused by ALV (Hayward et al. 1981), and fusion mRNAs, transcripts initiated within the ALV LTR and extending through myc-coding sequences, were found in tumor cells. Very high levels of LTR-c-myc fusion mRNAs were detected, apparently reflecting the high degree of transcriptional activity of the ALV LTRs in bursal cells. Transcription factors that activate high-level ALV LTR transcription in bursal cells, such as those that bind to EFII-CCAAT/enhancer sequences in the LTR, may be one determinant of the tissue specificity of tumors seen in association with particular retroviral infections. Insertion of active LTR promoters with adjacent 5′ retroviral splice sites can also lead to oncogene activation through the formation of aberrantly spliced oncogene mRNAs resulting in altered proteins. Examples of this leader insertion model include the generation of variant mRNAs for the epidermal growth factor receptor, c-erbB, generated by ALV insertions in avian erythroblastosis (Goodwin et al. 1986), and alterations in Myb protein structure in MLV-induced leukemias (Shen-Ong et al. 1986).

With the analysis of more integration sites in tumors, it became apparent that a strict model of “promoter insertion” was not universally applicable in describing activation of cellular oncogenes by inserted proviral LTRs. In particular, MMTV LTRs were identified both 5′ and 3′ to cellular genes, originally referred to as the int genes (Nusse and Varmus 1982; Peters et al. 1983; for review, see Nusse 1991). The proviruses integrated 5′ of the genes were often in the opposite transcriptional orientation. In these cases, direct transcription from the MMTV LTR cannot account for the transcriptional activation of int gene expression. Furthermore, some MMTV proviruses are located at considerable distances (>10 kb) from the associated int gene. Nevertheless, the int genes adjacent to MMTV integration sites exhibited marked transcriptional activation. On the basis of these observations of int gene activation, the concept of enhancer insertion was developed. In this case, the inserted LTR serves as an enhancer for transcription of the cellular gene. Like natural enhancers adjacent to cellular genes, the LTR exerts its enhancing activities in an orientation- and position-independent manner and can act at a distance.

As described above, the mammary-specific enhancer element in the distal MMTV LTR is likely to be involved in activation of int gene expression during tumorigenesis. Both enhancer and promoter insertion events have also been demonstrated for hematopoietic malignancies caused by MLV and feline leukemia virus (FeLV), and multiple cellular genes have been shown to be transcriptionally activated by integrated MLV proviruses. Insertional activation of proto-oncogenes has been demonstrated to occur over strikingly long distances. MLV insertions in rat lymphomas have been shown to activate expression of the c-myc gene over 30–270-kb distances (Lazo et al. 1990). The molecular mechanisms underlying these long-range effects on transcription are unclear, but presumably, they are related to those that regulate long-distance effects on cellular promoters, as discussed above.

Integrated proviruses may also activate expression of genes other than oncogenes. In somatic cells of the mouse, numerous examples of alterations of gene expression due to the insertion of intracisternal A particles (IAPs) have been detected (Chapter 8. Endogenous retroviruses have apparently been used during the course of evolution to activate certain genes in a tissue-specific manner in different mammalian species. An integrated MLV-like LTR is responsible for male-specific expression of the mouse sex-limited protein (Loreni et al. 1988; Stavenhagen and Robins 1988). A steroid hormone response element in the LTR, in conjunction with adjacent enhancer binding proteins, mediates androgen-specific activation of this gene. Similarly, the insertion of endogenous retroviral sequences is apparently responsible for the tissue-specific expression of human salivary amylase (Ting et al. 1992). A parotid-specific enhancer element exhibiting specific binding to parotid nuclear proteins was identified in retroviral sequences mapping between the 5′ LTR and the start of the gag gene. Interestingly, different mechanisms are responsible for salivary-gland-specific expression of this gene in other species; in the mouse, for example, retroviral sequences do not play a part in tissue-specific expression.

Retroviral integration can also inactivate the expression of cellular genes. Integration may be associated with interruption and/or deletion of host-cell gene sequences and production of aberrant mRNAs. Examples of gene inactivation are more difficult to identify based on the requirement for inactivation of both alleles in order to detect a phenotype for most genes. Nevertheless, a number of examples have been identified both in the induction of mouse mutations by integrated endogenous proviruses (see Chapter 8 and in oncogenesis. In an early experimental system, MLV infection resulted in inactivation of v-src expression from an integrated RSV provirus in transformed cells (Varmus et al. 1981). An example of oncogenic effects of retroviral insertion through inactivation of gene activity is seen in leukemogenesis induced by Fr-MLV and spleen focus-forming virus (SFFV). Integration of proviral DNA into the p53 tumor suppresser gene with resultant loss of gene expression has been identified in leukemic cell lines and leukemic spleens (Mowat et al. 1985; David et al. 1988; Ben-David et al. 1990). In some tumors, integration is associated with deletions of p53-coding sequences. In others, integration into intronic sequences results in the synthesis of aberrant p53 mRNAs, likely through altered splicing and polyadenylation due to the insertion of the viral RNA processing signals.