On board a raft or boat in the retrovirus sea

Significance

The main significance of these long-term studies lies in establishing chicken Rous sarcoma virus (RSV) genome integration into a rodent species cell genome in the absence of infectious virus production. Infectious virus, however, can be rescued by fusion of RSV-transformed rodent cells with permissive chicken fibroblasts, which provide some functions critical for full virus expression. The nature of chicken factors required and the role of viral envelope changes are discussed.

Abstract

This article summarizes the essential steps in understanding the chicken Rous sarcoma virus (RSV) genome association with a nonpermissive rodent host cell genome. This insight was made possible by in-depth study of RSV-transformed rat XC cells, which were called virogenic because they indefinitely carry virus genetic information in the absence of any infectious virus production. However, the virus was rescued by association of XC cells with chicken fibroblasts, allowing cell fusion between both partners. This and additional studies led to the interpretation that the RSV genome gets integrated into the host cell genome as a provirus. Study of additional rodent virogenic cell lines provided evidence that the transcript of oncogene v-src can be transmitted to other retroviruses and produce cell transformation by itself. As discussed in the text, two main questions related to nonpermissiveness to retrovirus infection remain to be solved. The first is changes in the retrovirus envelope gene allowing virus entry into a nonpermissive cell. The second is the nature of the permissive cell functions required by the nonpermissive cell to ensure infectious virus production. Both lines of investigation are being pursued.

Working for several years on the first virogenic rat XC tumor cell line during the early 1960s in Prague, Czechoslovakia, I was isolated, like a lone man on a raft. The only encouragement came from my boss urging me to finally complete this work. Nevertheless, the work progressed, I attracted several good PhD students, and in the process acquired a boat. During the Soviet occupation of our country in 1968, I was expelled from the boat and found myself again alone on a rough raft. The profound changes in 1989, known as the Velvet Revolution, brought new challenges. I was elevated by public vote to the leadership of our Academy and directorship of our Institute, thus lifting me almost completely out of the retroviral sea. In 1997, after completing my service to reorganizing our Academy and Institute in the spirit of democratic changes, I handed over my group and started to build a new raft made of impermeable balsa logs. At present, I am again enjoying good sailing in favorable weather and proper wind.

My election to the National Academy of Sciences (NAS) is of great honor. Despite being founded at the height of civil war, the NAS has maintained an independent and progressive spirit, which is in contrast to my experience. After completing a study on the virogenic nature of XC cells, I was asked by Bob Huebner in the early 1960s to submit an article to PNAS. But in my own country, my request to be allowed to publish in PNAS was rejected for political reasons.

Background

The subject of my work is deeply rooted in early attempts to understand tumorigenesis via tumor virus and genetic research approaches, postulated respectively by Peyton Rous (1911) (1), who was first to thoroughly characterize a virus producing sarcomas in chickens, and Theodor Boveri (1914) (2), who recognized that disequilibrium in the normal distribution of cell genetic makeup repositioned in chromosomes leads to malignant cell formation. Boveri’s idea became known as “somatic cell mutation theory.” For decades, the cell mutation theory dominated oncology, until it became clear that oncogenic viruses, like retroviruses, integrate into the host cell genome, thus incorporating their oncogenic potential into the infected cells. Of crucial importance was the discovery that retroviral cancer genes, called oncogenes, correspond structurally to their counterparts in normal cells named proto-oncogenes (3). The transition from a proto-oncogene to an oncogene requires certain proto-oncogene activation steps ensuring constitutive expression and is facilitated by recombination with a retrovirus.

The somatic cell theory thus merged with the virus theory of cancer. Nevertheless, both theories are now confronting each other again. A series of oncogenes and other genes (driver genes) contributes especially to early carcinogenesis steps. However, progression to malignancy, metastatic process included, is being defined at the level of cancer cell progression and related to stepwise selection of progressively growing cell clones surpassing normal tissue barriers. Even as cancer research advances and grows in sophistication, retroviruses remain a focus and inspiration for many areas such as gene expression, RNA processing, mediation of virus and cell-to-cell interactions, and gene therapy.

Early Work with Rous Sarcoma Virus

My first encounter with Rous sarcoma virus (RSV) occurred in 1953, when I worked as a volunteer in the tissue culture laboratory of Helena Keilová under the supervision of Mojmír Brada. This virus was used in that laboratory and also at the Institute of Organic Chemistry and Biochemistry as a tool for cell transformation, and both places studied the metabolic consequences of this change. To a student who liked cell biology, chicken cell RSV transformation was a highly impressive process, suggesting that the virus is endowed with a strong capacity to change cell morphology and behavior within a short period. This impression has remained with me for my whole life and acted as a driving force in my experimental work.

The metabolic experiments we conducted were inconclusive, and we were about to terminate them. Then, the department was taken over by Milan Hašek, coauthor of immunological tolerance. After some disagreements, we concentrated our program on immunological tolerance to RSV, a field opened by Morten Simonsen and Bob Harris in London, who found that turkeys made tolerant to chicken antigens by intraembryonic injection of chicken blood became susceptible to RSV-induced oncogenesis. We confirmed these data and extended them to ducks (4), assuming that RSV tumors acquire some chicken antigens, virions included, and therefore suppression of anti-chicken antigen immunity favors the RSV tumor growth in foreign avian species.

RSV-Induced Rodent Tumors: XC Cell Story

Avian experiments prompted an attempt to induce immunological tolerance to RSV in mammals. We did not obtain such tolerance but discovered a rare sarcoma induction, as previously described by George Svet-Moldavsky (5). It was hard to interpret such findings because no infectious RSV was detected in these sarcomas, and therefore they might have arisen by a “hit and run” mechanism. However, careful testing of our tumor samples led to the encouraging observation that in one of the rat tumors, designated as XC, the virus was detectable, and when minced sarcoma tissue was inoculated in chickens, it produced fast-growing sarcomas containing infectious RSV (6). I assumed that this orphan finding might provide a good opportunity to study the RSV genome in transformed cells, but I did not succeed in getting any collaborative support. In a further three papers, I made the essential observation that the RSV genome in XC cells remained fixed even after 23 passages of tumor tissues in rats, implying a highly stable association with the host cell. I tackled another critical objection centered on the possibility that RSV did not trigger tumor formation but secondarily infected tumor cells of other etiology. This notion did not hold up because RSV injected into transplantable rat tumors or applied together with the carcinogen benzopyrene in both cases showed no RSV activity, and therefore the virus exerted no affinity to tumorigenesis triggered by other factors (7). In a further set of experiments, we verified by immunology and karyology that XC cells were of rat origin. Furthermore, I demonstrated that, to obtain virus production, it was necessary to inoculate chickens with 10⁵–10⁶ structurally intact XC cells. This number of XC cells remained constant over 23 XC tumor cell passages. Conversely, when the cells were broken by three cycles of freezing-thawing, they lost their tumorigenicity in chickens (8).

In later stages of XC research, I attracted Dušan Šimkovič from the Bratislava Oncology Institute for collaboration. He focused on establishing the XC cell line in tissue culture, which enabled cell cloning experiments that we performed jointly. It turned out that each XC cell clone retained the ability to produce RSV tumors in chickens when injected as a cell suspension (9).

Finally, together with my colleagues, we attempted a broad search for infectious RSV presence in RSV tumors and cells. Not surprisingly, no such virus was found either in high-speed pellets from XC culture medium or in the cell fraction corresponding to that of virus particles. Moreover, in rats bearing regressing XC tumors, no virus-neutralizing antibodies were present.

Based on 4-y XC cell study, we proposed that the RSV genome in mammalian cells behaved as stable, newly acquired genetic information, and we therefore classified XC cells as virogenic cells that carry the viral genome in proviral state (10). Since then, I remained in written contact with Howard Temin (11), who acknowledged our results in his Nobel lecture as independent proof of the RSV provirus state (12).

In summary, the results of long-term XC and other RSV-transformed mammalian cell lines provided convincing evidence for RSV genome stability. Moreover, the viral genome was retained in all cell lines and clones in the absence of any infectious virus production. Therefore, we called such cells virogenic because they harbor the integrated RSV provirus.

I submitted a detailed account of our XC and RSV work to Folia Biologica (Praha), including all citations of the original works (13). The essential contributions of XC cells to understanding the retrovirus–cell interaction are shown in Fig. 1 with reference also to ref. 14.

Fig. 1.

Genesis of RSV-transformed XC rat cells and the ways of rescuing infectious virus from them. I obtained the XC tumor cell line after inoculation of neonatal rats with chicken RSV tumor tissue. Even after long-term passaging, the XC cells, in the absence of infectious virus production, kept their capacity to induce chicken RSV-containing tumors when chickens were inoculated with structurally intact XC cells. The stability of the RSV genome that was maintained in every XC cell population, including monocellular clones, led to the conclusion that the RSV genome is integrated in the XC cell genome as a provirus. In the next step, RSV virus rescue by cell association was analyzed, revealing that the cell association leads to cell fusion and the chicken-XC cell hybrids (heterokaryons) become virus producers. The chemical proof of the provirus presence was provided by transfection experiments in which XC cell DNA was introduced into chicken cells, triggering RSV production (24). Working on the same issue, we established that RSV obtained after transfection belongs to the same subgroup as RSV rescued from XC cells by other means (14), thus demonstrating that DNA transmitted the provirus of XC cell origin.

I should remark on the situation in retrovirology in Eastern Europe and Russia at the end of the 1950s and beginning of the 1960s. In Czechoslovakia, in addition to our group, there were the Thurzo’s Institute of Experimental Oncology and an independent Říman’s group focused on biochemistry of leukemia viruses. We maintained a friendly relationship with Lev A. Zilber and his group and the George Svet-Moldavsky laboratory. Both went through Soviet concentration camps, called by the Russians “school of forestry,” but kept their liberal views. I still treasure the gift from George with the inscription “Don Quixote” of the year 1957 in the Russian alphabet. In East Germany, we had further partners in the renowned Arnold Graffi Institute. Thus, at the beginning, I did not feel totally lost at sea.

RSV Rescue from Virogenic Cells

My close contact with Western retrovirology was significantly enhanced by the International Avian Tumor Virus Conference in Durham, NC, in 1963, which was attended by eminent retrovirologists. Here, for the first time, I met Harry Rubin, who established and applied quantitative measurement of RSV cell transformation and, together with his pupils, provided a deep insight into retroviral genetics (15). Also present were his collaborators: Howard Temin, Peter Vogt, Saburo Hanfusa, and many others such as Fred Prince and National Institutes of Health (NIH) virology group attendants from additional countries. Therefore, I felt like I was on a boat on the retrovirus sea and highly valued the effort invested into research. I realized that it would be hard to compete in areas that require well-defined culture media, plastic, pure chemicals, and freezing technology, all of which we were lacking, and so I came to the conclusion that we had to preferentially stick to our models and ask questions solvable by our comparatively less advanced technology.

Our next point of interest was the mechanism of virus rescue by association of XC cells with chicken cells. As I proposed (16), the underlying mechanism should include transmission of the virus via intercellular bridges or better, by fusion with permissive chicken fibroblasts. Using inactivated Sendai virus treatment of mixed XC and chicken fibroblast cultures, we detected a clear increase in RSV rescue (17). During a stay in London at Imperial Cancer Research Fund, I performed quantitative experiments showing a 100-fold increase in the efficiency of RSV rescue and obtained a good correlation between the degree of virus rescue and the number of heterokaryons (18). Finally, the study of individual heterokaryons provided evidence that they produced infectious RSV (19).

In parallel to these experiments, we aimed to extend our experimental models beyond XC cells. Together with my new PhD students, namely Pavel Chýle, Václav Klement, Pavel Veselý, and Ivo Hložánek, we obtained new tumor cell lines of in vitro RSV-transformed rodent cells derived from rat, mouse, and Syrian hamster cells. All of them exhibited the virogenic characteristics previously described in XC cells. Of particular interest were primary Chinese hamster cells analyzed by karyology (done by L. Donner) before, during, and after RSV transformation. The results of extensive study provided clear evidence that RSV cell transformation was not accompanied by any detectable chromosomal change (20), thus supporting the idea that the virus itself was directly responsible.

Transformation experiments led to other general conclusions, namely that the most efficient way to achieve transformation was cocultivation of RSV-transformed chicken fibroblasts with rodent fibroblasts. The transformed rodent cells overgrew the chicken cells and their mammalian origin could be confirmed by karyology. We also proposed that cocultivation might contribute to direct virus transfer into mammalian cells (21), which became a subject of present-day interest in the context of HIV infection (22, 23).

Based on the previously mentioned data, I became fully confident that the RSV genome was integrated in transformed cells. There remained the question of chemical definition of the provirus. In those days, virogenic cells provided the best starting model for such studies because they eliminated any possibility of virus contamination. At the Czechoslovak Biology Conference, we agreed to cooperate with Miroslav Hill, who was focused on optimization of mammalian DNA transmission to chicken fibroblasts and presented the results of his studies. In fact, we agreed that he would move from Brno to our institute, where the transfection experiment would be performed. However, the turn of history changed our plans. As a result of the Soviet occupation of Czechoslovakia in 1968, Hill left our country and settled in Paris, where he published the first data on successful chicken cell transfection with XC DNA (24). In the turmoil of the occupation, we proceeded, although with some delay, extending the XC DNA result to an additional virogenic line, and we characterized the virus specificity belonging to the same subgroup C [Prague (PR) RSV C] as the virus we had previously rescued by fusion with chicken cells. Later, we addressed an additional topic related to transfection (25, 26).

Finally, I would like to stress that in no case were we successful in RSV transformation of human cells, in agreement with the findings of others. Whether human cells lack some required cofactor or whether they have evolved some function(s) suppressing the omnipotent RSV oncogene (src) remains a challenging issue for future research.

Summarizing the results of several of our studies focused on RSV tumor induction in various rodent species, we concluded that in more than half of the tested tumors, unaltered proviral genes are integrated, and, in 21%, a defective virus is present that can be rescued only in the presence of a helper virus (27).

We lacked functional information about RSV provirus integration sites, which probably regulate the degree of provirus expression. However, comprehensive density gradient analysis clearly suggested that the expressed proviruses are located in active chromatin regions (28).

The src Oncogene

An ultimate question was hanging over retrovirology. Does an acutely transforming retrovirus harbor a specific transforming gene? Persuasive findings about the existence of such a gene were provided by virus genetics (29) and molecular biology (30). We were complete outsiders because of the lack of any standard tissue culture equipment and ³²P nucleoside triphosphates, essential for the preparation of high-specific activity DNA probes. However, already in 1968, I proposed that some exceptional RSV-transformed rodent cells might contain only part of the RSV genome responsible for cell transformation (31). First, we studied one such mouse cell line from which no virus was rescuable—even after helper virus complementation—but which retained a specific transplantation antigen (TSTA) characteristic of RSV tumors of different species origin (31). We found, in collaboration with Marcel Baluda, that it carried about one third of the RSV genome (32). Despite the encouraging start, we revealed that the mouse tumor cell line used underwent karyological changes such as anomalous metacentric chromosomes and showed fluctuations in hybridization experiments. Therefore, I decided to induce a series of fresh tumors, which might yield more stable RSV integration patterns, including that represented by integration of the transforming RSV part already designated as v-src. One hamster tumor cell line (H-19) met all of the criteria predicted for a tumor line transformed by the RSV oncogene. Its provirus was highly deleted and expressed only v-src mRNA (33). We then rescued this RNA by H-19 fusion with chicken fibroblasts infected by Rous-associated virus (RAV)-1 helper virus. Under such conditions virions were formed, which transformed chicken cells by integrating v-src in the chicken genome site characteristic for each transformation event (34). In one case, we detected a recombinant between v-src and RAV-1 in which the v-src function was preserved due to the emergence of a new splice acceptor site (34). The complete nucleotide sequence of H-19 provirus revealed that it exactly matched a reverse transcript of v-src mRNA that had become integrated in hamster cells, producing a hexanucleotide direct repeat typical of α-retrovirus integration (35). Collectively, these findings define v-src as a minimal autonomous transforming RSV unit. In agreement with this conclusion, we established that purified DNA from the cloned provirus produced fast growing sarcomas in chickens accompanied by metastasis formation (36).

The hamster cell line (H-19) harboring just v-src produces revertants with high frequencies (10⁻⁶ nontransformed revertants per cell per generation, the v-src DNA of which is highly methylated). Conversely, when demethylated, the provirus resumes tumorigenic activity, which clearly points to the conclusion that reversible epigenetic modification is involved in the reversion (37). Interestingly, v-src was integrated in H-19 cells in an already methylated genomic region, which, however, became demethylated, possibly as a result of oncogene integration (38). This particular genomic site thus offers further analysis of the switched-on switched-off methylation status triggered by an oncogene.

In conclusion, we provided convincing evidence that v-src mRNA can be captured and incorporated into a nonacutely transforming (helper) virus, reverse transcribed, and integrated regularly into the host cell genome. The integrated oncogene can be efficiently silenced by methylation.

Why Are Mammalian Cells Nonpermissive for RSV Replication?

Despite decades of work, we still do not know why mammalian cells do not support RSV replication. At present, we are concentrating on functional studies related to nonpermissiveness to RSV because we are in possession of a series of virogenic cell lines differing in provirus expression. In our laboratory, we also cloned the subgroup C cell receptor (39), which facilitates elucidation of PR-RSV-C interaction with cells, even with cells of distant origin.

Thus far, our main tool has been represented by Chinese hamster cells transformed with this virus (RSCh cells) due to their previous detailed characterization. As described by Lounkova et al. (40), at least two blocks in RSV production were identified in RSCh cells: one at the level of unspliced viral RNA export from the nucleus to the cytoplasm and the other related to env gene expression. Both are relieved by cell fusion with chicken fibroblasts. Evaluation of the effect of protein synthesis inhibitors on virus rescue revealed clearly that the indicator chicken partner cells used for fusion require protein synthesis for their rescuing activity.

We are aware that nonpermissiveness is a multifaceted issue. It has become an urgent topic for us, as even different cell types of the same species differ in their permissiveness to retroviruses, HIV included.

Because of the nonpermissiveness complexity, we turned our interest to the Syrian hamster cell line H-20 containing one functional RSV provirus. According to our original data (41), this cell line expresses the processed gag gene product and synthesizes infectious virus detectable only by ultracentrifugation of higher amounts of tissue culture supernatant. At present, using a combination of proper methods, we are trying to complement the lack of chicken cell factor(s) using transduction of chicken cell extracts into hamster H-20 cells, where proteins needed for infectious virus production might be absent.

We are not only targeting the nature of nonpermissiveness due to the absence or lack of cell factors. Another important issue related to the changes in the virus has appeared as a result of its mutation in the envelope gene (env). Under strong selection, several env gene mutations spanning the surface (SU) and transmembrane (TM) RSV domains were observed, allowing mammalian cell infection (42–44). In our experiments we use PR RSV-C, which was passaged twice through rodent cells. This virus originally suffered eight amino acid changes that do not match alterations described thus far.

α-Retroviruses are supposed to be functionally simple compared with other retroviral taxa, meaning that they express only virion proteins, where complex viruses like HIV encode a number of proteins not found in virions, whose job is to counter cellular antiviral factors like APOBEC3G and tetherin. As a result of this, α-retroviruses do not develop resistance to classical dominant-negative nonpermissive factors. An extreme situation represented by RSV mammalian cell transformation suggests that RSV expansion to mammals is also limited by other reasons, namely by the absence of avian factors in mammalian cells required for full RSV expression.

α-Retroviruses possess some common features with HIV with regard to envelope gene composition and function, such as a two-step mechanism of virus entry in the cell (45), which make them important tools for studying general retrovirology. The products of different retrovirus genome regions, especially gag and env, are involved in nonpermissiveness of certain cell lines to retrovirus infection (46, 47). To overcome this nonpermissiveness, env and/or gag gene products require interaction with cell-specific factors. As reported by Lounková et al. (40), such factors also play an important role in the nonpermissiveness of cells of foreign species. Despite the rarity of transclass transmission, γ-retrovirus transmission represented by REV and amphotropic MLV (48–50) to rodents also represents an additional well-documented case of this phenomenon, which should be taken as an additional warning of retrovirus expansion. Because of the wealth of information about RSV, it can be used to clarify the cell factors determining the virus fate in cells of foreign species.

Does retrovirology remain a sea, or is it better classified as an ocean? In favor of the latter are the observations that about half of our genetic information arose by reverse transcription and at least 8% of our genome is represented by integrated proviruses. To clarify how this happened and what consequences it brings about remains an important task not only for virology but for biology in general. We can expect many new discoveries in this fascinating field. Settled on a well-built raft, I will follow these issues as long as possible and try to discover new findings and ideas to prevent retrovirology to be diluted in the ocean of general knowledge and fall into oblivion.