University of Southern California Dissertations and Theses

In vivo evolution of recombinant feline leukemia viruses (FeLV) toward FeLV subgroup B viruses in thymic lymphomagenesis

(USC Thesis Other)

In vivo evolution of recombinant feline leukemia viruses (FeLV) toward FeLV subgroup B viruses in thymic lymphomagenesis

PDF

Download

Open document

Flip pages

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content IN VIVO EVOLUTION OF RECOMBINANT FELINE LEUKEMIA VIRUSES
(FeLV) TOWARD FeLV SUBGROUP B VIRUSES IN
THYMIC LYMPHOMAGENESIS
by
MARTA KEITH BECHTEL
A Dissertation Presented to the
FACULTY of the GRADUATE SCHOOL
UNIVERSITY of SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree of
DOCTOR OF PHILOSOPHY
(Biochemistry and Molecular Biology)
August 1997
Copyright 1997 Marta Keith Bechtel
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
UNIVERSITY OF SOUTHERN CALIFORNIA
THE GRADUATE SCHOOL
UNIVERSITY PARK
LOS ANGELES. CALIFORNIA 90007
This dissertation, written by
• • • • • • • • • • T a • > < • • • • • • •
under the direction of IlJ U u Dissertation
Committee, and approved by all its members,
has been presented to and accepted by The
Graduate School, in partial fulfillment of re
quirements for the degree of
DOCTOR OF PHILOSOPHY
tte Studies
D ate AugusL 8;
LX^....A
T J Jn O i c L v r ^ t T s a n .
C/L^ _
^
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
DEDICATION
This work is dedicated to my family. My husband,
Thurston Neville Bechtel, who patiently encouraged,
advised, and supported me with his love and
understanding. My son, Geoffrey Keith Bechtel, who also
demonstrated great patience and an understanding far
beyond his years while encouraging me throughout this
endeavor, and who is certainly my most precious and
significant contribution to the world. You both made
great sacrifices in order to allow me to chart a new
course in life and I appreciate it more than words can
express.
This work is also dedicated to the many friends and
family who encouraged me to pursue my dreams ; most
especially my sisters, Estrellita and Micky, and my
niece Anna. Lastly, this work is dedicated to my Mom,
Hilda Gladys Cardenas de Keith, who set an example in
life I strive to emulate every day. I love you all.
ii
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
ACKNOW LEDŒ Œ NTS
I am pleased to acknowledge the many people who
have assisted me in achieving this goal in life. My
fellow lab members: Rebecca Sheets, PhD., who left me a
great project to expand upon; Rakesh Pandey, PhD., a
helpful and patient colleague in getting me acquainted
in the lab; Ananta Ghosh, PhD., for his considerable
technical expertise; and Lili Li, M.D., for her help and
expertise in automated sequencing studies. Since in
vivo studies comprised a large portion of this work, I
also gratefully acknowledge the valuable collaboration
with our colleagues at Ohio State University, Dr.
Lawrence Mathes, Dr. Andrew Phipps, and Kathleen Hayes.
A special thanks to Zoltan Tokes, PhD., and the
Department of Biochemistry and Molecular Biology for
giving me the chance to become a graduate student and
realizing an ambition I harbored for so long. A special
appreciation is acknowledged to all the faculty at USC
School of Medicine. They are a caring and supportive
group who taught me how to think creatively. A valuable
skill I had not expected.
I also wish to thank Sandy Hosteller,
Administrative Director, Office of Scientific Affairs,
and the many fellow graduate students I had the pleasure
of working with while in the School of Medicine Graduate
iii
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
student Association, especially Traci Newman and my good
friend, Giselle Lim, .. .what a team we made.
Lastly but most importantly, I wish to thank my
committee members, W. French Anderson, M.D., Michael H.
C. Lai, M.D., PhD., Michael Stallcup, PhD., and Daniel
Broek, PhD., who sacrificed their valuable time,
patience, and energy to make sure I developed as a
scientist. Most especially I thank my major advisor,
Pradip Roy-Burman, PhD., who oh so patiently guides,
nudges and coaxes his students along. I came to
graduate school with a narrow focus from research in
industry, but my eyes were opened to another world of
research which has only whetted my appetite for more.
Thank you all.
IV
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
TABLE OF CONTENTS
Page
Dedication........................................ 11
Acknowledgments................................... Ill
List of Figures.................................... vll
List of Tables.................................... vlll
Abstract.......................................... Ix
Chapter 1: Introduction........................... 1
Historical Background. ........ 1
Retroviral Model for Carcinogenesis 7
FeLV as a Model System............ 12
Thesis Rationale and Hypothesis.... 13
Experimental Strategies........... 14
Chapter 2: Feline Leukemia Virus Variants In
Experimentally-Induced Thymic LSA..... 16
Abstract........................... 16
Introduction....................... 18
Materials and Methods............. 21
Results............................ 28
Discussion......................... 45
Chapter 3: In Vivo Selection of Specific Recombinant
Feline Leukemia Virus Species......... 52
Abstract.......... 52
Introduction....................... 54
Materials and Methods............. 57
Results............................ 64
Discussion......................... 79
Chapter 4i Nucleotide Sequence Conversion of
Recombinant Feline Leukemia Viruses
(rFeLVs) to FeLV Subgroup B Viruses.... 83
Abstract........................... 83
Introduction....................... 85
Materials and Methods............ 88
Results............................ 91
Discussion......................... 101
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table of Contents (Continued)
Chapter 5: Recombinant Feline Leukemia Virus Variants
Demonstrate Altered In Vivo Cell Tropism 105
Abstract............................ 105
Introduction........................ 107
Materials and methods.............. Ill
Results............................. 121
Discussion.......................... 138
Chapter 6 : Epilogue
Summary and Conclusions............ 143
Future Directions................... 149
References.......................................... 155
VI
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Fig. 1.
Fig. 2.
Fig. 3.
Fig. 4.
Fig. 5.
Fig. 6.
Fig. 7.
Fig. 8.
Fig. 9.
Fig. 10.
Fig. 11.
Fig. 12.
Fig. 13.
Fig. 14.
Fig. 15.
Fig. 16.
Fig. 17.
LIST OF FIGORES
Page
Representative PCR Analyses for Tissue DMA
Isolated from an Infected Cat............. 32
Presence of MGPNL Sequence in All Clones
Examined................................... 39
Depiction of Epitope Sequence Mutation
in the Background of enFeLV Homology..... 41
PCR Detection of FeLV RNA Levels in Sera
Collected Over Infection Timecourse...... 43
Strategy for Sequencing the SU-encoding env
Gene of Recombinant-specific PCR Products. 63
Detection of rFeLV env in cats with
Experimentally-Induced LSA................ 65
PCR Detection of FeLV RNA Levels in Early
Sera Samples of Cat 4746-5................ 67
Schematic Representation of Recombination
Structural Motifs......................... 70
In Vivo Selection of rFeLV Species in Sera
4746-5 over Infection Timecourse......... 78
Schematic Summary of Nucleotide Sequencing
for the 1.2 Kb rFeLV env Gene............. 93
Sequence comparison of Amino Acid Chamges
between rFeLV Inocula and rFeLV Clones.... 100
Amino Acid Comparison of rFeLV/DB
Variants................................... 122
In Vitro Cell Tropism Studies............. 128
Indirect IFA Detection of FeLV Antigen.... 131
PCR Analysis of rFeLV/D8 Provirus in BM... 134
Deduced Amino Acid Comparison for rFeLV/D8
Variants in BM at 45 Weeks pi............. 136
Two-Dimensional Model for FeLV-B/GA SU.... 152
V I X
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
list of tables
Page
Table 1. Serologic Peurameters in Co-Infected Cats.. 30
Table 2. Detection of Recombinant FeLV env in
Tissues of Co-infected Cats............... 35
Table 3. Summary of 3 ' Recombination Junctions..... 71
Table 4. Summary of Nucleotide Substitutions
Conserved Between FeLV-B Isolates and
In Vivo-derived rFeLV Clones............ 95
Table 5. Summary of FeLV In Vitro 50%
Neutralization Titers ................... 126
Table 6. Summairy of PCR Analyses of Proviruses
in Litter I rFeLV/D8-infected Cats...... 133
v m
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
ABSTRACT
The primary objective guiding the study described
in this dissertation was to gain a better understanding
of the temporal process for in vivo evolution and
selection of recombinant FeLV species. The experimental
strategy was based upon the hypothesis that specific
recombinant FeLVs (rFeLVs) evolve in vivo such that
certain recombinants gain selective advantage, resulting
in their outgrowth from a genotypic mixture of
heterogenous but closely related recombinant species. A
logical extension of this hypothesis was that such
predominating recombinants may represent species with
greater leukemogenic potential. The specific approaches
to test this hypothesis were: (i) in vivo studies with a
defined mixture of rFeLVs either with or without FeLV-A
helper virus, (ii) focusing on the surface glycoprotein
(SU) portion of the env gene, comparison of nucleotide
sequences for the parental rFeLVs to those of in vivo-
derived rFeLVs, and (iii) subsequently, comparison of
those nucleotide sequences for the above in vivo-derived
rFeLVs to sequences for rFeLVs isolated from
experimentally—induced emd naturally-occurring thymic
tumor DMA, in order to determine whether a common
pattern of selection exists for certain rFeLV species.
In summary, these analyses demonstrated a total of 19
ix
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
point mutations which were scattered throughout the
endogenously-derived SU portion of the rFeLV env gene, a
region encoding approximately 80% of the N-terminal
portion of the SU, and which resulted in sequence
identity to exogenous FeLV-B isolates. Early selection
for rFeLV species containing this set of point mutations
in the SU region was followed by a more gradual
selection for rFeLV species harboring relatively greater
amounts of endogenously-derived SU sequence, resulting
in the predominance of rFeLV species with recombination
structural motifs similar to all natural FeLV-B
isolates. The work described herein supports the
hypothesis that certain rFeLV species do evolve and
predominate in vivo, resulting in rFeLV species bearing
considerable resemblance to exogenous FeLV-B isolates.
The presence of these specific rFeLV species in the
various thymic tumor DNAs employed in this work
potentially implicate these recombinant species as
proximal leukemogens in thymic lymphosarcoma disease
progression.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
CHAPTER 1
INTRODUCTION
Historical Background
Feline leukemia virus (FeLV) was first discovered
in a cluster of cats with lymphosarcoma (LSA) (Jarrett
et al., 1964) and was subsequently purified from plasma
of an infected cat with LSA (Kawakeuni et al., 1967).
The observed clustering of feline LSA cases in which
FeLV was isolated, suggested that an infectious agent
was responsible for feline leukemia. Evidence for the
horizontal or contagious transmission of FeLV was
confirmed when it was demonstrated that virus isolated
from cats with LSA could infect human cells, cemine
cells, as well as feline cells in tissue culture
experiments (Jarrett et al., 1969). Since FeLV was
shown to be a cause of leukemia in natural feline
infections, this challenged the oncogene theory proposed
by Huebner and Todaro (1969). The oncogene theory was
based primarily on observations with inbred laboratory
rodents, and stated that retroviruses were endogenous
viruses present in host cell genomes which were
vertically transmitted to offspring as a Hendelian
trait. The subsequent development of FeLV antisera and
an immunofluorescence assay for FeLV antigens, confirmed
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
that FeLV was contagiously transmitted (reviewed in
Hardy et al., 1976).
The ncune, feline leukemia virus, refers to its
historical association with lymphomagenesis; however,
FeLV has a broad spectrum of pathogenic effects ranging
from suppressive (immunodeficiency) to proliferative
(lymphoma) (Hoover and Mullins, 1991). The primary
route of viral transmission is through saliva and nasal
secretions in social exchange (reviewed in Roy-Burman,
1996; Hoover and Mullins, 1991; Mullins and Hoover,
1990). It is estimated that up to 50% of cats become
infected in their lifetime.
Currently, the clearest known correlates to FeLV
susceptibility are the cat^s age and its immune system
status at the time of exposure. During the first few
weeks after FeLV exposure, an interplay between virus
and host determines infection outcome. The dynamics of
FeLV infection can be divided into four categories of
host/virus relationships. Category 1 describes
approximately 30% of infected cats which develop
persistent viremia. Such persistently infected cats
present active viral shedding with diagnostic indicators
of ELISA( + ) and IFA(+) (for FeLV gag product antigens),
2uid negative virus neutralizing (VN) antibody. Most of
these cats (95%) will progress to FeLV-related disease
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
within three years post-infection (pi). Category 2
describes regressive or self-limiting infections in
which VN antibody is developed, infection is cleared,
and the cats become ELISA(-) and IFA(-). This occurs in
approximately 60% of infected cats, with most developing
no disease. Categories 3 and 4 describe atypical
diagnostic patterns of infection comprising the
remaining -10% of cases which may or may not progress to
disease. Anemia and immunodeficiency are the most
frequent disease manifestations within persistently
infected cats, while leukemia-lymphoma occur in
approximately 10-15% of all cases.
FeLVs are enveloped, positive-strand RNA viruses
with simple proviral genomic structures containing 5'-
LTR-gag-pol-env—LTR-3 ' organization following
integration into the host cellular DNA. The LTRs (long
terminal repeats) contain the regulatory sequences
necessary for transcription. The gag gene encodes viral
core structural proteins; pol gene products include the
viral functional proteins such as protease (PR), RNA-
dependent DNA polymerase or reverse transcriptase (RT)
euid integrase (IN); the env gene encodes viral envelope
proteins, surface glycoprotein (SU) and transmembrane
(TM) protein. Exogenous FeLV isolates are divided into
three interference subgroups based upon the ability of
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
one FeLV subgroup to prevent superinfection by a second
FeLV subgroup (Sarma and Log, 1973; rev. Hoover and
Mullins, 1991). This property of differentiating
between the three exogenous FeLV subgroups, is
attributed to sequence differences present on the virion
SU. Ecotropic FeLV subgroup A (FeLV-A) grows only in
feline cells, is present in all natural infections and
in general, is minimally pathogenic in the absence of
other variants (rev. Roy-Burman, 1996). However, one
FeLV-A variemt which is replication defective due to a
mutation in the env gene, induces an immunodeficiency
syndrome similar to AIDS in 100% of specific-pathogen-
free (SPF) cats when inoculated along with an FeLV-A
helper virus (Roy-Burman, 1996). Subgroups FeLV-B and
FeLV-c are polytropic viruses which infect feline cells
as well as cells from heterologous species. FeLV-C
virus is believed to have arisen from FeLV-A via genetic
mutations, primarily in the env gene, while FeLV-B
viruses are known to have arisen by recombination of the
FeLV-A env gene with endogenous FeLV-like sequences
inherited in the cat genome. FeLV-C is rarely isolated
(1%) and is always associated with FeLV-A. FeLV-C
isolates and the molecularly cloned FeLV-C/Semna are
able to consistently induce aplastic anemia in SPF cats
(Roy-Burman, 1996). FeLV-B isolates have a variety of
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
pathogenic properties, most notably their prevalence in
LSA (-60%) relative to healthy but infected cats (-30%)
(Jarrett et al., 1978; Tzavaras et al., 1990; Sheets et
al., 1993; Tsatsanis et al., 1994). However, whether
FeLV-B actually contributes to leukemogenesis is
currently unclear. It is possible that the presence of
FeLV-B in LSA may merely be a consequence of long
standing infection allowing more opportunity for
recombination events to occur. (This is of course an
active area of research in our leda.)
In addition to the three exogenous subgroups, there
are approximately 15 copies of endogenous FeLV (enFeLV)
proviruses per haploid genome in the domestic cat (Roy-
Burman, 1996). A comparison of full-length proviral
structures for the exogenous FeLVs and enFeLV show that
major distinguishing features include differences in LTR
sequences within the U3 region, frameshift and nonsense
mutations in the gag and gag~pol regions, as well as
scattered amino acid substitutions or deletions in the
enFeLV env gene (Soe et al., 1983; Soe et al., 1985;
rev. Roy-Burman, 1996). Transcription studies using 5'
enFeLV LTRs show the presence of promoter and enhancer
functions, but with variable transcriptional activity
due to the influence of negative cls-acting cellular
sequences indicating enFeLV expression is tightly
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
regulated in the host cell environment (Berry et al.,
1988; Roy-Burman, 1996).
As mentioned above, FeLV-B viruses are known to
have arisen from recombination within the FeLV-A env
gene such that variable amounts of the FeLV-A env,
starting from the 5' end, were substituted with
corresponding portions of env sequence from endogenous
retroviral elements. A comparison of env gene sequences
between enFeLV clones and isolates from all three
exogenous FeLV subgroups A, B, and C, show large regions
of highly conserved sequence interspersed with regions
of variable sequence (Kumar et al., 1989; Roy-Burman,
1996). Starting from the N-terminal portion of the
enFeLV SU sequence, several of these variable regions,
specifically variable regions II-V, show extensive
similarity to corresponding exogenous FeLV-B isolates,
but not to exogenous FeLV-A or FeLV-C isolates. The
enFeLV sequence subsequently diverges from all exogenous
FeLV subgroup isolates in the c-terminal variable
regions of the SU as well as in those varieible regions
present in the transmembrane (TM) domain of the env
gene, neuaely variable regions VI-X. In spite of the
fact that the enFeLV gag gene carries a number of
mutations, the high degree of nucleotide sequence
conservation in the pol gene, in addition to the
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
retention of an open reading frame in nearly the entire
pol region upstream of the enFeLV env gene, suggests
that homologous recombination may be initiated within
the pol gene and proceed into the adjacent env gene from
its 5' coding sequences (Pandey et al,, 1991; Roy-
Burman, 1996).
Retroviruses as models for carcinogenesis
Since cancer develops as a multi-step process, the
study of retroviral pathogenesis, which also follows a
complex, multifaceted progression, has provided insight
into the cumulative genetic events associated with
spontaneous human cancer (reviewed in Fan, 1994).
Retroviruses remain one of the best model systems for
the study of carcinogenesis in a whole organism, since
retroviruses induce tumors at very high efficiencies.
Following the initial isolation of avian retroviruses
(Rous, 1911) and the subsequent discovery of cellular
proto-oncogenes (Bishop and Varmus, 1984), retroviruses
have proven to be valuable tools in understanding
mechanisms of carcinogenesis. In regard to
carcinogenesis, retroviruses fall into two categories,
(i) acutely transforming viruses and (ii) chronic non
acute viruses (reviewed in Fern, 1994). Nonacute
retroviruses are standard replication-competent viruses.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
while acutely transforming retroviruses are replication-
defective viruses derived from nonacute retroviruses.
Acutely transforming retroviruses carry oncogenes which
allow them to rapidly induce tumors. First described in
the avian system involving Rous sarcoma virus (Rous,
1911), acutely transforming retroviruses induce cancer
through the process of transduction. A number of models
have been proposed to describe the transduction process.
All models involve proviral integration, and subsequent
co-packaging of one viral RNA strand along with a
cellular mRNA, followed by either template switch or
illegitimate recombination during reverse transcription
(reviewed in Neil et al., 1991). The ultimate result
however is that viral genomes of acute-transforming
retroviruses are altered such that a cellular proto
oncogene is incorporated into the viral genome.
Frequently, the cellular proto-oncogene replaces one or
more of the viral genes (often the env), resulting in a
replication-defective virion carrying a host gene which
has become a viral oncogene (v-onc). Integration of an
acutely transforming retrovirus leads to unregulated
expression of the oncogene product and loss of normal
cell growth and/or differentiation.
Chronic, non-acute transforming retroviruses do not
contain oncogenes and so must induce neoplasia via other
8
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
mechanisms. These possible mechanisms include
insertional activation and/or mutagenesis of cellular
proto-oncogenes. Insertional activation or mutagenesis
may occur when the viral DNA integrates within the
coding region of a cellular gene, resulting in disrupted
normal cellular gene expression while the viral LTR
promoter may drive overexpression of a truncated
cellular product. Alternatively, proviral integration
in a noncoding region adjacent to a cellular proto
oncogene may also drive unregulated expression of a
fully functional gene product via the retroviral LTR
promoter and/or enhancer elements. In addition to
transduction or insertional mutagenesis processes
associated with proviral integration into host cellular
DNA, there is also a correlation between the
pathogenicity of particular proviral env genes and
development of cancer (Sheets et al., 1993; Athas et
al., 1995; reviewed in Neil et al., 1991; and Fan,
1994).
In regard to the pathogenic potential associated
with the env gene, the well-studied murine leukemia
virus (MuLV) system has provided considerable insight
into those molecular genetic mechanisms potentially
involved in lymphoma induction (reviewed in Fan, 1996).
One such mechanism involves the binding of recombinant
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
viral SU protein to certain cell surface receptors. For
example, infection of an IL-2-dependent T-lymphoma cell
line with polytropic recombinant MuLV leads to IL-2-
independent growth. Another example is the highly
oncogenic MuLV variant, spleen focus-forming virus
(SFFV), which has been linked to erythroleukemia
induction. The replication-defective SFFV contains a
large deletion in the env gene, interrupting cleavage
between the SU and TM polyprotein product, resulting in
a fusion product which mimics erythropoietin (Epo)
growth factor by binding to the erythropoietin receptor
(EpoR) and thereby establishing an autocrine loop of
Epo-independent, unregulated cell growth.
Reminiscent of FeLV-B viruses, the env gene of
ecotropic MuLVs is able to recombine with endogenous
MuLV-like sequences in the murine genome, resulting in
polytropic recombinant viruses which have 5' portions of
the env gene derived from endogenous sequence and 3'
portions derived from the infectious ecotropic parent,
and which are able to bind cell surface receptors other
than those recognized by the parental ecotropic MuLV
(reviewed in Roy-Burman, 1996). Mink cell focusing-
forming viruses (MCF) are examples of such polytropic
recombinant MuLVs. However, unlike FeLV-B viruses, MCF
viruses require an additional genetic event to induce
10
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
thymie lymphoma as demonstrated in the ÂKR mouse model.
This additional event involves a separate recombination
event in which the MCF virus derives the U3 region of
its LTR from endogenous MuLV-related sequences. While
it is currently unclear as to how MCF viruses contribute
to disease progression, a model has recently been
described regarding the interplay of MuLVs and MCFs
(Lavignon and Evans, 1996). Evidence from their in vivo
study indicates polytropic MCF viruses arise from
ecotropic MuLVs in the pre-leukemia stage of disease
progression and that these polytropic MCF genomes are
initially pseudotyped within ecotropic virions, followed
by subsequent incorporation of polytropic genomes into
polytropic virions later in disease progression. From
this evidence, it can be speculated that pseudotyping of
polytropic genomes within ecotropic virions may
facilitate the initial entry into target cells of
polytropic genomes harboring recombinant env and LTR
sequences, increasing the chance for insertional
mutagenesis of a nearby cellular proto-oncogene.
Additionally, superinfection of ecotropic-infected cells
by polytropic virions may also play a role in tumor
development by increasing replication efficiency. Since
this study demonstrated the emergence of polytropic
virions prior to leukemia induction, this model also
11
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
provides evidence suggesting polytropic virions do
contribute to leukemogenesis and are not merely a
consequence of prolonged infection. Thus, mechanisms
employed by MuLVs and MCF viruses have provided a
foundation upon which the study of FeLV-mediated
pathogenesis can be built.
FeLV as a Model System
FeLV infection of cats has proven to be a powerful
model for understanding the pathogenesis of human cancer
and immunodeficiency for a ntunber of reasons. First,
FeLV is a natural pathogen endemic to an outbred
mammalian species and so may more accurately mimic human
disease processes as compared to those of inbred animals
such as the murine model (Mullins and Hoover, 1990).
Second, as a non-acute oncogenic retrovirus, FeLVs are
known to cause a variety of diseases, both proliferative
and suppressive, providing a versatile model for diverse
studies (Hoover and Mullins, 1991). Third, FeLV has
been studied since the 1960's, so it represents a well-
characterized retroviral model for natural pathogenesis
which has been important in elucidating human retroviral
disease (Hardy, 1990). Fourth, similar to human
retroviruses, FeLV is genetically variable, particularly
the FeLV-B viruses, and subtle differences in viral
12
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
genotypes may have significant relevance to disease
progression (Neil et al., 1991).
Thesis Rationale and Hypothesis
As described earlier, FeLV-B viruses are known to
be overrepresented in LSA as compared to infected but
otherwise healthy cats, suggesting a role for these
viruses in disease progression. FeLV-B viruses are also
known to be recombinants between portions of the
exogenous FeLV-A env gene and corresponding endogenous
FeLV-like elements, which can lead to a mixture of
variant recombinant env genotypes, depending upon
recombination sites employed (Sheets et al., 1992; Roy-
Burman, 1996). Some in vitro-generated recombinant
mixtures have been characterized previously in our lab,
and similar recombinant env gene motifs have been
observed by our lab in naturally-occurring LSA (Sheets
et al., 1992; 1993). However, scattered differences
exist between env gene nucleotide sequences of in vitro-
generated recombinant mixtures and those of natural
FeLV-B isolates. While studies using natural FeLV
isolates have been valuable in identifying viral
pathogenic determinants, this approach has not provided
a clear understanding of how emd which viral genotypes
may evolve in vivo since either the genotype of the
13
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
originally infecting virus was unlcnown or only single
FeLV-B isolates were employed. In addition,
considerable evidence suggests that nucleotide sequence
variations present in the SU portion of the env gene
play a role in retrovira 1 -mediated disease progression
(rev. in Fan, 1996 and in Roy-Burman, 1996). The
hypothesis which led to the experiments described herein
states that specific recombinant FeLVs (rFeLVs) evolve
in vivo such that certain recombinants gain selective
advantage, resulting in their outgrowth over a mixture
of viruses. Such predominating recombinants may
represent species with greater leukemogenic potential.
Experimental Strategies
The overall goal of this work was to gain an
understanding of the process of in vivo evolution for
rFeLVs in an infected host. Focusing on the SU region
of the env gene, the specific aims in this work were:
(i) to address the disparity between env gene nucleotide
sequences for in vitro-generated rFeLVs emd those of
reported FeLV-B isolates, and (ii) to determine whether
certain recombinant species predominate in vivo from an
initial mixture of recombinant species. To achieve
these aims, a mixture of in vitro—generated recombinants
were inoculated into neonatal cats with or without FeLV-
14
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
A helper virus and molecular genetic analyses of the SU-
encoded portion of the env gene were employed to
characterize the recombinant proviruses subsequently
derived from the infected cats. Finally, the above
results were then compared to those obtained from two
additional cats with experimentally-induced thymic
tumors and one cat with a naturally-occurring thymic LSA
to determine whether similar observations could be
extended to suggest a common pattern of selection may
exist for certain rPeLV species in different infected
hosts.
15
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
CHAPTER 2
Feline Leukemia Virus Variants in Experimentally-Induced
Thymic Lymphoseurcomas
ABSTRACT
This study was initiated to evaluate the in vivo
infectivity and pathogenicity of a group of recombinant
feline leukemia viruses (rFeLVs) previously generated by
in vitro-forced recombination between a FeLV subgroup A
virus (FeLV-A) and an endogenous FeLV (enFeLV) envelope
(env) element. To determine infectivity of these
rFeLVs, neonatal cats were inoculated with rFeLVs alone
or in combination with FeLV-A. Results showed the
recombinant viruses were able to efficiently replicate
in vivo only when administered along with FeLV-A. Of
six co-infected cats, three developed thymic
lymphosarcomas, one severe aplastic anemia, and two
cachexia and depression; all were viremic and
seroconverted shortly after inoculation. While both
virus types were detected in virtually all tissues
examined from these tumor-bearing cats, there was a
particularly noteworthy sequence change in the rFeLVs.
It is known that exogenous FeLV isolates carry a
conserved neutralizing MGPNL epitope in the middle of
16
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
the surface glycoprotein (SU) domain encoded by the env
gene. In contrast, parental recombinant viruses used to
inoculate these cats harbored the enFeLV-derived MGPNP-
encoding sequence at this position. However, all in
vivo-derived recombinants displayed a C-»T transition
mutation within the enFeLV nucleotide sequence
background resulting in a Pro217Leu amino acid change
corresponding to the exogenous MGPNL epitope, while the
env-encoded flanking sequence was that of the parental
recombinant virus. These results suggest that viruses
harboring the MGPNL epitope have an in vivo
proliferative advantage. The data also provide an
explanation for the conservation of this epitope in
exogenous FeLVs despite the existence of variant forms
in enFeLV proviral elements with which they can
recombine.
17
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
INTRODUCTION
Within the three exogenous, horizontally-
transmitted FeLV subgroups A, B and C (Sarma and Log,
1971, 1973), FeLV-B viruses are known to be generated by
recombination of FeLV-A env sequences with corresponding
but varied endogenous FeLV (enFeLV) proviral elements
(Elder and Mullins, 1983; Kumar et al., 1989; Overbaugh
et al., 1988; Sheets et al., 1992, 1993; Stewart et al.,
1986; Rohn et al., 1994). Previously we obtained
recombinant FeLVs (rFeLVs) by a process of induced
recombination between a molecular clone of FeLV-A/6IE
(Donahue et al., 1988) and env sequences from an enFeLV
clone, CFE-6, (Soe et al., 1985) in transfected feline
H927 fibrosarcoma cells (Sheets et al., 1992). The
recombinemt viruses derived from this in vitro-forced
recombination process had large portions of the FeLV-A
env gene, beginning from the 5' end, replaced with
sequences from the endogenous element, displaying
polytropic cell tropism similar to FeLV-B viruses, but
different from the ecotropic nature of FeLV—A viruses.
Although in vitro replication properties and overall
sequence structure of the env gene for these generated
recombinants were very similar to FeLV-B (Kumar et al.,
1989; Pandey et al., 1991; Sheets et al., 1992), there
were scattered sequence variations, including a change
18
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
in the nucleotide sequence encoding a prominent MGPNL
pentapeptide which is highly conserved in the middle of
the SU surface glycoprotein in exogenous FeLVs (exPeLVs)
and which represents a binding site for a neutralizing
monoclonal antibody, C11D8 (Elder et al., 1987; Grant et
al., 1983). In most rFeLVs, the portion of the env gene
backbone acquired during the recombination event
contained the endogenous-like sequence, which encodes a
Pro217 rather than the exogenous Leu217 of the mature
SU.
Early epidemiological surveys reported that cats
with lymphosarcomas (LSAs) or leukemias had a higher
incidence of harboring FeLV-B than asymptomatic FeLV-
infected animals (Jarrett et al., 1978). More recently,
direct evidence based on molecular genetic analyses was
obtained for the prevalence of FeLV-B-like viruses in
naturally-arising FeLV-related LSA (Levy et al., 1993;
Sheets et al., 1993; Tsatsanis et al., 1994). Thus, it
was of interest to examine the in vivo infectivity and
pathogenicity for a group of recombinants we previously
generated through in vitro recombination between FeLV-A
and enFeLV (Sheets et al.,1992). Additional impetus for
the study was derived from the observation that similar
recombinemts between FeLV-C and enFeLV could augment the
development of FeLV-C-related erythrocyte aplasia in
19
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
inoculated cats or alter in vivo tissue tropism (Mathes
et al.,1994; Chakrsüaarti et al., 1994), as well as from
the finding that certain types of recombination could
yield FeLV env polyproteins defective in intracellular
processing and transport properties (Bechtel et al.,
1994). In the work described here, we present evidence
that recombinant viruses, analogous to certain FeLV-B
natural isolates, appear to replicate and establish
viremia in vivo only when administered in association
with FeLV-A, emd coexist with FeLV-A as proviruses in
these inoculated cats. Additionally, an intriguing
possibility is indicated by our finding that transition
mutations in the nucleotide sequence encoding the
endogenous M6PNP epitope are altered to that of the
exogenous FeLV-B-like nucleotide sequence encoding the
MGPNL epitope. Expression of this exogenous Pro217Leu
residue may be necessary to gain in vivo replication
efficiency. This finding may explain why there is
strict conservation of the Leu217 residue in the mature
SU of exFeLVs despite the existence of variant forms in
enFeLV proviral elements.
20
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
lU^TERIALS AND METHODS
Viruses
Recombinant FeLVs were previously produced by
coexpression of a plasmid containing infectious,
minimally pathogenic FeLV-A/61E (Donahue et al., 1988)
along with a plasmid containing different chimeric
enFeLV constructs, and selected for their ability to
propagate in human cells resistant to infection by the
parental FeLV-A (Sheets et al., 1992). The recombinant
virus pool, previously designated A x r6env-LTR (Sheets
et al., 1992), now abbreviated and henceforth referred
to as rFeLVs, was examined for in vivo biological
properties. The rFeLVs contained multiple crossover
sites in the env gene; major subpopulations previously
identified in this pool were recombinants at A, B, E, F
and G sites, such that the length of enFeLV env
substitution at the 5' end of env increased from site A
(enFeLV CFE-6 nucleotide env position 586 in the mature
SU) to site G (CFE-6 env position 856 in the mature SU),
and that sites E to G reside downstream of the exFeLV
MGPNL-encoding sequence while site D (CFE-6 env position
637) resides within the epitope-encoding sequence
(Sheets et al., 1992; Kumar et al., 1989). The
recombinant virus mixture was initially generated in
feline fibrosarcoma H927 cells, selected in human
21
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
fibrosarcoma HT1080 cells and subsequently expanded in
HT1080 cells (Pandey et al., 1991; Sheets et al., 1992).
Virions released in culture supernatant fluids were
monitored by enzyme-linked immunosorbent assay (ELISA)
for presence of the virion p27 capsid protein. The
uncloned FeLV-A/Rickard plasma inoculum was taken from a
cat challenged at 9 weeks of age and collected at 14
weeks of age. This challenge virus represented the
eighth in vivo passage of FeLV-Rickard, which was
originally derived from a tumor homogenate and
determined to be a FeLV-A subtype by direct interference
assay. The titer of all infectious viruses was
determined by the clone 81 S+L- focus induction assay of
Fischinger et al. (1974). One infectious unit of FeLV
was represented by one focus-forming unit (FFU).
Animal studies
Four newborn, specific pathogen-free (SPF) kittens,
one to two days of age, from the breeding colony of Ohio
State University, were inoculated intraperitoneally with
1125 FFU of cell-free virus prepared from cell culture
fluid of rFeLV-infected HT1080 cells. The same virus
preparation was also used to inoculate three weanling
(8- to 10-week old) SPF cats, which received the cell-
free virus by d combination of intravenous and
22
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
intraperitoneal infections. The older cats also
received 10 mg/kg methylprednisolone acetate in an
attempt to facilitate the esteiblishment of viremia. In
order to more accurately mimic natural conditions
observed in FeLV-B infection of cats, six neonates
(4746-1 to 4746-6) were inoculated with 4500 FFU of the
rFeLV mixture along with 90 FFU of the FeLV-A/Rickard
plasma preparation. Blood seunples were collected at
biweekly intervals post-inoculation (pi) except the
first interval, which was omitted for the newborn groups
due to their small size. All inoculations and blood
specimen collections were performed under ketamine
anesthesia (25 mg/kg). Hematologic parameters were
determined at each sampling interval when amtigenemia
and antibody responses were also measured. At necropsy,
various tissues were collected for histopathologic
examinations, molecular genetic analyses for provirus
sequences, and FeLV expression status.
Antigenemia was determined by a commercial ELISA
for the p27 capsid protein (Synbiotics Corp., San Diego,
CA) in plasma. ELISA values were determined using an
ELISA plate reader (Molecular Devices) at 650 nm
wavelength; values were expressed as optical density,
and were not corrected for background. Positive and
negative controls included with each ELISA had values of
23
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
0.0387 ± 0.0042 and 0.421 ± 0.119, respectively. The
positive control had a calculated p27 protein content of
121 ug/ml.
The FeLV antibody response was measured by a live
cell indirect immunofluorescence assay on the FL-74
feline T-cell lymphoma cell line chronically infected
with FeLVs representing each of the A, B, and C
subgroups (Essex et al., 1971; Mathes et al., 1994).
Titers were expressed as the reciprocal of serum
dilution. Sera from uninfected SPF control cats gave
titers of < 2. Each assay contained a positive control
serum with a titer of 1024. Assays were discarded when
the positive control serum titer deviated more than one
dilution from the expected value.
Genomic DMA extraction and detection of proviruses
High molecular weight genomic DMA was extracted
from tissues of infected cats which developed LSA as
previously described (Sheets et al., 1993). Tissues
included: tumor, bone marrow, lymph node, thymus, spleen
and brain. Primer set PRB-1 and PRB-2 was used to
specifically eunplify recombinant env sequences from
genomic DMA by polymerase chain reaction (PCR) using a
total of 40 cycles. Similarly, exogenous FeLV-A-
specific env sequences were amplified by use of another
24
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
primer set, RB-59 and PRB-2 (Sheets et al,, 1993). The
resultant PCR products were run on agarose gels and
stained with ethidium bromide.
Total RNA isolation and reverse transcription-PCR (RT-
PCR) amplification
Total RNA isolation from sera and tissue samples
was performed using Clontech's total RNA isolation kit
(Palo Alto, CA). RT reaction was performed using
Moloney murine leukemia virus reverse transcriptase (M-
MLV RT) (Gibco-BRL, Gaithersburg, MD) and oligo(dT)
primer (Promega, Madison, WI) to synthesize cDNA. For
each sample studied, 5 ug total RNA was used in a 50 ul
RT reaction. RT-PCR without RNA or without reverse
transcriptase enzyme were included in each experiment,
serving as negative controls. For detection of FeLV-A-
specific fragments in sera, 5 ul cDNA was used in the
first round of amplification employing the 5' primer
RB59 and 3' primer RB52 (Sheets et al., 1993) in 35 PCR
cycles. One ul PCR product from the first round of
eunplification was then amplified in a second round of 35
amplification cycles using the nested primers, 5' RB326
and 3' RB17 (Sheets et al., 1993), giving a 963 bp
product. The RB326 primer is oriented in the sense
direction identical to positions 6368 to 6391 of the
25
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
published FeLV-A/61E strain sequence (Donahue et al,,
1988). To detect the presence of rFeLV fragments in
sera, first-round amplification utilized 5' primer RB56
and 3' primer RB52 (Sheets et al., 1993) in 35 PCR
cycles. Second-round amplification employed the
recombinant-specific nested primers RB53 (S') and RB17
(3') (Sheets et al., 1993) using 35 PCR cycles, and
yielding an 888 bp fragment. First-round amplification
using tissue cDNA utilized the same primers as employed
with the sera samples. Second-round amplification used
RB326 (S') and RB19 (3') (Sheets et al., 1993) for FeLV-
A-specific expression, yielding a 1,533 bp product.
RB53 (S') and RB19 (3') primers were used for rFeLV-
specific expression, yielding a 1,377 bp product. PCR
products from the second-round of amplification were
analyzed by 2% agarose gel electrophoresis and DNA bands
visualized by ethidium bromide staining.
Cloning and sequencing of PCR products for rFeLV—
specific env fragments
PCR products generated from tissue DNA with the
PRB-l/PRB-2 primer set were cloned into the TA-cloning
vector (Invitrogen, San Diego, CA). White colonies were
screened for the 1.5 kb insert by restriction analyses,
and positive clones were manually sequenced as reported
26
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
earlier (Pandey et al,, 1991; Sheets et al., 1992). The
sequence for each of the recombinant env clones was
compared to those of parental CFE-6 enFeLV (Kumar et
al., 1989) and exogenous FeLV-A/6 IE ( Donahue et al.,
1988).
27
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
RESULTS
Infectivity and pathogenicity studies
In the three weanling SPF cats and four neonatal
SPF cats inoculated only with rFeLVs, none became
viremic or seroconverted within twelve weeks of
observation. Results indicated that this pool of
rFeLVs, like most natural isolates of FeLV—B (Jarrett
and Russell, 1978; Rojko et al., 1979; Hoover et al.,
1980), failed to induce viremia in any of the weanling-
age or newborn cats (data not shown). In the six
neonates co-challenged with both FeLV-A/Rickard and
rFeLVs, all six were antigenemic at the 4-week pi
sampling point as determined by ELISA for viral capsid
protein, and all had seroconverted as determined by
live-cell-indirect immunofluorescence assay for
detection of FeLV antibody levels ( Table 1 ). As noted
in Table 1, antigenemia generally increased in all six
cats, with levels reaching maximal at approximately 6
weeks post-infection (pi) and generally declining by 40
weeks pi. Antibody titers were maximal at 8-10 weeks pi
and subsequently declined to below detectable limits of
the assay by 20 weeks pi. Serologic results observed in
the six cats were comparable to those reported in early
studies (Jarrett and Russell, 1978; Jarrett et al.,
1978; Russell and Jarrett; 1978). Three (4746-1, 4746-
28
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2, and 4746-5) developed thymic LSA and were euthanized
to collect tumors and other tissues at 44, 75, and 90
weeks pi, respectively. One cat (4746-3) developed
severe aplastic anemia and was euthanized at 37 weeks
pi. The remaining two cats (4746-6 and 4746-4)
developed cachexia and depression, and were euthanized
at 19 and 78 weeks pi, respectively.
From historic challenge studies, of 33 cats
inoculated at 8-10 weeks of age with FeLV-A/Rickard
inoculum alone, 30 of the 33 weanling cats developed
chronic, life-long viremia (unpublished data; personal
communication, L.E. Mathes and P. Roy-Burman). Of those
chronically viremic cats observed from 26 weeks pi to 52
weeks pi, 2 of 19 developed lymphomas, while 2 of 5 cats
held greater than 52 weeks developed tumors.
29
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
7}
CD
■D
O
Q .
C
g
Q .
■D
CD
C/)
W
o'
3
0
5
CD
8
■D
( O '
3 "
1
3
CD
3 .
3"
CD
CD
■D
O
Q .
C
a
O
3
■D
O
CD
Q .
■D
CD
( / )
( / )
TABLE 1
Serologic Parameters Observed in Co-infected Cats
Cat #
4 6 8
Plasma p:
10
7 levels*
20
/antibody t
30
iters^ at
40
weeks post
50
-inoculât:
60
on
72 80 90 100
4746-1 .3 7 3/1 2 8 .781/64 .784/256 .711/4096 .403/<2 .874/HD ° .408/<2 .337/<2 .153/<2 .230/<2 B(75>**
4746-2 .3 9 0/1 2 8 1.0 5/1 2 8 .821/256 .428/64 .436/<2 .812/HD .285/<2 .386/<2 201/<2 .217/<2 .201/<2 .493/<2 B t90)
4746-a .6 5 0/1 2 8 .809/128 .662/256 .433/1024 .341/<2 ND/ND B(37)
4746-4 .3 9 5/1 2 8 .917/128 .782/256 .519/128 .267/<2 .858/ND .235/<2 .342/<2 .276/<2 .194/<2 B(78)
4746-5 .6 0 0/1 2 8 .936/256 .780/256 .320/128 .315/<2 .750/ND .372/<2 B(44)
4746-6 .404/128 .884/128 .714/256 .483/32 B (19)
*FeLV p27 capsid protein in the plasma was datectad by tha ELISA mathod.
^Antibody titers in the sera were determined by indirect immunofluorescence on cat lymptacma cell line, EL-74, which is chronically infected
with all three subtypes of FeLV. Refer to text for further details and interpretation of data in table.
'^Not determined.
W
O
Detection of both FeLV—A and recombinant emr genes in
tissue DNA of cats with LSA
The PCR strategy was similar to what was previously
described (Sheets et al., 1993). Specifically, PCR
amplification detected rFeLV env sequences in genomic
DNA extracted from tumor or other tissues of the
infected cats using a rFeLV-specific primer set PRB-1
and PRB-2. Similarly, FeLV-A proviral env sequences
were detected with FeLV-A-specific primer set RB59 and
PRB-2. PCR results for tissue DNAs from one tumor-
bearing cat are illustrated in Fig. 1. Both FeLV-A and
rFeLV env sequences were detected in the DNA of all the
various tissues tested. Expected-size PCR products of
about 1.45 kb for recombinant env and 1.75 kb for FeLV-A
env were generated (Sheets et al., 1993). Similar
results were obtained with tissue DNA from the other two
tumor—bearing cats, 4746—1 and 4746-2 (data not shown).
The presence of FeLV-A and recombinant proviruses in
both tumor and normal tissue DNA suggested that rFeLVs,
like FeLV-B, could be infectious in vivo when inoculated
along with FeLV-A, and were able to manifest broad
tissue tropism.
31
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
•1.5 Kb
1 75 Kb
FIG. 1. Representative data of PCR analyses for
tissue DMA isolated from infected cat 4746-5. (A) PCR
products using rFeLV-specific primer set PRB-1 and PRB-
2. DNA in lanes: 1, bone marrow; 2, spleen; 3, brain;
A, thymus; 5, lymph node; 6, lymphosarcoma ; 7, a
positive control (A x r6env-LTR-infected HT1080 cells);
8, a negative control (uninfected SPF cat DNA); 9, water
blank; and N, DNA markers. Arrow indicates position of
the approximately 1.5 kb band in marker which closely
corresponds to the expected band size from recombinant
proviruses. The rFeLV-specific primer set has
previously been shown to not amplify FeLV-A (Sheets et
al., 1993). (B) PCR products with FeLV-A-specific
primer set RB-59 and PRB-2. Similar analyses of tissue
DNAs as in panel A with DNA in lanes 1-6, 8, 9, and H
being the same as (A); 7, a positive control (FeLV-A-
infected feline H927 fibroblast cells). Arrow indicates
the position of 1.75 kb expected product. This FeLV-A-
specif ic primer set was shown to not amplify pBHMl (a
FeLV-B plasmid), nor uninfected H927 cells (Sheets et
al., 1993).
32
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Identif ication of recombinant proviral crossover sites
in tissue DMA
PCR products generated to detect recombinant env
genes were cloned into the TA cloning vector and
sequenced by priming with Ml 3 universal forward and
reverse primers as well as with oligonucleotides
representing conserved sites internal to the insert. In
a previous report, several recombination points. A, B,
E, F and G, were found in clones of PCR products from
rFeLV-infected HT1080 cells (Sheets et al., 1992).
However, this earlier study was limited since only eight
clones were examined, and the possibility remained that
all potential variants were not identified. In the
present study a larger number of clones were sequenced
after eunplif ication of tissue DMA from the infected
cats. A total of 47 clones were analyzed, representing
two to three different tissues from the three LSA-
positive cats. Those results are summarized in Table 2.
Recombinant species with crossover sites at A, B or E
were not readily detected in the tissues analyzed.
Although no definitive claim could be made based upon
PCR and cloning analyses, it appeared that some species
with longer enFeLV env substitution, (e.g., recombinants
at sites F and G, ) could have in vivo replication
advantage over other species such as sites A, B or C
33
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
recombinants. However, this could not be the only
criterion for replication efficiency, since the site D
intermediate species, which was not originally detected
in the inocula was readily observed in tissues of two of
the three cats. Additionally, another variant
recombination site, located slightly downstream of site
D, which we termed here as D', was detected in one tumor
DNA. Of all the recombinants, site G recombination type
was present in all tissues of the three cats examined in
this work, and appeared to be a primary, if not
exclusive, component in a single cat (4746-5) (Table 2).
Since all three tumor, bone marrow and spleen tissues of
this cat contained this site G recombinant, at least in
this animal the site G recombinant species apparently
had replication advantage over other species.
34
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
CD
■D
O
Q .
C
g
Q .
■D
CD
C/)
W
o'
3
0
3
CD
8
■D
( O '
3"
1
3
CD
3 .
3"
CD
CD
■D
O
Q .
C
a
O
3
■D
O
CD
Q .
TABLE 2
D e te c tio n o f R e c o m b in a n t FeLV SU S equences i n T is s u e s o f In f e c t e d C a ts
C a t
Number
T is s u e
Exam ined
T o t a l
# o f C lo n e s
Sequenced
C ro s s o v e r S i t e
(Num ber o f C lo n e s )
4746-1 Tumor 8
D(l); D’(2); F(3); G(2)
Bone Marrow 8 0(1); G(7)
4746-2 Tumor 7 0(3); F(2); G(2)
Bone Marrow 6 0(1); G(5)
4746-5 Tumor 6 G(6)
Bone Marrow 6 G(6)
Spleen 6 G(6)
■D
CD
( / )
( / )
w
Mutation detected in nucleotide sequence encompassing
the enFeLV MGPNP pentapeptide
All sequenced exFeLVs are known to contain a MGPNL
pentapeptide epitope near the middle of SU. This
pentapeptide, a binding site for a monoclonal antibody,
C11D8, was found to be varied (MGPNP or MGPDP) in enFeLV
loci present in cat genomic DNA (Kumar et al., 1989;
Sheets et al., 1993) and expressed as env message in
feline lymphoid cells (McDougall et al., 1994). The
rFeLV pool employed in this study was shown to harbor
the enFeLV MGPNP sequence at this position in those
clones analyzed (Sheets et al., 1992). The site D
recombination point, which resides within the nucleotide
sequence encoding this pentapeptide, was previously
found to also be a frequent site of mutation (Sheets et
al., 1992; 1993). Earlier studies demonstrated that
alterations in this region allowed neutralization escape
by the C11D8 monoclonal antibody (Nicolaisen-Strouss et
al., 1987; Sheets et al., 1992). The sequence of this
region was, therefore, scrutinized for potential
alterations in the recombinant species propagated in
vivo. The finding was striking. Of 47 PCR clones
sequenced, 41 harbored a nucleotide change resulting in
conversion from MGPNP to MGPNL. The six clones without
amino acid sequence change already contained the MGPNL
36
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
epitope because the site D recombination point is
located just upstream of the pentapeptide's C-terminal
amino acid (i.e., Pro217 in the mature SU) encoded in
the CFE-6 sequence. Results illustrated in Fig. 2 show
that recombinants at sites D', F and G, with crossover
junctions downstream of the pentapeptide epitope,
exhibited a C-*T transition mutation relative to enFeLV
in the nucleotide sequence encoding the last amino acid.
This transition mutation is consistent with conversion
of the Pro217Leu sequence in all clones examined.
The issue of mutation occurring in the background
of parental CFE-6 sequence is further illustrated in
Fig. 3. Site F and site G recombinants were chosen to
highlight this point since they contain an extensive
amount of enFeLV env substitution. While homology to
endogenous sequence was retained up to the crossover
junction in these two illustrated cases, the Pro217Leu
substitution was invariably manifested. Results were
similar with other F and G recombinants sequenced.
37
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
FIG. 2. Presence of HGPNL sequence in all clones
examined. Comparison of sequence encompassing the HGPNL
pentapeptide epitope is shown for each of the
representative classes of recombinants with crossover
sites D, D', F and G, which were detected in tissues of
the tumor-bearing cats. (A) Sequence alignment is
relative to enFeLV clone, CFE-6, with numbering based
upon the mature SU sequence as reported (Kumar et al.,
1989). (B) Sequence changes relative to CFE-6 are
shown, with identity indicated by (. ). The rFeLV env
gene, with locations of the various identified crossover
sites in the mid-region of the SU-encoding domain, are
illustrated for reference. Recombination site D falls
within the pentapeptide region but sites F and G reside
farther downstream in the C-terminal half of SU (Sheets
et al., 1992). Site D', first detected in this study,
appears to be just a few nucleotides downstream of site
D.
38
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Q>
Pi u
t
m
39
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
FIG. 3. Illustration of sequence mutation in the
background of enFeLV homology. To emphasize this point,
only recombinants having crossover points at site F and
G are presented due to retention of extensive enFeLV env
sequences. Nucleotide variations in FeLV-A/61E or the
recombinants (site F, site G) as compared to the enFeLV,
CFE-6, are presented. Nucleotide residues identical to
CFE-6 are indicated by (. ) and deletions by (-). The
MGPNL or MGPNP region is shaded, and the difference in
the fifth amino acid between parental FeLV-A and enFeLV
is presented in bold typeface. The positions of
recombination sites F and G are marked by bars above the
relevant sequence. The env sequence, beginning from the
top left and ending at the bottom right, represent
published sequence number 6662-6889 of 61E (Donahue et
al., 1988) and number 770-1030 of CFE-6 (Kumar et al.,
1989).
40
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
770
I
C F E - 6 (?P A A T nA fT :aT T arr« ira 7 n n Y 7 a fM rr M '^1T ^r''% gj|w nM
...........
Site F
Site G
CFE-6 rrrnraTrraaaraaTrTr&aaTa«ammrrmQaaTaATaccTcaccaTCCCCaAGGcaaf:
61E C.&......... C..G___ AA. . .GGCGA.C. .GAGG..... ACG. .T
Site F
Site G
CFE-6 GGAGGCACCCCAGGXAXAACTCITGTZAATGCCTCCATTGCCCCICXAAGCACCCCTGTC
61E JL* * G # • • • #A#——————————————————G* * TG * ■ • • • •
Site F
Site G
F
I------------------------------- 1
CFE-6 anrrrrficaarypTfrnaAarryraTarMgarafifiaaa'pafigTTOaTOaaTTTAGTfirAfififig
61K ...A. ..TGG......... G. .T.... C...G...................A. A--
Site F .....................................................A. A. . .
Site G ............................................................
CFE-6 ACATAICTAACIZTAAAT6TC
61E C...G.C.......C.
Site F C G.C...... C.
Site G .........G.C......C.
41
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
viral SNA detection in sera and tissues of infected cats
A representative experiment demonstrating the
levels of FeLV viral SNA observed in sera from infected
cat 4746-5 is presented in Fig. 4. The top panel
depicts the rFeLV-specific SNA pattern while the bottom
panel shows FeLV-A-specific SNA detected in the sera.
FeLV-A viral SNA was detected at the first collection
time point of 4 weeks pi. Although the band was faint,
and not readily visible in this reproduction (bottom
panel), repeated analysis confirmed its presence at 4
weeks pi. The presence of FeLV-A viral RMA in the serum
appeared to vary with time since no band was present at
8 weeks pi, while at 12 weeks pi a strong band was
visible. This cyclical pattern for FeLV-A continued,
with low levels detected at 16 to 24 weeks, followed by
higher levels observed at 30 to 43 weeks pi. At the
last collection time point of 44 weeks pi, a low level
was again detected. In contrast, rFeLV viral RNA levels
shown in the top panel (Fig. 4) indicated a relatively
consistent pattern. A band was first observed at 8
weeks pi, and the rFeLV viral RNA level appeared to
remain fairly uniform throughout the 44-week time
course. While this RT-PCR was not quantitative per se,
the analysis was designed to be semi—quantitative in
that the total RNA for each serum sample was
42
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
rFeLV RNA
W M h s p i
888 bp
FeLV-A RNA
W M k s p i
FIG. 4. PGR detection of FeLV RMA levels in sera
collected from cat 4746-5 from 4 to 44 weeks pi. The
top panel (rFeLV RNA) depicts an 888 bp product from
recombinant-specific nested PCR amplification. The
bottom panel (FeLV-A RNA) shows a 963 bp product from
FeLV-A-specific nested PCR amplification. In both
panels, the lanes contain the following samples reading
from left to right: 4 weeks pi, RT(-) ; 4 weeks pi,
RT(+); 8 weeks pi, 12, 16, 20, 24, 30, 38, 43, 44 weeks
pi and 44 weeks pi, RT(-), respectively. The lack of a
band in the lane containing the FeLV-A/Rickard plasma
inoculum demonstrate these are recombinant-specific
primers. Recombinant FeLV (rFeLV) cDNA served as
positive control, followed by uninfected H927 cDNA as
negative control, water, emd 100 bp DMA ladder,
respectively. In the bottom panel the positive control
is FeLV-A/61E (FeLV-A RNA). All other lanes in the
bottom panel contain the same samples in the order
described in the top panel.
43
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
standardized to 5 ug of total RNA in the RT reaction,
and the cDNA used in the PCR reaction was also
equivalent for each seunple. This effort ensured
relative viral RNA levels were reflected for each serum
sample. In addition, the same cDNA preparations were
used in both experiments depicted in Fig. 4. Moreover,
the consistent levels of rFeLV RNA observed from 8 to 44
weeks pi in Fig. 4 (bottom panel) serve as an internal
control for the demonstration of equivalent RNA levels
present in this series of experiments. Detection of
FeLV viral RNA in the FeLV-A/Rickard plasma inoculum is
also shown in Fig. 4. No rFeLV RNA was detected in the
plasma inoculum (top panel), while a strong FeLV-A-
specific band was observed in this inoculum (bottom
panel).
FeLV-A expression and rFeLV expression were
detected in cDNA from all tissues evaluated (data not
shown). Tissues analyzed for FeLV expression were
tumor, lymph node, spleen, bone marrow, thymus, and
brain.
44
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
DISCUSSION
FeLV-A is seen in all natural infections and its
presence is believed to be required for the in vivo
infectivity of FeLV—B or FeLV-C (Jarrett et al., 1978;
Jarrett and Russell, 1978; Sanaa et al., 1978). Studies
in which the molecularly cloned FeLV-B/Geurdner-Amstein
strain is administered alone indicate its inability to
induce viremia in weanling age or newborn cats even in
the presence of corticosteroids (Rojko et al., 1979;
Hoover et al., 1980). Similarly, while the FeLV-C/Sarma
strain is capable of infecting newborn cats, it fails to
induce viremia in animals beyond one week of age (Riedel
et al., 1986). In the present study, the lack of in
vivo infectivity as observed with our in vitro-generated
rFeLVs, and which exhibited genetic similarity to FeLV-B
viruses, is consistent with the reported nature of FeLV-
B infectivity. As noted above, our results also suggest
efficient in vivo proliferation of rFeLVs appear to be
achieved only in the presence of FeLV-A helper virus.
These results, taken together with earlier findings of
genetic make-up and expemded in vitro cell tropism for
such recombinants (Overbaugh et al., 1988; Pandey et
al., 1991; Sheets et al., 1992) further corroborate the
conclusion that FeLV—B viruses arise in nature through
45
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
genetic recombination between enFeLV and exogenous FeLV-
À.
There is evidence that FeLV-B viruses are
associated with leukemogenesis (Hardy, 1990; Jarrett and
Russell, 1978; Tzavaras et al., 1990; Sheets et al.,
1993; Tsatsanis et al., 1994; Rohn et al., 1994). In
the current study, three of six cats, challenged with
rFeLVs and FeLV-A/Rickard, developed detectable thymic
LSAs within a period ranging from 40 to 86 weeks post
inoculation. In another investigation, 2 of 19 cats
inoculated only with the helper FeLV-A/Rickard virus
inoculum and kept under study for up to 52 weeks pi
presented with lymphomas (data not shown). Although
these control experiments were parts of other studies
(personal communication from L.E. Mathes), a few of the
cats were also observed for longer than 52 weeks. Two
of 5 such cats exhibited hematopoietic malignancies
(data not shown). In vivo studies with cloned FeLV-
A/61E, the parental FeLV-A used in the in vitro-
generation of our recombinants, indicate a low frequency
of thymic tumor induction (4 of 28 cats) as previously
reported (Overbaugh et al., 1992 ; Quackenbush et al.,
1990; Rohn et al., 1994). Although a larger study over
a longer period of observation, as well as a parallel
control study including FeLV-A alone, will be necessary
46
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
to define the significance of rFeLVs in lynphomagenesis,
at present there is an indication that the presence of
rFeLVs may potentiate malignancy development.
While analyzing tissue DNAs obtained from tumor-
bearing cats, it was of much interest to find the
nucleotide sequence for the MGPNL epitope, rather than
the expected enFeLV-encoded MGPNP sequence of the CFE-6
recombination partner in recombinants which displayed
recombination points downstream of the epitope (e.g.,
D’ , F and G sites). So far all exFeLV molecular clones
sequenced, including four FeLV-As, one FeLV-C and three
FeLV-Bs, have been found to contain the MGPNL
pentapeptide in the middle of the SU glycoprotein (Neil
et al., 1991). In contrast, examination of enFeLV cDNA
clones, as well as two enFeLV provirus clones, reveal a
variant sequence of MGPNP or MGPDP in this position
(Kumar et al., 1989; McDougall et al., 1994; Neil et
al., 1991; Sheets et al, 1993). Thus, in consideration
of the origin of FeLV-B viruses from enFeLV elements, it
has been difficult to explain the retention of MGPNL
epitope in FeLV-B subgroup isolates. Now however, our
data suggest that in vivo conversion of the epitope may
be responsible for these observations. In the midst of
homology to the enFeLV CFE-6 substituted region, its
MGPNP sequence is converted to MGPNL as deduced from a
47
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
C-»T transition in the codon of the fifth amino acid.
Apparently this has happened to all recombinant virus
clones examined in each of the three LSA-positive cats
studied. Additionally, the coexistence of FeLV-A seems
to be required for the observed nucleotide changes,
since in another study, when newborn cats were
inoculated with another subpopulation of rFeLVs alone
(prepared In vitro by selection of neutralization
resistance to C11D8 monoclonal antibody), none of the
cats developed viremia or seroconverted (unpublished
data presented in Chapter 5). However, recombinant
proviruses could be detected by PCR in the bone marrow
of these cats 45 weeks pi, and sequencing of the PCR
clones revealed unchanged MGPNP sequence consistent with
the rest of the portion of enFeLV (CFE-6) sequence.
Thus, efficient in vivo replication of the rFeLVs in the
presence of FeLV-A may be a requisite for epitope
conversion, potentially due to point mutations from
increased cycles of replication, followed by in vivo
selection for their growth advantage.
Although the above scenario seems to us the most
likely explanation, other reasons for the observations
cannot be excluded. For example, recombinant viruses
which we detected in the cats might be different, newly
generated recombinants formed in vivo through
48
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
recombination of FeLV-A with enFeLV elements. If any of
the multiple enFeLV loci retained the MGPNL sequence and
was indeed a partner in recombination, it might be
difficult to distinguish a newly formed recombinant from
the potentially homologous CFE-6 sequence. This
possibility is at least partly deemphasized by our
finding that the FeLV-A/Rickard plasma inoculum did not
contain any detectable recombinant viruses as measured
by nested PCR. Another possibility may be that viruses
detected in the cat were actually present in the rFeLV
pool and then outgrew others in vivo without any induced
changes in the epitope region. This, however, does not
affect our proposition that recombinant viruses with
MGPNL epitope are indeed products of recombination
between enFeLV and FeLV-A but with subsequent mutation
in the epitope area (Sheets et al., 1992).
Since tissues from terminally—ill infected cats
were used in PCR amplification and sequence analysis,
the question of when during infection timecourse the
epitope sequence alteration occurred could not be
addressed. However, analysis of sera from one tumor-
bearing cat over a 44-week time course indicated the
presence of rFeLV at virtually all collection time
points. Further cloning and sequencing of these serum-
derived RT-PCR products should provide clues as to when
49
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
epitope sequence alteration occurred in these co
infected cats (see Chapter 4). Another interesting
aspect is the apparent difference in detection levels
between the helper FeLV-A and rFeLVs RNA levels in the
sera from cat 4746-5 over the infection timecourse.
FeLV—A displayed a cyclical pattern while rFeLV levels
remained fairly constant, except for an apparent short
delay in the first detection point of rFeLV. The
fluctuating pattern of FeLV-A virus in serum may suggest
attempts to clear FeLV-A infection by the host's immune
response. Conversely, a fairly uniform detection of
rFeLVs may indicate the inability of the host to
effectively clear infection by the recombinants, further
corroborating early data on poor immunologic responses
in cats infected with FeLV-B (Russell and Jarrett,
1978). It is noteworthy that despite the conservation
of one of the major neutralizing epitopes (ie, MGPNL) in
its SU, FeLV-B fails to elicit a strong immune response
in infected cats. Perhaps the close similarity of the
N-terminal half of FeLV-B SU to that of an endogenously
expressed enFeLV protein may play a role in manifesting
some degree of immune tolerance.
Finally, our results underscore the importance of
biological conservation of the MGPNL pentapeptide
sequence in the middle of the SU glycoprotein for
50
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
exogenous, horizontally-spread FeLVs. Although
recombinants with altered pentapeptide can efficiently
replicate in vitro (Pandey et al., 1991; Sheets et al.,
1992), they appear to be selected against in vivo. This
region constitutes a prominent binding site for a
monoclonal antibody, and conceivably this conformational
epitope, may also be necessary for additional protein-
protein interactions facilitating one or more important
events, such as efficient polyprotein folding, receptor
recognition, membrane insertion, membrane fusion and
virus maturation. The factors that regulate these
interactions may more likely be encountered under in
vivo conditions than in vitro cell culture conditions
where pentapeptide chemges may have little significant
effect on virus replication efficiency. In addition to
the observed epitope conversion, it is also demonstrated
here that rFeLV replication efficiency in vivo may also
be dictated by the site of recombination. Some rFeLV
species appear to have an in vivo proliferative
advantage. One such recombinant, site G rFeLV, detected
in all three tumor-bearing cats, could potentially
represent em exemple of replication efficiency coupled
to a strong leukemogenic function.
51
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
CHAPTER 3
In Vivo Selection of Specific Recombinant
Feline Leukemia Virus Species
ABSTRACT
Ecotropic FeLV-A is known to recombine with
endogenous FeLV (enFeLV) env elements yielding
recombinant polytropic FeLV-B viruses which in turn are
frequently associated with feline lymphosarcomas (LSA).
In this portion of the study, we examined the in vivo
selection of recombinant FeLV (rFeLV) viruses in three
experimentally-induced thymic LSAs, and compared these
results with findings derived from a naturally-occurring
thymic tumor. Two of the three experimental cats were
challenged with a preparation of FeLV-A/Rickard, while
one cat received this FeLV-A along with a mixture of in
vitro-generated rFeLVs. The FeLV-A/Rickard preparation
used in this study was shown to be free of detectable
rFeLVs since no recombinant products were observed
following nested PCR analyses. PCR amplification and
cloning of the surface glycoprotein (SU) portion of the
proviral env gene was performed using the four tumor
DNAs, emd multiple clones of rFeLV-specific PCR products
were exeunined by nucleotide sequence analyses. Evidence
is presented indicating there is in vivo selection for
52
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
specific recombinant species containing relatively
greater eumounts of enFeLV-derived SU sequence and that
this in vivo selection process for certain recombinant
species appears to be rather gradual, occurring over the
timecourse of infection. The recombinants eventually
emerging as the predominant species may have in vivo
selective advantage, suggesting a role for these
particular recombinant species in the FeLV-mediated
disease process.
53
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
INTRODUCTION
Early epidemiologic surveys reported polytropic
FeLV-B viruses in approximately 30% of field isolates,
and always in association with ecotropic FeLV-A (Jarrett
et al., 1978), while FeLV-B viruses were found in
approximately 60% of cats developing LSA (Jeurrett et
al., 1978; Tzavaras et al., 1990; Sheets et al., 1993;
Tsatsanis et al., 1994), implicating FeLV-B viruses in
the disease process. In one of these studies, direct
molecular genetic analyses performed in our lab showed
the incidence of recombinémt, FeLV-B-like viruses in
naturally-occurring LSAs (Sheets et al., 1993). The
findings were: -75% of thymic lymphomas (11 of 15); 75%
of alimentary lymphomas (3 of 4); and 33% of
multicentric lymphomas (2 of 6) contained recombinant
FeLV proviruses in cats shown to be FeLV-positive by
immunofluorescence assay while only 1 of 22 FeLV-
negative naturally-occurring LSA contained recombinant
FeLV provirus. These results were corroborated by
another study in which 63% of natural feline lymphoma
were found to harbor recombinant env as determined by
blot hybridization using a FeLV-B-specific probe
(Tsatsanis et al., 1994). The structural motifs for the
recombinant proviral env genes found in the naturally-
occurring LSAs were similar to those produced in vitro
54
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
by our lab (Sheets et al., 1992). These recombination
structural motifs involved enFeLV-derived sequence at
the 5' end of the env gene prior to crossing over to
exogenous FeLV-A-like sequence. More thorough
characterization of the recombination junctions found in
the in vitro-derived recombinants indicated a limited
region, encompassing approximately 270 nucleotides in
the middle of the SU domain of the env gene, was
preferentially utilized in the generation of these
recombinants. The in vivo infectivity of these in
vitro-generated recombinants was evaluated in a co-
infection study with FeLV-A helper virus in which three
of six cats developed LSA (described in Chapter 2).
While the rFeLVs used in this study contained crossover
sites ranging from site A (least amount of enFeLV
substitution) to site G (greatest amount of enFeLV
substitution), only sites D to G were detected in the
tissue DMA, suggesting an in vivo selective advantage
for particular recombinant species harboring relatively
greater amounts of enFeLV-derived sequence. Here, we
extend our analysis and present evidence that specific
rFeLV species containing relatively more endogenously-
derived SU sequence appear to predominate from a mixed
virus population during in vivo infection and that these
55
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
particular recombinant species are predominant in both
experimentally-induced as well as naturally-derived LSA.
56
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
MATERIALS AND METHODS
Viruses
The rFeLV pool was previously described (Pandey et
al., 1995). As a reminder, eight clones originally
sequenced from this rFeLV pool were shown to contain
multiple recombinant species with various crossover
sites in the env gene. One clone was a site A
recombinant, one clone a site B, and two clones each had
crossover sites E, F, and G (Sheets et al., 1992). Site
A recombinants contained the least amount of enFeLV
substitution from the 5' end of env, while site G
recombinants contained the most enFeLV substitution.
Animal studies
In vivo infection experiments using specific
pathogen-free (SPF) kittens from the breeding colony at
Ohio State University, were performed as previously
described (Pandey et al., 1995). Briefly, six neonates
(4746-1 to 4746-6) inoculated with the rFeLV mixture and
the FeLV-A/Rickard plasma preparation resulted in three
of six cats developing thymic tumor. In addition, two
other SPF kittens, 3315 and 3541, were challenged with
the FeLV-A/Rickard alone. Tumors were collected at
necropsy from cats 3315 and 3541 at 72 weeks pi and 74
weeks pi, respectively.
57
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Tissues and cell lines
DNÀ isolation from the naturally-occurring thymic
tumor tissue, Pll, as well as its source and
characterization were previously described (Levy et al.,
1993; Sheets et al., 1993). High molecular weight DMA
was extracted from tumor and bone marrow tissues
obtained at necropsy, and from control cell lines as
reported earlier (Pandey et al.,1995). The two thymic
LSAs experimentally induced by infection with FeLV-
A/Rickard were previously described (Levy et al., 1993).
The various control cell lines included: FeLV-A/6IE-
infected feline H927 fibrosarcoma, uninfected H927, and
rFeLV-infected human HT1080 fibrosarcoma. Total RNA
isolation was performed using Clontech's total RNA
isolation kit (Palo Alto, CA) as reported earlier
(Pandey et al.,1995).
Amplification and cloning of proviral DMA
Detection of recombinant env proviral sequences in
tumor DMA of cats 3315 and 3541, and in cDNA from cat
3541 tumor tissue, was accomplished by polymerase chain
reaction (PCR) with the endogenous-specific 5' sense
primer RB53 and FeLV-A-specific 3' emtisense primer RB19
(Sheets et al., 1993), employing 45 cycles of
amplification. PCR products were analyzed by
58
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
electrophoresis on 1% agarose gels and the 1,376 bp
recombinant-specific fragment visualized by ethidium
bromide staining. To clone the SU-encoding portion of
recombinant env from the various tissue DNAS, PCR
amplification was performed using a 5' sense primer
(H19) recognizing a sequence at the pol/env junction
(pF6A 5877-+5S93) which is conserved between enFeLV and
exogenous FeLV-A (GTAGACGGAGTTGCTGC) and the 3'
antisense primer RB19 (Sheets et al., 1993), yielding a
2.0 Kb product. These 2.0 Kb PCR products were cloned
into Invitrogen's TA cloning vector (San Diego, CA).
Screening for rFeLV-specific clones in this mixture was
performed by PCR using the recombinant-specific primer
set, RB53 and RB17. RB53 is an enFeLV-specific 5' sense
primer and RB17 is an FeLV-A-specific 3' antisense
primer (Pandey et al., 1995) which yields an 888 bp
fragment.
Reverse transcription—polymerase chain reaction (RT-PCR)
amplif ication
Five ug total RNA was used in a 50 ul reverse
transcription (RT) reaction. RT-PCR without RNA or
without RT enzyme were included in each experiment,
serving as negative controls. Additionally, in all PCR
analyses, RT negative controls were included to ensure
59
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
no genomic DNA contamination of the cDNA samples. A
nested PCR strategy was employed for detection of FeLV
RNA levels in sera samples as previously described
(Pandey et al., 1995). In summary, the FeLV-A-specific
primer set, RB59 and RB52, was used in 35 cycles for
first-round amplification; the FELV-A-specific primer
set, RB326 and RB17, was used in 35 cycles for second-
round amplification, yielding a 963 bp product. To
detect rFeLV RNA in sera, first-round aunpl if ication
utilized the rFeLV-specific primer set, RB56 and RB52,
in 35 PCR cycles; second-round amplification employed
the rFeLV-specific primer set, RB53 and RB17, in 35
cycles, yielding an 888 bp fragment. To investigate the
in vivo selection of recombinant species over infection
timecourse, sera seunples for seven timepoints from cat
4746-5 were used in rFeLV-specific nested RT-PCR. The
888 bp PCR products were excised from the gel, purified,
and inserted into the TA cloning vector. To clone the
entire SU—encoding portion of recombinant env from sera
RNA of cat 4746-5 at 8 weeks pi, a nested RT-PCR
strategy was again employed. First-round amplification
used a 5' sense primer (HIS) recognizing a sequence at
the pol/env junction (pF6A 5840-^5860) which is conserved
between enFeLV and exogenous FeLV-A
(ACATATCGTCCTCCTGACCAC), along with a 3' antisense
60
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
primer (H20) ( GAAGGTCGAACTCTGGTCAACT ) complementary to
the exogenous U3 region located in the LTR (pF6A
8189-+8210), using 35 amplification cycles and resulting
in a 2.37 Kb PCR product. Second-round amplification
used 1 fil of product from first-round PCR with primer
set H19 and RB19 as described above. The 2.0 Kb
products from the second-round of amplification were
also inserted into the TA cloning vector and screening
for recombinant-specific clones was perfoznsted using the
rPeLV-specific primer set, RB53 and RB17 as described
above.
Nucleotide sequencing
For analysis of in vivo selection for specific
recombinant species over the 44-week infection
timecourse, ten to fifteen clones for each of the seven
sera timepoints from cat 4746-5, totalling 85 clones,
were sequenced manually using two internal sequencing
primers, MP-1 and MP-2 (Fig. 5), as reported earlier
(Pandey et al., 1991). Sequences were then aligned and
compared to those of enFeLV CFE-6 (Kumar et al., 1989)
and FeLV-A/61E (Donahue et al., 1988). For analysis of
the entire 1.2 Kb SU-encoding env region, two to four
clones were selected for each tissue seunple, as well as
seven clones from the rFeLV inocula, emd automated
61
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
sequencing performed using fluorescence-based cycle
sequencing with the ABI Prism 377 DNA Sequencer (Perkin
Elmer, Foster City, CA) and the ABI Prism Dye Terminator
Cycle Sequencing Kit (P/N 402080) as per the
manufacturer's protocol. To sequence the 1.2 Kb
fragments in both directions, four primers were used in
the automated sequencing strategy: AP-1, AP-2, AP-3, and
AP-4, as outlined in Fig. 5. Cycle sequencing employed
the GeneAmp 9600 thermal cycler in 25 rounds of
amplification and nucleotide sequences were resolved
using 4% acrylcunide gels. Nucleotide sequences were
initially read and analyzed using the ABI Prism DNA
Sequencer Analysis software. Version 3.0. All sequence
data were subsequently assembled, analyzed, and compared
to published sequences for FeLV-A/61E, FeLV-B/GA, FeLV-
B/ST, FeLV-B/RIC, and CFE-6 (Donahue et al., 1988? Elder
and Mullins, 1985? Kumar et al., 1989? Nunberg et al.,
1984) using the Genejockey software.
62
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
pol
TA vector
env
ABCD
n i l
EFG
III
TM^
1
A ^-2
M M 1
^TAveetor f e L V e n i f
Fig. 5 Strategy for sequencing the SU portion of the
env gene for recombinant-specific PGR products. Manual
sequencing of an -.9 Kb rPeLV fragment was performed
using MP-l and MP-2. These primers were previously
described as RB-38 and RB-39, respectively (Hathes et
al., 1994). Automated sequencing of the 1.2 Kb SU
fragment employed primers AP-1 to AP-4. AP-1 is
universal Ml3 reverse primer matching vector sequence at
the junction between vector and the 5' end of the
insert. AP-2 is a 3' antisense primer complementary to
enFeLV clone, CFE-6, at position 512-*531
(TTTACTGTGATGTAGTCCCA). AP-3 is a 5' sense primer
previously described as RB53 (Sheets et al., 1993) and
AP-4 is a 3' antisense primer complementary to a
sequence conserved between enFeLV and FeLV-A/61E env at
CFE-6 position 1212-»1227 (TGGGTCTTAGGAACAGT). Location
of recombination crossover sites A to G as previously
characterized are also included for reference (Sheets et
al,, 1992).
63
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
RESULTS
PGR detection of rFeLVs in tumor tissue of infected cats
Recombinant FeLV proviruses were detected in tumor
DNA of two cats, 3315 and 3541, challenged with a FeLV-
A/Rickard inoculum alone (Fig. 6). Parallel analysis of
cDNA derived from tumor tissue of these two cats showed
detectable rPeLV expression only in cat 3541, while no
rFeLV expression was detected in cat 3315 (Fig. 6).
Since it was possible the FeLV-A/Rickard preparation
harbored rFeLVs, an analysis using nested PCR for rFeLV-
specific products was conducted. As reported previously
(Pandey et al., 1995), no rFeLV fragments were detected
in the FeLV-A/Rickard preparation using this nested PCR
strategy, only FeLV-A-specific products were observed.
The FeLV-A control presented in Fig. 7 contains cDNA
derived from the FeLV-A/Rickard inoculum and confirmed
these findings. Our results suggested the rFeLVs
observed in tumor DNA of cats 3315 and 3541 arose de
novo. The presence of rFeLV proviral env sequences in
tumor tissue DNA of the naturally-occurring thymic
tumor, Pll, was previously described (Sheets et
al.,1993) and confirmed in this study (data not shown).
64
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
1.5 Kb—
Genomic DNA c D N A -> |
Fig. 6 PCR detection of rFeLV env in tumor DNA and cDMA
of cats with experimentally-induced LSA. Animals were
challenged with the FeLV-A/Rickard plasma inoculum
alone. The 1.37 Kb recombinant proviral product was
detected in tumor DNA of cats 3315 and 3541. DNA
extracted from FeLV—A/6lE-infected feline H927
fibrosarcoma cells and uninfected H927 cells served as
negative controls while DNA extracted from rFeLV-
infected human HT1080 fibrosarcoma cells served as
positive control. Expression of rFeLV was observed only
in tumor cDNA of cat 3541. For positive control, cDNA
from rPeLV-infected HT1080 cells was used, while the
negative control was 3541 cDNA without RT enzyme. M,
100 bp DNA ladder.
65
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Detection of FeLV RNA levels in sera of FeLV-co—infected
cat 4746-5. As reported earlier (Pandey et al., 1995),
cat 4746-5 was one of six cats co-infected with both the
parental rFeLV pool and the FeLV-A/Rickard helper virus.
Serum samples were obtained from this cat at two-week
intervals throughout the timecourse of infection,
beginning at 4 weeks pi and ending at 44 weeks pi. One
objective of this study was to identify the earliest
timepoint for detection of rFeLV species, so additional
early timepoint seunpling intervals of 6 weeks euid 10
weeks pi were added between 4 weeks and 12 weeks pi. A
representative experiment demonstrating viral RNA levels
at the early timepoints is presented in Fig. 7. These
results confirmed our previous report that the earliest
timepoint at which rFeLV RNA could be detected in cat
4746-5 was eight weeks pi (Pandey et al., 1995). It
should be noted that while this nested RT-PCR was not
quantitative per se, the analysis was designed to be
semi -quantitative in that the total RNA for each sample
studied was standardized to 5 ug total RNA in the RT
reaction, and the cDNA used in the PCR reaction was also
equivalent for each sample. In addition, the same cDNA
prepsurations were used in experiments depicted in both
gels of Fig. 7 in order to ensure that relative viral
RNA levels were reflected for each sera timepoint.
66
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
FeLV-A
FéLV-A
Weeks pi
rFeLV
HT1080
Water
0.8.
8 10 12 RT-
0.96 Kb
rFeLV
, E bL V = A
-Weeks pi— I
6 8 10 12 R T^
ESLV
HT1080
Water
—0.9 K b
Fig. 7 PCR detection of FeLV RNA levels in early sera
samples of cat 4746-5. The top panel (FeLV-A RNA)
depicts a -960 bp product from FeLV-A-specific nested
PCR. The bottom panel (rFeLV RNA) shows a -900 bp
product from recombinant-specific nested PCR. In both
panels, the lanes contain the following, reading from
left to right : DNA marker; 4 weeks pi; 6 weeks; 8 weeks;
10 weeks ; 12 weeks pi; and 12 weeks pi, RT(-),
respectively. In the top panel, cDNA from FeLV-
A/Rickard plasma preparation served as positive control,
while rFeLV cDNA is a negative control. The lack of a
band in the lane containing rFeLV demonstrates these are
FeLV-A-specific primers. Other controls are uninfected
HT1080 CDNA and water. In the bottom panel (rFeLV RNA),
one negative control is the FeLV-A/Rickeird. A lack of
band in this leune demonstrates the primer set is
recombinant-specific and that no recombinant products
could be detected in the FeLV-A/Rickard inoculum by this
nested PCR strategy. All other lanes in the bottom
panel contain the same samples in the order described in
the top panel.
67
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Recombination junctions determined for rFeLV clones
The rFeLV mixture used for ±n vivo challenge
experiments was initially shown to contain a spectrum of
3' recombination sites. Of the eight clones originally
sequenced to determine their 3' recombination junctions,
1 of 8 was found to be a site A recombinant, 1 was a
site B, 2 were site E, 2 were site F and 2 of the 8
clones were site G recombinants (Sheets et al., 1992).
In our current study, we extended characterization of
crossover sites within the initial rFeLV mixture to
include both 5' and 3' recombination junctions for an
additional seven clones from this parental rFeLV
inocula. Results of this sequence analysis are
presented in Fig. 8 and a summary of the 3'
recombination junctions sites for those clones derived
from the rFeLV inocula is presented in Table 3.
As noted in Fig. 8, the 5' junctions for clones 3
and 4 from the rFeLV pool could not be ascertained in
this study since enFeLV-like sequence was found at the
5* end of their inserts. Three of seven clones (clones
1, 2, and 5) from the rFeLV pool had 5' crossovers in
the region near the pol/env junction, while two (clones
6 and 7) had 5' junctions near the start of the SU-
encoding portion of the env sequence. As presented in
68
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 8. Schematic representation of the various 3'
recombination structural motifs observed within the SU
region analyzed as compared to FeLV-A/61E (Donahue et
al., 1988), enFeLV CFE6 (Kumar et al., 1989) and the
three known FeLV-B isolates, FeLV-B/GA (Nunberg et al.,
1984), FeLV-B/ST (Nunberg et al., 1984) and FeLV-B/RIC
(Elder and Mullins, 1985). FeLV-A/61E sequence is
indicated by solid black, deletions within FeLV-A
sequence relative to the endogenous CFE6 sequence are
also shown. CFE6 sequence is depicted in gray, with
deletions within CFE6 sequence as compared to FeLV-A
also presented. Both the 5* and 3' recombination
junctions are shown. Recombination sites for the FeLV-B
isolates are based upon their reported sequence. The 5'
junction for FeLV-B/ST is still unknown. For cat 3541,
tumor DNA clone 1, the location of its 3' recombination
junction is unknown (see text for details). The sera
sample used from act 4746-5 was obtained at 8 weeks pi
and the bone marrow (BH) sample from cat 4746-5 was
obtained during necropsy at 44 weeks pi.
69
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Schematic of Observed Recombination Structural Motifs
ABCD
p — n
E F G
TM.
• • • FeLV-A/61E
. . enFd.VICF&«
... FeLV-BA8A
... FeLV-B/ST
. . . FeLVe/RIC
» . . done 1 rFeLV Inoeula
Cione2
Clones
t. . don#4
Clones t . .
•. . Clones
' • . Clone/
. . Clone 1 , Cat 4746-6, Sera
I. . doneZ
. . Clones
. . Clone 1 , Cat 4746-6, BM
■ .. Clone 2
!.. Clones
>. . Clone 1 , Cat 4746-6, Tumor
». . Clones
t. . Clones
!.. Clone 1,CatP11,Tumor
>.. Clones
I.. Clone 1,Cat3315,Tumor
>. . done 2
!.. Clones
... Clone 1,Cat3641,Tumor
. . Clones
. . Clones
.. Clones
70
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table 3: Summary o f the 3' Recombination Junction Sites
DESCRIPTION 3* JUNCTION SITE
FeLV-B/GA E
FeLV-B/ST >G
FeLV-B/RIC E
rFeLV Inoeula Clone 1 B
Clone 2 F
Clone 3 E
Clone 4 G
Clone 5 F
Clone 6 G
Clone 7 A
Cat 4746-5, Sera Clone 1 E
Clone 2 E
Clone 3 E
Cat 4746-5, BM Clone 1 G
Clone 2 G. Clone 2 had a
3' junction well upstreeua of site A and so was
designated G.
Thus, all three FeLV-B isolates have large eu&ounts of
endogenous-like sequence comprising the SU portion of
the env gene. Demonstrating similar recombination
structural motifs to the exogenous FeLV-B isolates,
clones derived from tumor DNA of all four cats showed
that 9 of 12 (75%) clones had either site F or G
recombination junctions, and 6 of 12 (50%) were site G
recombinants, while only 3/12 (25%) had 3' recombination
junctions of sites A to D.
In Vivo selection of recombinant FeLV species. Another
importamt question we wished to address in this study
74
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
was the pattern of in vivo selection for various
recombinant species. Combining our earlier report
(Sheets et al,, 1992) with current results, we
demonstrated that the rFeLV pool used in this in vivo
co-infection study contained a spectrum of recombination
sites. We also recently reported that site D, F, and G
recombinemts were the predominant species types detected
in various tissues of three tumor-bearing cats receiving
both the rFeLV mixture and FeLV-A helper virus,
indicating some in vivo selection had occurred (Pandey
et al., 1995). To investigate this pattern of in vivo
selection, we cloned and sequenced rFeLV RT-PCR products
derived from sera of cat 4746-5 at various timepoints
throughout the infection timecourse. Results summarized
in Fig. 9 indicate that at 8 weeks pi, the timepoint at
which rFeLVs were initially detected in sera, 2/14
clones were site D recombinants, 2/14 were site E, 3/14
were site F, 3/14 were site G, and 4/14 clones had
crossover sites greater than site G (>G). At the next
timepoint of 12 weeks pi, 7/11 clones were site F
recombinants while 4/11 clones were site G recombinant
species. By 16 weeks pi, the in vivo selection pattern
indicated the vast majority (12/15) were of the site G
recombinauat species type, with a minority of the clones
representing other species (2/15 were site C species and
75
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
1/15 was a site F species). Throughout the remainder of
the infection timecourse ^ site G recombinant species
gradually emerged as the predominant species in the sera
of this particular cat.
76
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Fig. 9 In vivo selection of rFeLV species in sera of
cat 4746—5 over the 44-week timecourse of infection.
Panel A presents a histogram showing the number of
clones for each recombinant species detected at each of
the sampling intervals as described in the text. Panel
B diagrams the rFeLV env gene, showing the various 3'
recombination crossover sites previously described
(Sheets et al., 1992).
77
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
u- 111 û o
CO
CM
4----1 ---- 1 ----h
CM
i h
CM
S9UOIO jo JoquinN
«
c
o
S)
s
>
_l
0 )
A
I.
?
U9
-C L
1U
§.
I '
3S
s t
ffi
78
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
DISCUSSION
We recently observed an apparent preference for
specific recombinant species containing relatively
greater amounts of enFeLV-derived sequence following in
vivo selection in three tumor-bearing cats co-infected
with a rFeLV mixture and FeLV-A helper virus (Peuxdey et
al., 1995), encouraging us to pursue these issues
further. Here we report that in the course of in vivo
infection, rFeLV species containing longer enFeLV-like
env substitution predominate over recombinants with
lesser amounts of enFeLV-derived substitution. In the
co-infection study which included cat 4746-5, the rFeLV
pool used as challenge virus initially contained
recombinants with crossover sites ranging from site A to
site G (Sheets et al., 1992), yet the predominant
crossover sites detected in vivo were sites D, F, and G
(Pandey et al., 1995). We proposed two hypotheses to
explain this observation. Firstly, in vivo selection
for certain recombinant species could be an early event
resulting in immediate selective advantage for specific
recombinants over others, resulting in the rapid
emergence of predominant species early in the timecourse
of infection. Alternatively, a gradual outgrowth from a
mixture of all the recombinant species may occur during
the timecourse of infection, resulting in the slow
79
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
emergence of predominant species. Our observations
suggest gin eunalggun of these two possibilities. The
earliest timepoint at which rFeLV RNA could be detected
was 8 weeks pi; at this timepoint site D to site G
recombinant species were detected. These results
indicate that at least in cat 4746-5, recombinant
species with crossover sites upstream of site D are, for
the most pgirt, selected against very early in the course
of infection. Subsequently, we observed a gradual
selection for site G recombinants resulting in their
eventual predominance over other recombinemt species.
Interestingly, while the source of the rFeLVs could not
be determined in the naturally-occurring thymic tumor,
Pll, both clones from this cat were site G recombinants.
Moreover, site G recombinants were the predomineint
species recovered from cat 3315 in spite of the fact
that the rFeLVs detected in this cat had apparently
arisen de novo over the natural course of in vivo
infection. We thus propose that recombinants with more
endogenously-derived SU sequences, such as site G
species, may have an in vivo selective advgmtage. It is
possible via the mechanism of immune tolerance that
expression of additional endogenously-derived sequence
on the virion surface glycoprotein may confer some
resistance to host immune surveillance. In this regard,
80
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
work previously reported by our lab described the low-
level expression of enFeLV elements in placental and
fetal lymphoid tissues, as well as some FeLV-negative
lymphomas (Busch et al., 1983; Roy-Burman, 1996). It
has also been reported that truncated enFeLV env are
highly expressed in feline lymphoma cell lines and
normal hemopoietic tissues (McOougall et al., 1994).
Thus, it is conceivable that enFeLV env expression
during fetal development and/or tissue-specific
expression in the thymus during T-cell maturation, may
induce tolerance to rFeLVs containing large portions of
endogenously-derived SU sequences. An alternative
explanation for the apparent in vivo selective advantage
for certain recombinant species is that variously
substituted SU regions may have different affinities for
FeLV-B receptors expressed on feline target cells,
potentially influencing in vivo outgrowth of particular
rFeLV species in a fashion similar to what is reported
here. There is, however, no direct evidence in support
of either of these points at this time. In summary, we
propose that some early in vivo selection occurs against
recombinant species containing lesser amounts of
endogenously-derived sequence, followed by a gradual
selection for recombinants harboring relatively greater
amounts of endogenously-derived sequence, potentially
81
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
implicating this species in the multi-step process of
FeLV-mediated leukemogenesis.
82
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
CHAPTER 4
Nucleotide Sequence Conversion of Recombinant Feline
Leukemia Viruses (rFeLVs) to FeLV Subgroup B Viruses
ABSTRACT
While ecotropic FeLV-A is known to recombine with
endogenous FeLV (enFeLV) env elements yielding
polytropic FeLV-B viruses, there are scattered
differences between enFeLV env elements and
corresponding sequences of exogenous FeLV-B isolates.
To address this disparity, we examined the in vivo
evolution and selection of recombinant FeLV (rFeLV)
viruses in three experimentally-induced thymic LSAs, and
compared these results with findings derived from a
naturally-occurring thymic tumor. Two of the three
experimental cats were challenged with a preparation of
FeLV-A/Rickard, while one cat received this FeLV-A along
with a mixture of in vitro-generated rFeLVs. The FeLV-
A/Rickard preparation was shown to be free of detectable
rFeLV levels since no recombinant products were observed
following nested PCR analyses. PCR amplification and
cloning of the surface glycoprotein (SU) portion of the
proviral env gene was performed using the four tumor
DNAs, and multiple clones of rFeLV—specific PCR products
83
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
were examined by nucleotide sequence analyses.
Substitution at a total of 19 nucleotide positions
relative to enFeLV sequence were found scattered
throughout the SU portion of the env gene in these
clones. Most interestingly, this set of 19 point
mutations led to complete nucleotide and amino acid
sequence identity with known FeLV-B isolates. Our
results also indicate these mutational events occur
early in the In vivo evolution of recombinant viruses,
suggesting this set of point mutations is strongly
selected and may be important for efficient FeLV—B
replication.
84
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
INTRODUCTION
FeLVs are naturally-occurring ^ horizontally-
transmitted, mammalian type-C retroviruses (oncovirinae)
capable of causing a spectrum of diseases in the outbred
domestic cat population (Hardy, 1993). Exogenous FeLVs
are divided into three subgroups, designated A, B, and
C, based upon viral interference and neutralization
assays (Sarma and Log, 1971; 1973). While ecotropic
FeLV-A viruses have been shown to have low pathogenicity
and occur in 100% of field isolates (Jarrett and
Russell, 1978; Jarrett et al., 1978; Sarma et al.,
1978), polytropic FeLV-B viruses are estimated to occur
in approximately 30% of field isolates (Neil et al.,
1991), are always isolated along with FeLV-A, FeLV-C, or
both, and are overrepresented in LSAs (Jarrett et al.,
1978; Levy et al., 1993; Sheets et al., 1993; Tsatsanis
et al., 1994). There are also multiple enFeLV proviral
elements known to exist in the feline genome (Roy-
Burman, 1996), but while these enFeLV proviruses are all
defective due to various mutations or deletions, they
generally retain transcriptional activity (Busch et al.,
1983; HcDougall et al., 1994; Soe et al., 1983; 1985).
Polytropic FeLV-B viruses are known to have arisen
via recombination between FeLV-A env sequences and
corresponding but varied enFeLV proviral elements (Elder
85
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
et al., 1983; Kumar et al., 1989; Overbaugh et al.,
1988; Pandey et al., 1991; Roy-Burman^ 1996; Sheets et
al., 1993; Soe et al., 2983; Tsatsanis et al., 1994).
In spite of this fact, a comparison of the env
nucleotide sequence for the cloned enFeLV, CFE-6 (Kumar
et al., 1989), with those of three molecularly cloned
FeLV-B isolates (Elder and Mullins, 1985; Nunberg et
al., 1984) shows distinct differences scattered
throughout the region, differences which are generally
conserved in all three FeLV-B isolates in spite of their
diverse geographic origins (Mullins and Hoover, 1990).
We were interested in studying this paradox as well
as the nature of rFeLV species which appear to be
selected during the temporal course of in vivo infection
and recently reported results of our initial
investigation (Pandey et al., 1995). We presented
nucleotide sequence analysis of rFeLV proviral DNA
obtained from three tumor-bearing cats, co-infected with
the FeLV-A/Rickard and a pool of in vitro-generated
rFeLVs, demonstrating a C-+T transition mutation in the
env gene resulting in an amino acid change from the
enFeLV-derived Pro217 to that of the exogenous Leu217.
This Leu217 is strictly conserved in all naturally-
occurring exogenous FeLV molecular clones thus far
sequenced (Neil et al., 1991) and resides at the end of
86
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
an epitope previously characterized as a major
neutralizing epitope in the mid-SU region (Elder et al.,
1987; Grant et al., 1983).
In this chapter evidence is presented indicating
conversion from enFeLV env sequence to exogenous FeLV-B
sequence is an early event in the in vivo evolution of
rFeLVs, suggesting this set of point mutations
responsible for sequence conversion is strongly selected
and may be important for efficient FeLV-B replication.
87
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
MATERIALS AMD METHODS
Viruses
The rFeLV pool was previously described (Pandey et
al., 1995) and in Chapter 2. Briefly, this rFeLV pool
was shown to contain multiple recombinant species such
that the length of enFeLV substitution from the 5' end
of env increases from crossover site A (CFE-6 nucleotide
position env746) to site G (CFE—6 nucleotide position
envioie) (Kumar et al., 1989; Sheets et al., 1992).
Animal studies
In vivo infection experiments using specific
pathogen-free (SPF) kittens from the breeding colony at
Ohio State University, were performed as reported
earlier (Pandey et al., 1995). Briefly, six neonates
(4746-1 to 4746-6) inoculated with the rFeLV mixture and
the FeLV-A/Rickard plasma preparation resulted in three
cats developing thymic tumor. In addition, two other
SPF kittens, 3315 and 3541, were challenged with the
FeLV-A/Rickard alone and tumors were collected at
necropsy at 72 weeks pi and 74 weeks pi, respectively.
Tissues and Cell Lines
DMA isolation from the naturally-occurring LSA
tumor tissue, Pll, as well as its source and
88
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
characterization was previously described (Levy et al.,
1993; Sheets et al., 1993). High molecular weight DNA
was extracted from tumor and bone marrow tissues
obtained at necropsy, and from control cell lines as
reported earlier (Pandey et al., 1995). The six thymic
lymphosarcomas experimentally induced by infection with
FeLV-A/Rickard were previously described (Levy et al.,
1993). The various control cell lines included: FeLV-
A/61E-infected feline H927, uninfected H927, and rFeLV-
infected human HT1080 fibrosarcoma. Total RNA isolation
was performed using Clontech/s total RNA isolation kit
(Palo Alto, CA) as previously described (Pandey et al.,
1995).
Amplification and cloning of proviral DNA
Detection of recombinant env proviral sequences in
tumor DNA of cats 3315 and 3541, and in cDNA from cat
3541 tumor tissue, was accomplished by polymerase chain
reaction (PCR) as described in Chapter 3.
Reverse transcription-polymerase chain reaction (RT-PCR)
amplification
The RT-PCR amplification strategy is described in
Chapter 3. A nested PCR strategy was employed for
detection of FeLV RNA levels in sera seunples as
89
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
previously described (Pandey et al., 1995) and outlined
in Chapter 3. To clone the entire SU-encoding portion
of recombinant env from sera RNA of cat 4746-5 at 8
weeks pi, a nested RT-PCR strategy was again employed
and is described in Chapter 3.
Nucleotide sequencing
For analysis of the rFeLV env region, multiple
clones were selected for each tissue seunple, as well as
seven clones from the rFeLV inocula, and automated
sequencing performed as described in Chapter 3.
90
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
RESULTS
Recombinant FeLV nucleotide changes observed in vivo
correspond to sequences for exogenous FeLV-B isolates
To analyze the entire SU portion of the env gene,
automated sequencing was performed utilizing four
primers (ÀP-1 to AP-4) spanning this region (Fig. 5).
Results presented in Fig. 10 show nucleotide
substitutions were observed at a total of 19 positions
relative to parental rFeLV clones and enFeLV sequences.
Changes at these 19 positions were not clustered in any
particular region, nor were these substitutions confined
to regions characterized as variable regions, but
instead were scattered throughout the 1.2 Kb region
emalyzed. Twelve of these substitutions also resulted
in altered amino acid residues. A summary of these
observed substitutions is presented in Table 4.
Nucleotide changes at all 19 positions, as well as
relevant putative amino acid changes, corresponded with
reported nucleotide and amino acid sequences for three
molecularly cloned FeLV-B isolates. At 12 of 19
positions, there was strict conservation between
nucleotide sequences for the three FeLV-B isolates and
those sequences for all in vivo—derived clones, whether
naturally-occurring or experimentally—induced thymic
91
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Fig. 10 Schematic summary of nucleotide sequencing for
the 1.2 Kb rFeLV emr gene indicating the positions of 19
mutations found scattered throughout the region
analyzed. The 19 mutations are highlighted in gray
within the magnified sequence regions. Sequences are
aligned relative to the enFeLV clone, CFE6 and numbering
is as reported (Kumar et al., 1989). Only nucleotides
differing from CFE6 are shown; (.) indicates no change
in nucleotide sequence. The presumed amino acid
sequence for CFE6 is presented above its nucleotide
sequence, and only differences in amino acids as
compared to CFE6 are shown. Panel À presents portions
of the 5' half of the SU-encoding region showing
mutations at positions 1 to 11 as observed in five
magnified sequences. Panel B presents six magnified
sequences from the 3' half of the rFeLV SU-encoding env
showing eight mutation positions (12 to 19) observed in
this region. Recombination crossover sites A to G are
shown for reference, along with the approximate pol/env
junction, the start of the mature SU, and start of the
transmembrane (TM) domain.
92
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
A
polenv
A l CD E
j
FG
r y y \
/ yv
, «rFeLV env
93
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Fig. 10 (Continued)
• •rFeLV env
r t » ~ e / 9 t
. « % » ■ J i . . . . . . . . c . . . . . . . . e . . .
tWmXff aOCBEA
t o “
1 a o a » . J U . A e
t a o M
« 7 4 C -8 a n A . • « t a p i
3 a
< 7 4 e > 8 M
1 CXotta » . 1 . #A . . . . . . . . C » . . ^ ^ 5 # . . . . . .C . .♦
1 aoo# .............» c » . .......e.#.
X ^Xoaa . .A. .A........CT..iBP...... .c...ce
4 7 4 C -8
2 CXooaa . .A. .1. ...... .c. . . f . . .... .c. .
Pll
t a o o M
331»
3 a
SS4X t a m t
1 a
X aOBa ..». .A. ...... .......C. ..
94
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table 4. Summary of nucleotide substitutions as presented
in Figure 10*.
pol env
SU^
m " ,u. M l HI
ABCO EFO
Position Substitution Location*
Amino Acid
Change
1 A-»G 191 I-»V
2 T-*C 225 I-*T
3 K-*G 266 T-^A
4 hr*G 278 Z-*H
5 C-^T 349
—
6 G-+A 379
-
7 GhC 392 A-»P
8 T-C 457
-
9 C-»T 460
—
10 A-K5 498 K-»R
11 C-À 511
—
12 A-*G 554 N-»S
13 C-*T 615 T-*-I
14 &+A 685
—
15 C-»T 810 P-*L
16 T-*C 867 I-*T
17 C-*T 878 p-»S
18 C-*T 940
—
19 A-^ 1016 T-*A
“Nucleotide numbering based upon reported CFE-6 sequence
(Kumar et al., 1989). *Substitutions found conserved
between FeLV-B isolates and all in vivo-derived clones are
designated in Bold typeface.
95
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
tumors were analyzed. Those twelve positions: 1, 2, 2,
7, 8, 12, and 14-19, as presented in Fig. 10, are also
noted by bold typeface in Table 4. At the remaining
seven positions, these substitutions were observed in
three of four cats studied, but were not strictly
conserved. For example, at position 4 in Fig.IDA, the
A-K5 transition mutation was observed in all clones from
tumor DNA of cats Pll, 3315, and 3541, as well as sera
from cat 4746-5, but this mutation was not consistently
observed in bone marrow (BM) or tumor DNA of cat 4746-5.
A similar pattern was also apparent at positions 5, 6,
and 9, as shown in Fig. lOA. Seven nucleotide mutations
did not result in amino acid changes, yet three of these
changes (at positions 8, 14, and 18) were strictly
conserved in nucleotide sequences for known FeLV-B
isolates as well as in all in vivo-derived clones
analyzed here. The remaining four synonymous mutations
(at positions 5, 6, 9, and 11), known to be conserved in
the three molecularly cloned FeLV-B isolates, were also
observed in the majority of in vivo-derived clones
analyzed here; however, these four changes were not
strictly conserved in all the in vivo-derived clones
(Fig. 10). Similar to results reported by Rohn et ai.
(1994), the majority of these substitutions were
transition mutations, with only two transversion
96
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
mutations noted. Transition mutations comprised 89%
(17/19) of the substitutions, with A-K3 type mutations
predominating (32%), followed by C-»T (26%), T-K3 (16%),
and G-^A (11%).
Since nucleotide sequences for the rFeLV mixture
were known, those nucleotide changes observed between
clones from the rFeLV inocula and the rFeLVs recovered
from cat 4746-5 tissues suggested these changes most
likely arose in vivo. It is also interesting to note
that these changes occurred at the earliest timepoint (8
weeks pi) at which rFeLVs could be detected in the sera,
suggesting these changes were important in the early
evolution of rFeLVs for in vivo replication efficiency.
The eunino acid changes noted between the rFeLV inocula
and those rFeLVs recovered following in vivo infection
are presented in Fig. 11. For purposes of illustration,
only site G recombinants are shown. While none of the
eight amino acid changes were present in the rFeLV
inocula, all eight changes were observed in the rFeLVs
recovered from cat 4746-5 following in vivo co—infection
with FeLV—A helper virus. These amino acid changes
corresponded to eight of ten emino acid residues
conserved between the three cloned FeLV-B isolates.
Curiously, in regard to the two remaining amino acid
residues conserved in the FeLV-B isolates (ie. Met73 and
97
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Argl46 ), the nucleotide substitutions leading to these
two amino acid chsuiges were observed in in vivo-derived
rFeLV clones from the naturally-occurring LSA, Pll, as
well as clones from the two FeLV—A/Rickard-infected
cats, 3315 and 3541, and in clones from sera of cat
4746-5, but not in all bone marrow and tumor-derived
clones from 4746-5. The significemce, if any, for this
difference in tissue expression is unclear. This
difference may merely be an artifact or potentially
represent some pattern of tissue-specific expression.
98
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Fig. 11 Sequence comparison showing changes observed
between the rFeLV inocula and recombinant clones
recovered from various tissues of cat 4746-5 co-infected
wito the rFeLV inocula along with FeLV-A/Rickard.
Alignment is presented relative to enFeLV CFE-6 env for
the mature SU (Kumar et al., 1989) and identity to CFE6
indicated by (.) *For purposes of illustration, only
sequences for site G recombinants are presented.
Sequences are also compared to the known FeLV-B
isolates: FeLV-B/Gardner-Arnstein (FeLV-B/GA) (Nunberg
et al., 1984), FeLV-B/Snyder-Theilen (FeLV-B/ST)
(Nunberg et al., 1984), and FeLV-B/Rickeurd (FeLV-B/RIC)
(Elder and Mullins, 1985). A total of eight amino acid
changes, highlighted in gray, were observed in clones
from cat 4746-5 within the region analyzed. All eight
of these changes correspond to sequences reported for
the three known FeLV-B isolates. The two amino acid
residues. Met73 and Argl46, which are conserved in
cloned FeLV-B isolates, but were not strictly conserved
in all rFeLV clones from cat 4746-5, are also noted.
The amino acid residues highlighted by dashed lines
contain locations of nucleotide residues which were
synonymous or "silent" mutations also conserved in all
clones analyzed and which corresponded to conserved
nucleotide residues present in all exogenous FeLV-B
isolates. Varieüale regions II-V, contained within the
sequence presented, are noted for reference and the
locations of the various 3' recombination crossover
sites as previously characterized (Sheets et al., 1992)
are also shown for reference.
99
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
g&M ÎTilîil .^ V * .
3 m 8 *
\ w * "H m
100
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
DISCUSSION
Given that earlier studies indicate FeLV-B viruses
arose via recombination between ecotropic FeLV-A and
enFeLV-like env sequences (Elder and Mullins, 1983;
Kumar et al., 1989; Overbaugh et al., 1988; Rohn et al.,
1994; Sheets et al., 1992; 1993; Roy-Burman, 1996), a
curious disparity exists between cloned enFeLV-like
sequences and corresponding env sequences for exogenous
FeLV-B isolates; a disparity which is conserved in all
three reported FeLV-B isolates. Here we report that
nucleotide chemges at a total of 19 positions were
observed in the SU portion of the env gene between in
vivo-derived rFeLV clones emd corresponding enFeLV
sequence and, additionally, this set of 19 point
mutations correspond to sequences conserved between the
three cloned and sequenced FeLV-B isolates. We also
report here that this set of 19 point mutations was
observed in tumor DNA of all four cats, whether the
tumor source was experimentally-induced or naturally-
occurring.
Within the group of three experimentally-induced
tumor-bearing cats, one (4746-5) had been co—infected
with FeLV-A/Rickard emd the rFeLV pool, while two (cats
3315 and 3541) had received FeLV-A/Rickard alone. In
studies involving cat 4746-5, nucleotide sequences were
101
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
known for the rFeLV pool used as challenge virus,
allowing us an opportunity to monitor the evolution of
nucleotide changes over the timecourse of in vivo
infection. We observed that when the rFeLV pool was
administered along with FeLV-A helper virus, a total of
19 nucleotide substitutions were detected within the in
vivo-derived rFeLV env in regions encompassing those
portions derived from enFeLV elements. Furthermore,
these substitutions were observed at the earliest
timepoint at which rFeLVs could be detected in the sera,
suggesting there is strong selection occurring during
the early in vivo evolution of rFeLVs. Regarding the
rFeLVs detected in clones from the two FeLV-A/Rickard-
infected cats, while we cannot completely rule out the
possibility that the FeLV-A/Rickard inoculum may have
harbored rFeLVs, no rFeLVs were detected in this
inoculum using a nested PCR strategy, suggesting that
the rFeLV clones recovered from the FeLV-A/Rickard-
infected cats arose de novo during the course of in vivo
infection. Within this set of 19 point mutations, a
group of twelve mutations was strictly conserved in all
recombinant clones analyzed and also correspond to
conserved sequences observed between the reported FeLV-B
isolates. Three of these strictly conserved mutations
were "silent", suggesting these changes may be important
102
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
for RNA secondary structure ^ perhaps leading to more
efficient genomic packaging or virion assembly.
A current paradigm describing the complex process
of genetic diversity associated with retroviruses
suggests that retroviruses exist within a dynamic
equilibrium of heterogenous but closely related
variants, poised to allow subpopulations of variemts to
predominate under environmental (selective) pressures
(Coffin, 1992; Katz and Skalka, 1990). As a result,
moleculaur clones obtained from a single infected host
may represent a "snapshot" of retroviral species
predominating under specific selective conditions, and
such molecular clones can thus only be defined as an
"isolate" or "strain". In spite of this tendency for
considerable genetic diversity, there are selective
forces imposing limitations to sequence variation, often
by structural constraints and functional competency
(Coffin, 1992; Katz and Skalka, 1990; Domingo et ai.,
1996). With this in mind, some interesting points can
be noted in regard to the set of 19 point mutations
found scattered throughout the SU-encoding env gene of
in vivo-derived rFeLV clones. First, this set of
mutations did not segregate to emy particular region,
but instead these mutations were found scattered
throughout both conserved and variable regions,
103
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
encompassing an area of about 80% of the N-terminal
portion of the SU. Second, as observed in sera samples
obtained throughout infection timecourse for one cat,
this set of mutations was strongly selected in that all
19 of these point mutations were observed at the
earliest timepoint at which rFeLVs could be detected.
Third, this set of mutations resulted in complete
sequence identity with known FeLV-B isolates. Taken
together, these results would suggest that this set of
19 mutations may represent functional and structural
constraints required for the conversion of recombinant
FeLVs into exogenous FeLV-B viruses.
104
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
CHAPTER 5
Recombinant Feline Leukemia Virus Variants
Demonstrate Altered In. Vivo Cell Tropism
ABSTRACT
Many reports describe the lack of efficient in vivo
infactivity for FeLV-B in the absence of FeLV-A, as well
as the fact that FeLV-B is always found in association
with FeLV-A in natural infections. Potential helper
functions suggested for FeLV-A include replication
rescue for replication-defective viruses, as well as
assisting in horizontal spread and/or immune evasion for
FeLV-B viruses. In an in vivo study employing cats
challenged with rFeLVs amd FeLV-A helper virus, we
recently observed multiple nucleotide substitutions
scattered throughout the SU-encoding env domain in
recombinant FeLVs (rFeLVs) recovered from tissues in one
of these cats. The substitutions found in these in
vivo-derived rFeLVs resulted in complete sequence
identity with known FeLV-B isolates, raising the
question: can these substitutions occur in the absence
of FeLV-A co-infection. In this current study we
confirmed the subpopulation of rFeLVs employed here was
replication-competent and demonstrated FeLV-B-like in
vitro functional activity. We next investigated whether
105
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
nucleotide substitutions could occur in vivo in the
absence of FeLV-A. Here we demonstrate the presence of
rFeLV provirus in various tissues of rFeLV-infected
cats, but without detectable viremia, apparent rFeLV
expression, or significant seroconversion. However,
altered cell tropism was observed. Using indirect
immunofluorescence detection of FeLV antigen, three of
four rFeLV-infected cats showed focal infection in the
marginal zones (T-cell areas) of splenic follicles while
a FeLV-A/Rickard-infected cat showed viral antigen
associated with the germinal center (B-cell area) of
splenic follicles. These results indicate that at least
this subpopulation of rFeLV variants is able to
establish latent or chronic in vivo infection, albeit a
regressive, self-limiting one, capable of apparent
altered cell tropism. Subsequent nucleotide sec[uence
analysis of proviral DNA obtained from bone marrow of
these rFeLV-infected cats showed none of the nucleotide
substitutions observed when rFeLVs were co-infected with
FeLV-A, suggesting rFeLV viruses are capable of
esteüDlishing a limited in vivo infection in the absence
of FeLV-A co-infection, but that efficient productive
infection, such as occurs in the presence of FeLV-A, is
probably necessary for the in vivo conversion of rFeLV
nucleotide sequence to that of FeLV-B sequence.
106
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
INTRODUCTION
FeLV is a natxirally-occurring, contagious
retrovirus of outbred cats, capable of causing both
proliferative (lymphoma) and suppressive
(immunodeficiency) disorders (Hoover et al., 1991).
FeLV-A is the dominant subgroup encountered in nature;
being readily transmissible, highly conserved and
demonstrating low pathogenicity (Jarrett and Russell,
1978; Stewart et al., 1986; Donahue et al., 1988).
FeLV-A is also known to give rise to FeLV-B viruses via
recombination of the FeLV-A env gene with corresponding
endogenous FeLV (enFeLV) sequences carried in the cat
genome such that, depending upon the site of
recombination, a mixture of variant FeLV-B genotypes
potentially arise over the course of in vivo infection
(Elder and Mullins, 1983; Overbaugh et al., 1988; Soe et
al., 1985; reviewed in Coffin, 1992).
Since FeLV-B is always isolated with FeLV-A in
natural infections, plus the observation that FeLV-B
viruses appear to be more pathogenic than FeLV-A in that
FeLV-B is overrepresented in thymic lymphoma (Hoover et
al., 1976; Jarrett and Russell, 1978; Jarrett et al.,
1978; Tzavaras et al., 1990; Sheets et al., 1993;
Tsatsanis et al.,1994), it is thought that FeLV-A acts
as helper virus during in vivo infection. One potential
107
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
helper function provided by FeLV-A is enhanced
replication efficiency. It is possible variant FeLV-B
genomes may be either replication-competent or
replication-defective so that concurrent FeLV-A
infection is necessary for rescue. It is also proposed
that FeLV-A co-infection may aid in the horizontal
spread and/or evasion of host immune surveillance
(Hoover and Mullins, 1991). Due to the apparent
dependence on concurrent FeLV-A co-infection in order
for FeLV-B to establish efficient in vivo infection, few
studies have addressed the pathogenesis of FeLV-B
viruses in the absence of FeLV-A (Neil et al., 1991).
Earlier, our lab described the generation of a
mixture of FeLV-B-like recombinants which were
propagated through a process of forced in vitro
recombination and selection. Within this mixture, a
subpopulation of variants were found to be resistant to
neutralization by a monoclonal antibody (mAb)
recognizing an exogenous epitope present in the middle
of the virion surface glycoprotein (SU), an epitope
highly conserved in all exogenous FeLV isolates,
regardless of subgroup, and previously characterized as
a major neutralizing determinant (Elder et al., 1987;
Grant et al., 1983; Nunberg et al., 1984; Sheets et al.,
1992). These variants harbored nucleotide mutations
108
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
within this epitope sequence, presumably allowing them
to escape recognition by this mAb. In another study,
our 12Ü3 described detection of rFeLV proviruses in tumor
DMA obtained from naturally-occurring feline
lymphosarcomas (Sheets et al., 1993). Sequence analysis
of DMA obtained from these rFeLV proviruses showed that
3 of 15 thymic tumors harbored mutations in this same SU
region of the env gene, providing some anecdotal
evidence suggesting that such mutations may potentially
permit escape from host immune surveillance, allowing
persistent infection to occur. In another report, a
molecularly cloned natural FeLV-B variant resistemt to
neutralization by this seune mAb, was shown to contain a
single nucleotide mutation three amino acids upstream of
this mAb-binding epitope (Micolaisen-Strouss et al.,
1987), providing further evidence that such naturally-
occurring variants containing mutations in the vicinity
of this conserved region may have a role in FeLV-
mediated disease. Finally, our lab recently described
results of an In vivo study in which three of six cats
co-infected with a mixture of rFeLVs and FeLV-A
developed thymic lymphoma (Pandey et al., 1995). We
observed multiple nucleotide substitutions scattered
throughout the SU-encoding env domain in the in vivo-
derived rFeLV clones. This set of 19 point mutations
109
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
resulted in complete sequence identity with known FeLV-B
isolates, with one of these nucleotide substitutions
occurring within this same highly conserved SU epitope.
These recent results prompted us to wonder whether such
nucleotide substitutions would happen in the absence of
in vivo FeLV-A co-infection. In the present study we
exploited the use of this mAb by selecting a
subpopulation enriched for such rFeLV variants and
describe their in vitro characterization, their pattern
of in vivo cell tropism, and results from nucleotide
sequence analysis of rFeLV proviruses recovered from
cats infected with these variants in the absence of
FeLV-A.
110
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
MATERIALS AND METHODS
Viruses
The rFeLV pool was previously described (Pandey et
al., 1995). This pool consisted of multiple recombinant
species with major subpopulations shown to have
crossover junctions at sites A, B, E, F and G, such that
starting from the 5' end of env, site A recombinants
harbored the least amount of enFeLV substitution and
site 6 recombinants contained the most enFeLV
substitution (Sheets et al., 1992). To obtain a
subpopulation of viruses from the rFeLV pool enriched
for site E to site G recombinants, the rFeLV pool was
treated with C1108, a monoclonal antibody (mAb)
recognizing the highly conserved exogenous MGPNL epitope
(Elder et al., 1987; Grant et al., 1983). Briefly, 1.5
ml of the rFeLV pool containing 4 x 10^ 50% cell culture
infectious dose (CCID50) was combined with 630 ng C11D8
(1.5 ml) to obtain a final mAb concentration of 210
(Mg/ml, followed by incubation for one hr at 37"c to
facilitate mAb:virus binding. The 210 f*g/ml mAb
concentration was previously determined to neutralize
the vast majority of virus (data not shown), leaving a
subpopulation of variant rFeLVs enriched for site E to
site G recombinants, and hereafter designated rFeLV/D8
variants. Virus stocks were cell culture supernatants
111
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
generated from human fibrosarcoma HT1080 cells as
previously described (Bechtel et al., 1994)).
Additional viruses used in this study were, FeLV-B/GA
(Nunberg et al., 1984 and FeLV-C/Sarma (Reidel et al.,
1986), and FeLV-A/61E (Donahue et al., 1988).
In vitro growth studies
To determine whether the rFeLV/D8 pool replicated
efficiently in vitro and demonstrated cell trop ism
similar to FeLV-B viruses, various human and feline cell
lines were infected with the rFeLV/D8 pool. Parallel
growth studies were conducted with FeLV-B/GA and FeLV-
C/Sarma viruses, serving as controls. Cell lines
employed in the growth studies were human HT1080 cells,
feline H927 cells, and the feline 3201B T-cell lymphoma
cell line, virus infection was accomplished using virus
at 1 CCIDbo/mI diluted in oMEM without FBS (MEM-SF) as
previously described (Bechtel et al., 1994). Cells were
maintained for up to four weeks to determine potential
viral infactivity. Periodically, culture supernatant
was collected, clarified by centrifugation and virus
production assessed by enzyme-linked immunosorbent assay
(ELISA) (Synbiotics) as previously described (Bechtel et
al., 1994). Cell counts and viability were determined
via hemocytometer using trypan blue dye exclusion
112
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
concurrently with the ELISA. Triplicate determinations
were performed for each time point and values averaged.
The level of virus production was expressed as a ratio
of the average ELISA OD450 vs number of viable cells x
10® as previously described (Bechtel et al., 1994).
In vitro neutralization studies
The in vitro neutralizing activity for the rFeLV/D8
pool was evaluated in parallel with FeLV-A/61E, FeLV-
B/GA, and FeLV-C/Sarma, using immune sera obtained from
three different cats. These cats, designated 4936, 4939
and 4942, had been challenged with FeLV-A/Rickard at 20-
24 weeks of age. As determined by ELISA and
immunofluorescence assay (IFA), cat 4936 was transiently
antigenemic between 3-5 weeks pi, cat 4939 was not
detectably antigenemic at any point, and cat 4942 was
weakly antigenemic only at 3 weeks pi. All three sera
had FeLV antibody titers of 1:512 to 1:1024. Sera,
collected at 14 weeks pi for use in these neutralization
assays, were filter sterilized (0.22 nm Millipore
syringe filters) and heated 30 min at 56"c to inactivate
any heat-labile, non-specific inhibitory serum
components. The rFeLV/D8 pool was propagated in both
human HT1080 cells and feline H927 cells so that the
H927-derived rFeLV/D8 virus could be assayed in parallel
113
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
with the ecotropic FeLV-A/61E, while the HTIO80-derived
rFeLV/D8 pool could be assayed in parallel with the
polytropic FeLV-B/GA and FeLV-C/Sarma viruses. Cells
were seeded into 96-well, flat-bottomed plates using 200
cells/well two days prior to virus infection. On the
day of infection, 6-8 serial twofold pre-dilutions of
sera were made in MEM-SF to target a final virus
neutralization range of 90% to 10%. Virus was also pre
diluted in MEM-SF to 2 CCIDgo/Ail, and equal volumes (0.3
ml) of serum and virus dilutions were combined to obtain
a virus concentration of 1 CCIDso//il in the final
serum:virus mixture. To examine any potential serum
effect on the cells, immune sera alone (1:2 in MEM-SF)
served as a control, while MEM-SF diluent served as
negative (cell) control and virus at 1 CCIDsq/ffl (in
MEM-SF) served as positive (virus) control. The serum-
virus mixtures and their controls were incubated one hr
at 37“c to allow virus neutralization to occur. To
confirm the virus was actually 1 CCIDs@/fil in the serum-
virus mixture, a virus titration was performed
concurrently with each neutralization assay. Serum-
virus mixtures and controls were added to at least
quadruplicate wells, and the remainder of the
neutralization assays and virus infactivity assays were
performed and assessed by ELISA as previously described
114
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
(Bechtel et al., 1994). ELISA values for replicate
wells were averaged and a percent neutralization derived
for each dilution using the formula:
% NeUt ^ n “ avq Vlrna control HUax - avo tm a t dll*n ELISA X 100
avg. Virus Control BLISX
A neutralization curve was generated for each serum-
virus mixture by plotting serum dilutions versus percent
neutralization. A 50% neutralization titer was
predicted from each curve using exponential curve-
fitting regression analysis.
In vivo cat studies
Two litters (4 kittens/litter) of specific
pathogen-free (SPF) kittens were challenged
intraperitoneally with the rFeLV/D8 pool using 3 x 10*
focus-forming units (ffu) at 4 days (Litter I, cats
4684-1 to 4684-4) and 8 days of age (Litter II, cats
4591-1 to 4591-4). Blood samples were collected at
biweekly intervals except the first interval, which was
omitted due to the animals' small size. All
inoculations and specimen collections were performed
under ketamine (25 mg/kg) anesthesia. Complete blood
cell counts were determined and samples evaluated for
viremia and anti-FeLV titer. Viremia in whole blood
smears was determined by indirect IFA and ELISA (Hardy
115
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
et al.y 1973; Hoover et al., 1978). Anti-FeLV titers
were determined by immunoblot and by live-cell IFA using
the chronically FeLV-infected FL-74 cell line as the
antigen source (Essex et al., 1971). Litter I was
euthanized at nine weeks post infection (pi) and
complete necropsies were performed on each animal
including histopathologic evaluation. In addition,
select tissues from Litter I cats were stained by
indirect IFA for FeLV antigen (see below). Bone marrow
was obtained via biopsy from Litter II cats at 45 weeks
pi.
Amplification, cloning and nucleotide sequencing
To confirm the rFeLV/D8 pool consisted primarily of
recombinants harboring site E to G crossover junctions,
high-molecular-weight genomic DNA was extracted from
~10^ HT1080 cells infected with the pool of rFeLV/D8
variants. PCR amplification in the mid-region of the
env gene was accomplished using primers MP-1 and MP-2
(Fig.5, Chapter 3). These PCR products were cloned into
the TA cloning vector and a total of 5 clones were
manually sequenced as previously described (Pandey et
al., 1995). Nucleotide sequences were then compared to
published sequences of CFE-6 and FeLV-A (Donaüiue et al.,
1988; Kumar et al., 1989). To analyze bone marrow
116
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
tissue, obtained via biopsy from Litter II cats at 45
weeks pi, for the presence of rFeLV proviruses,
recombinant-specific PCR products generated from Litter
II DNA samples were also cloned into the TA vector and
white colonies screened by restriction analyses.
Positive clones were manually sequenced with MP-1 and
MP-2 primers as described above cind the sequence for
each of the recombinant env clones compared to those of
CFE-6 (Kumar et al., 1989) and FeLV-A/61E (Donahue et
al., 1988).
To clone the SU-encoding portion of the recombinant
env gene from rFeLV/D8-infected-HT1080 cells (rFeLV/D8
Inocula) and from rFeLV/D8-infected cat 4591-2 bone
marrow DNA, PCR amplification was performed using a 5'
sense primer (H19) recognizing a sequence at the pol/env
junction and the 3' antisense primer RB19 as described
in Chapter 3, yielding a 2.0 Kb product. These 2.0 Kb
PCR products were cloned into the TA vector and screened
by PCR using the recombinant-specific primer set, RB53
and RB17, as described in Chapter 3. To sec[uence the
entire 1.2 Kb SU-encoding region, three recombinant-
specific clones from rFeLV/D8-infected HTIO80 cells, as
well as three recombinant-specific clones from cat 4591-
2 BM DNA were selected, and automated sequencing
117
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
performed using fluorescence-based cycle sequencing as
described in Chapter 3.
IFA detection of viral antigens in Litter I cat tissues
FeLV antigen was detected in tissues via an
indirect IFA technique (Rojko et al., 1978). The
following tissues, collected at necropsy from Litter I
cats, were fixed in absolute methanol: liver, kidney,
spleen, lymph node, bone marrow, large and small
intestine, testis/ovary, urinary bladder, tonsil,
salivary gland, pancreas, thymus and brain. The primary
antiserum was a polyclonal goat anti-FeLV (subgroup A, B
and C); the secondary antiserum was a fluorescein-
conjugated rabbit anti-goat IgG diluted in phosphate-
buffered saline solution containing 20% SPF cat serum as
a blocking agent. Tissue slides were counterstained
with Evan's blue.
PCR detection of recombinant proviral DMA in tissues of
cats challenged with rFeLV/D8 variants
High-molecular-weight genomic DNA, isolated from
necropsied Litter I cat tissues (cats 4684), were snap
frozen in liquid nitrogen. PCR amplification was
carried out using a 5' primer specific for endogenous
sequences (PRB-1) and a 3' primer specific for FeLV-A
118
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
(RB-17) encompassing the major neutralizing epitope and
all known FeLV-A recombination sites (Sheets et al,,
1992). The 3' primer, RB-17, corresponded to the
sequence at the boundary of SU and TM. Genomic DNA was
also extracted from bone marrow tissues obtained via
biopsy from Litter II cats (cats 4591 series) at 45
weeks pi. PCR amplification was performed using the
endogenous-specific 5' primer, RB53, and the 3' FeLV-A-
specific primer, RB-52, (Sheets et al., 1993) to amplify
recombinant-specific products, yielding a 1,393 bp
fragment. PCR products were electrophoresed on a 2%
agarose gel in TAB buffer and the gel stained with
ethidium bromide to visualize the bands.
Total RNA isolation from tissues and reverse
transcription-PCR (RT-PCR) amplification
Total RNA isolation from tissue samples was
performed using Clontech's total RNA isolation kit (Palo
Alto, CA). Reverse transcription (RT) reaction was
performed using Moloney murine leukemia virus reverse
transcriptase (M-MLV RT) (Gibco-BRL, Gaithersburg, MD)
and oligo(dT) primer (Promega, Madison, WI) to
synthesize cDNA. For each sample studied, 5 ug total
RNA was used in a 50 ul RT reaction. RT-PCR without RNA
or without reverse transcriptase enzyme were included in
119
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
each experiment, serving as negative controls. RT-PCR
analysis of glyceraldehyde-3-phosphate dehydrogenase
(GÀPDH) expression served as an additional control. The
previously described GAPDH primers (Gao et al., 1993),
suitable for both human and rat systems, were employed
with cat cDNA and yielded a 598 bp product. In all PCR
analyses, RT(-) controls were included to assure no
genomic DNA contamination of the cDNA samples. A nested
primer PCR strategy was employed for detection of FeLV
expression in tissue cDNA samples. For detection of
recombinant-specific FeLV fragments, first-round
amplification utilized the endogenous-specific 5' primer
RB56 and FeLV-A-specific 3' primer RB52 (Sheets et al.,
1993). One ul PCR product from the first round of
amplification was then amplified in a second round using
the nested primers, endogenous-specific 5' primer RB53
and FeLV-A-specific 3' primer RB19 (Sheets et al.,
1993), yielding a 1,376 bp fragment. PCR products from
the second-round of amplification were analyzed by 2%
agarose gel electrophoresis and DNA bands visualized by
ethidium bromide staining.
120
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
RESULTS
In vitro selection and nucleotide sequence analysis of
rFeLV/D8 varicuits following C11D8 mAb treatment
À subpopulation of rFeLVs was prepared in vitro
from the parental rFeLV pool using the MGPNL-recognizing
CUDS mAb to select against FeLVs expressing this
conformational epitope, leaving a subpopulation of
variant rFeLVs (rFeLV/D8s) enriched for non-MGPNL-
expressing viruses. Nucleotide sequence analysis of
five clones from this rFeLV/D8 pool was performed and
the predicted amino acid sequence for these five clones
was aligned with respect to FeLV-A/6IE and CFE-6 in the
SU region encompassing the epitope. Results summarized
in Fig. 12 show four of five clones had the enFeLV-
derived Pro217 amino acid at the fifth position in the
pentapeptide, and all five had the enFeLV-derived Het205
upstream of the pentapeptide. One clone (Clone 4) had
sequence homologous to FeLV-A within and downstream of
the epitope sequence but endogenous sequence upstream,
indicating this clone was a site D recombinant species.
In addition, one clone had a Gly214 to Arg214
substitution identical to an alteration previously
reported (Sheets et al., 1992). These results indicated
this pool was indeed enriched for non-MGPNL-expressing
variants.
121
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
CD
■ D
O
Q.
C
g
Q.
■D
CD
C/)
C/)
CD
8
■D
( O '
3 .
3"
CD
CD
■ D
O
Q .
C
a
O
3
"O
o
CD
Q .
200
FeLV-A/6IE VSRQVSTITPPQ
CFE-6
rFeLV/D8 pool
Clone 1
Clone 2
Clone 3
Clone 4
Clone 5
PDQKPPSRQSQTGSKVQTQRPQ----- TNESAPRSVAP-----TTVGPKRI
............IE.R.IPHH..GNGGTPGITLVNASI..LSTPV.PAS....
..lE.R.IPHH..GNGGTPGITLVNASI..LSTPL.PAS.
..IE.R.IPHH..GNGGTPGITLVHASI..LSTPL.PAS.
..lE.R.IP
..IE.R.IPHH..GNGGTPGITLVNASI..LSTPV
Fig. 12. Amino acid sequence comparison of FeLV-A/6IE and enFeLV CFE-6 with five
clones isolated from the variant rFeLV/D8 pool. Sequence numbering is from FeLV-A/6lE
beginning from the mature SU; (...) indicates homology to 6IE, while (--) indicates
gaps in sequence alignment (Kumar et al., 1989). The major neutralizing epitope is
highlighted in gray.
■D
CD
( / )
( / )
to
to
In vitro neutralization and cell tropism for rFeLV/D8
variants were similar to those for FeLV-B/GA
Neutralization rates for exogenous FeLVs and
rFeLV/D8 variants in the presence of immune cat serum
were determined. In multiple independent experiments,
serial two-fold dilutions of heat-inactivated sera from
three cats, previously challenged with a FeLV-A/Rickard
plasma pool, were incubated with different FeLV
preparations. These mixtures were then added to either
human HT1080 cells or feline H927 cells. The presence
of infectious virus was assessed using ELISA for FeLV
p27 antigen. Neutralization rates for the rFeLV/D8 pool
propagated in feline H927 cells was compared with
ecotropic FeLV-A/61E in H927 cells, while neutralization
rates for the rFeLV/D8 pool propagated in human HTIO80
cells were compared with those of the polytropic FeLV-
B/GA and FeLV-C/Sarma in HT1080 cells. A percent
neutralization was calculated for each serum dilution
and a neutralization curve generated for each virus-
serum mixture by plotting serum dilution versus percent
neutralization (data not shown). A 50% neutralization
titer was predicted from each neutralization curve using
exponential curve-fitting least squares regression
analysis. Quality of curve-fit was assessed by
confirming that the coefficient of determination (R^)
123
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
was >0.90. À summary of that analysis is presented in
Table 5. As expected, with each of the three cat sera,
exogenous FeLV-A and FeLV-C were the most strongly
neutralized viruses, having 50% neutralization titers
ranging from 1:108 to 1:668. The exogenous FeLV-B
neutralization rates ranged from 1:11 to 1:57 for the
three sera, and were comparable to the rFeLV/D8
preparations, which had 50% neutralization titers
ranging from 1:15 to 1:91. Interestingly, the rFeLV/D8s
derived from human HT1080 cells consistently
demonstrated 50% neutralization titers which were at
least twofold greater than the rFeLV/D8s derived from
feline H927 cells.
In vitro cell tropism for the rFeLV/D8s was also
assessed using cultures of feline T-lymphoma cells
(3201B), feline fibroblasts (H927) and human
fibrosarcoma cells (HT1080), in parallel with either
FeLV-B/GA or FeLV-C/Sarma. Results summarized in Fig.
13 indicated cell tropism for the rFeLV/D8 variants was
similar to those previously reported with its parental
recombinant pool (Sheets et al., 1992), and that the
rFeLV/D8 variants demonstrated equivalent cell tropism
and growth characteristics to the polytropic FeLV-B/GA
virus. Both the rFeLV/D8s and the FeLV-B/GA grew
equally well in human HT1080 cells and feline H927
124
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
fibroblasts ^ while neither the rFeLV/D8s nor the FeLV-
B/GA grew in the 3201B cells. Only FeLV-C/Sarma was
able to grow in the feline T—lymphoid 3201B cells.
125
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
TABLE 5
Summary of FeLV 50% Neutralization Titers
Virus
Description
Serum
4936
Ra.
Serum
4939
Ra.
Serum
4942
FeLV-A
FeLV-B
FeLV-C
rFeLV/D8 Pool"
rFeLV/D8 Pool®
1:122
1:40
1:183
1:91
1:45
0.95
0.92
0.97
0.97
0.97
1:668
1:11
1:108
1:67
1:15
0.98
0.99
0.99
0.98
0.91
1:391
1:57
1:252
1:74
1:34
0.91
0.96
0.97
0.96
0.92
“ Coefficient of determination, indicating quality-of-
fit
* * rFeLV/D8 pool prepared in human HTIO80 cells
° rFeLV/D8 pool prepared in feline H927 cells
126
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Fig. 13 In vitro cell tropism for rFeLV/D8 variants as
compared to FeLV-B/GA emd FeLV-C/Sarma. Cell cultures
of human HT1080 cells, feline H927 cells, and feline T-
cell lymphoma cell line, 3201B, were infected as
described in the text. The X-axis is number of days
post-infection (pi) of the cell cultures and were
maintained for 21 to 27 days pi. The Y-axis represents
a ratio of the OD450/# of viable cells as measured by
trypan blue dye exclusion as detailed in the text.
127
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Fig. 13. rPeLV/D8 In Vitro Cell Tropism
%
X
iS
"5
o
JQ
(0
>
*
o “
o
HT1080 Cells
5 10 15 20 25 30
H927 Cells
0.5
o.oé — Or-
o
o
■
□
1 I I I
5 10 15 20 25 30
2 -
1 -
/\
3201B Cells
\
\
/ m
0 ?---1 -^ n
5 10 15 20 25 30
Days Post Infection
UNINFECTED
FeLV-B
FeLV-C
rFeLV/D8
128
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
In vivo challenge experiments with rPeLV/D8 variants
Two litters, four cats each, were challenged with
rFeLV/D8 variants at 4 days (Litter I) and 8 days of age
(Litter II). Neither FeLV viremia nor antibody to FeLV
was detected in any of the blood samples collected
during the study. The four kittens comprising Litter I,
designated cats 4684-1 to 4684-4, were euthanized and
necropsied at 9 weeks pi. Tissues were collected and
analyzed for histologic lesions and viral antigen
expression. Hematology results were in the normal
range. Histologically, all four cats had prominent
white pulp in the spleen, and reactive lymph nodes. One
cat showed lymph node medullary hypoplasia. The
thymuses were well-developed, with beginning signs of
involution (degeneration) in one cat. The pancreases of
these cats showed a more nodular appearance than
expected for their age. The four kittens comprising
Litter II, designated cats 4591-1 to 4591-4, were
inoculated with rFeLV/D8s as 8 day old neonates.
Virologie, immunologic and hematologic parameters were
initially monitored weekly and later at monthly
intervals for 90 weeks. None of the plasma samples were
positive for FeLV p27 antigen as measured by ELISA.
Viremia was not detected in any of the blood samples
collected during the study (data not shown). Likewise,
129
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
none of the samples had measurable antibody as
determined by indirect IFA. In spite of a lack of
detectable viremia or anti-FeLV titers, virus antigen
was detected in lymphoid tissue of Litter I cats.
FeLV antigen distribution in rFeLV/D8-inoculated cats
Tissues collected at necropsy from Litter I at 9
weeks pi were stained for viral antigen expression.
Three of four cats showed antigen in the spleen
primarily in the marginal zone surrounding some splenic
follicles (Fig. 14A). By comparison, a positive control
spleen from a cat infected with FeLV-A/Rickard strain
showed very strong reactivity in the germinal center
area of the follicle (Fig. 143) rather than in the
marginal zone. Lymph nodes in three of four rFeLV/D8-
inoculated cats displayed antigen-positive cells
scattered throughout the paracortical region and cortex.
One of four cats showed staining of duct epithelium in
the salivary gland (Fig. 14C).
Recombinant proviruses were detected in DNA of rFeLV/D8-
infected cats
PCR amplification was performed to assess the
presence of recombinants in cats challenged with the
rFeLV/D8 pool.
130
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 14. Indirect immunofluorescence detection of
FeLV antigen expression in tissues from FeLV-infected
kittens. A) Spleen from an rFeLV/D8-infected kitten (9
weeks pi) illustrating FeLV antigens in T-cell areas
surrounding a germinal center, as contrasted to a (B)
spleen from an FeLV-A/Rickard-infected cat showing FeLV
antigen in a germinal center (B-cell areas). C)
Salivary gland from an rFeLV/D8-infected kitten and (D)
a spleen from an uninfected kitten.
131
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
For Litter I tissue DNA, recombinant-specific
primer set PRB-1 and RB-17 was used to amplify the
recombinant proviral elements. A product of expected
size (-870 bp) was detected in most tissues analyzed.
These results are summarized in Table 6. Spleen and
liver from all four cats were positive for this PCR
product, while thymus and bone marrow from three of four
cats, and lymph node tissues from two of three cats
tested were also positive for this PCR product.
However, the level of proviral sequences was apparently
quite low since in order to see a band in the gel using
20% of the total PCR product, 50 PCR amplification
cycles had to be performed. PCR was also performed
using recombinant-specific primers RB53 and RB52 (Sheets
et al., 1993) on DNA isolated from bone marrow tissue
obtained via biopsy at 45 weeks pi from the rFeLV/D8-
infected Litter II cats (4591-1 through 4591—4). A
product of expected size (-1.37Kb) was detected in all
four cats. A representative experiment showing this
result is illustrated in Fig. 15.
132
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
TABLE 6
Summary of PCR Analyses for Recombinant Proviruses in
rFeLV/D8-infected Litter I Cats
Tissue
Cat
4684-1
Cat
4684-2
Cat
4684-3
Cat
4684-4
Liver + + + +
Spleen + + + +
Thymus ND“ + + +
Bone Marrow + + + ND“
Lymph Node + ND*
+ NA**
“ PCR product not detected with Ethidium Bromide
staining
* * Tissue not available
133
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Uninfected DNA
FeLV-A
rFeLV
] r Water
1.6 Kb
Cat# 12 3 4
Fig. 15. PCR analysis of recombinant proviral DNA in
bone marrow of Litter II rFeLV/DS-infected cats 4591-1
through 4591-4 at 45 weeks pi. Listed from left to
right: M, 1Kb DNA Ladder; cats 4591-1 to 4591-4 at 45
weeks pi; uninfected cat 3488-2 DNA; FeLV-A-infected
feline H927 cell DNA; rFeLV-infected human HT1080 cell
DNA; water. PCR amplification product size was 1.37 Kb.
134
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
In addition, none of the four r FeLV/D8 - inf ected cats
seroconverted or became viremic throughout the course of
the study; none had detectable rFeLV proviral DNA in
peripheral blood lymphocytes (buffy coats) at 5 weeks or
32 weeks pi as measured by PCR; and none demonstrated
rFeLV expression in their bone marrow at the 45 weeks pi
timepoint as measured by nested RT-PCR (data not shown).
Nucleotide sequence analysis of clones from in vivo-
derlved rFeLV/D8 variants indicated no sequence changes
relative to the parental rFeLVs
The rFeLV PCR products from Litter II bone marrow
DNA at the 45 weeks pi timepoint were cloned and
sequenced. Four to five clones from each of the four
cat were analyzed and the putative amino acid sequences
aligned with respect to those of FeLV-A/61E (Donahue et
sü.,^ 1988) and enFeLV, CFE-6 (Kumar et al., 1989). Fig.
16 summarizes this comparison. All 19 clones had the
enFeLV-derived Pro217 at the fifth position of the
pentapeptide epitope and the enFeLV-derived Met205
adjacent and upstream of the epitope sequence. In
addition, one clone had the Gly214 to Arg214 alteration
as seen in the rFeLV/D8 inoculum (Fig.12).
As described in Chapter 3, multiple nucleotide
substitutions, scattered throughout the SU-encoding env
135
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
CD
■D
O
Q .
C
g
Q .
■D
CD
C/)
C/)
8
■D
3 .
3"
CD
CD
" D
O
Q .
C
a
O
3
"O
O
217
FeLV-A/6IE TVSRQVSTITPPQA
CFE—6 S... ..M.......
FeLV-B/GA S...
Cat 4591-1
1 Clone S...
4 Clones S...
Cat 4591-2
1 Clone 8...
1 Clone S...
3 Clones S...
Cat 4591-3
2 Clones
3 Clones
cat 4591-4
3 Clones S...
1 Clone S...
i PDQKPPSRQSQTGSKVSTQB
• •IB «R«0PHH
• «lE « R «DPHH
..lE.R.IpHH
..lE.R.ipHH
236 240
,
Q-----TNESAPRSVAP---- TTVGPKRIG
.GNGGTPGITLVNA.I..LSTPV.PAS....
.GNGGTPGITLVNA.I..LSTPV.PAS....
...lE.R.
NLK*
...lE.R.
...IB.R.iPHH
*..lE.R.gPHB
... lE.R.BBPHB
...IB.R.SpHB
.GNGGTPGITLVNG.I..LSTPV.PAS.
.GNGGTPGITLVNG.I..LSTPV.PAS.
.GNGGTPGITLVNG.I..LSTPV.PAS.
.GNGGTPGITLVNG.I..LSTPV.PAS.
.GNGGTPGITLVNG.I..LSTPV.PAS.
.GNGGTPGITLVNG.I..LSTPV.PAS,
.GNGGTPGITLVNG.I..LSTPV.PAS.
.GNVGTPGITLVNG.I..LSTPV.PAS.
(D
Q .
T 3
(D
(/)
(/)
Flg. 16. Deduced amino acid sequence comparison of FeLV-A/6IE (Donahue et al., 1988}
and enFeLV CFE-6 (Kumar et al., 1989) with 19 clones obtained from bone marrow DNA of
biopsied rFeLV/D8-infected cats at 45 weeks pi as described in text. Sequence
identity to FeLV-A/6IE is indicated by (.), while changes are indicated by the
appropriate amino acid letter designation. Deletions are indicated by (-) and the
translation termination codon is indicated by an (*). Sequences highlighted in gray
are described in the text.
H»
W
Ol
region, were observed in vivo when rFeLVs were co
infected along with FeLV-A as helper virus. However,
when rFeLV/D8 variants were inoculated in the absence of
FeLV-A, none of these nucleotide substitutions were
noted. Results indicated the rFeLV/D8 preparation
contained enFeLV-like sequence throughout the SU domain
analyzed and that this enFeLV-like sequence was
unchanged in the recombinant-specific clones obtained
from bone marrow DNA of cat 4591-2 after 45 weeks pi
(data not shown).
In summary, the rFeLV/D8 variants demonstrated
efficient in vitro replication and cell tropism
characteristics typical of FeLV-B viruses, but in vivo
challenge with rFeLV/D8 alone resulted in an inefficient
and unproductive infection with undetectable viremia,
minimal seroconversion, no rFeLV/D8 expression, and in
contrast to what was previously observed in rFeLV:FeLV-A
co-infection studies, there were no significant
nucleotide sequence substitutions in the SU domain of
these in vivo-derived rFeLV clones. However, some
altered cell tropism was observed in vivo.
137
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
DISCUSSION
Our lab previously described the generation of a
rFeLV pool through in vitro-forced recombination and the
observation that approximately 25% (6 of 24) of the
clones analyzed from this pool harbored mutations in a
nucleotide seguence encoding a highly conserved epitope,
allowing these variants to evade neutralization by a mAb
recognizing this epitope (Elder et al., 1987; Grant et
al., 1983; Sheets et al., 1992). In another study, our
lab described detection of rFeLV proviruses in DNA from
naturally-occurring thymic LSAs in which 20% (3 of 15)
of the tumors contained mutations in this same region
(Sheets et al., 1993). These results along with the
work of others (Nicolaisen-Strouss et al., 1987) led us
to speculate that variant rFeLV subpopulations
containing altered sequences in the vicinity of this
conserved epitope may be able to evade host immune
surveillcince and establish persistent in vivo infection.
The rFeLV/D8 variants employed in this study, were
selected using the C11D8 mAb which recognizes this
conserved epitope, resulting in a subpopulation enriched
for viruses harboring alterations in this region. Since
we had recently observed that a number of mutational
events occurred in rFeLV proviruses following co
infection with FeLV-A, we wished also to learn whether
138
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
nucleotide substitutions would occur within in vivo-
derived rFeLV proviruses in the absence of FeLV-A helper
virus.
Initial sequence analysis of the rFeLV/D8 variants
employed in this study indicated their ability to evade
neutralization by C11D8 was at least partially due to
nucleotide substitutions in the vicinity of the epitope
domain. While these rFeLV/D8 variants were selected
based upon their resistance to neutralization with this
mAb, a more important issue was whether these variants
were resistant to neutralization by natura1ly-occurring
antibodies present in immune cat sera. This issue was
investigated using cat immune sera for in vitro
neutralization assays. Results demonstrated that while
there was high neutralizing activity against exogenous
FeLV-A and FeLV-C, the exogenous FeLV-B/GA isolate and
the recombinant rFeLV/D8 variants were poorly
neutralized.
One potential helper function for FeLV-A is
replication rescue of defective variants, we therefore
investigated the competency of the rFeLV/D8 variants for
in vitro growth and cell tropism. We found that
replication efficiency as measured by in vitro growth
emd cell tropism was not altered for the rFeLV/D8
variants, but was comparable to the parental recombinant
139
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
population (Sheets et al., 1992) and to the FeLV-B/GA
isolate.
We next conducted in vivo studies using this
rFeLV/D8 pool in the absence of FeLV-A in a short
infection timecourse of 9 weeks pi (Litter I) and a
longer timecourse of 45 weeks pi (Litter II). Our
results indicated that in spite of no obvious viremia
and with minimal seroconversion observed, immunostaining
and PCR analysis showed the presence of rFeLV proviruses
in the tissues of rFeLV/D8-infected cats. A previous
report described a cat experimentally infected with an
FeLV subgroup A, which was both aviremic and anti-FeLV
negative but showed focal FeLV antigen expression in
lymphoid tissues (Hayes et al., 1989). Such cats may be
examples of in vivo recombination-generated viruses with
altered neutralizing properties occurring naturally. In
the present study, rFeLV/D8s appeared to be able to
infect cells of the splenic marginal zone (Fig. 14A) as
well as cells in the salivary gland (Fig. 14C), while
FeLV/Rickard primarily infected cells in germinal
centers (Fig. 143), suggesting a possible altered
cellular tropism. The lack of detectable antigenemia in
the rFeLV/D8-infected cats may be due to low virus
expression or selection of a virus phenotype that is
highly cell-associated. Inefficient virus replication,
140
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
as well as the cell-associated phenotype, may account
for the lack of detectable antibody to FeLV caps id
antigen by ELISA. Our lab has previously described
another example of altered in vivo cell tropism in which
cats were infected with either FeLV-C alone or a mixture
of FeLV-C recombinants along with parental FeLV-C
(Chakrabarti et al., 1994; Roy-Burman, 1996). Brain
tissues of infected cats that died of aplastic anemia
were examined by immunostaining. Positive staining of
the central nervous system (CNS) capillary endothelial
cells was observed in cats inoculated with the virus
mixture, while in contrast, brain tissue derived from
cats infected with FeLV-C alone had no such staining.
In vitro, human endothelial cells derived from retina
(REC) or brain (EEC) could be infected by the virus
mixture, while these cells were resistant to infection
by FeLV-C alone. Subsequent analysis showed a
recombinant in which two-thirds of the SU was derived
from the enFeLV sequence, indicating the CNS endothelial
cell tropism was probably due to presence of the
recombinant SU. Taken together, these results suggest
that recombinants, whether FeLV-C-derived or FeLV-B-
like, are capable of demonstrating altered cell tropism.
Since there is ample evidence that cell receptor choice
is primarily dictated by the N-terminus of the virion SU
141
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
(Roy-Burmany 1996) it is likely the altered cell tropism
of these recombinants is mediated by the enFeLV-derived
portion of the SU.
In summary, data presented in this study indicate
that replication-competent rFeLV variants are generally
able to escape immune surveillance in vivo and establish
a very limited infection in the edasence of FeLV-A, while
demonstrating some altered cell tropism. It is possible
an altered-expression phenotype may also contribute to
the survival of the recombinant virus. However, in the
cibsence of efficient infection and viral replication,
such as occurs during concurrent in vivo FeLV-A
infection, those nucleotide substitutions resulting in
conversion of rFeLV env sequences to those of FeLV-B-
type sequences is not observed.
142
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
EPILOGUE
SUMMARY AND CONCLUSIONS
Overview
The primary objective guiding the study described
in this dissertation was to gain a better understanding
of the temporal process for in vivo evolution of rFeLV
species. The experimental strategy was based upon the
hypothesis that specific recombinant FeLVs (rFeLVs)
evolve in vivo such that certain recombinants gain
selective advantage, resulting in their outgrowth from a
mixture of recombinants. A logical extension of this
hypothesis was that such predominating recombinants may
represent species with greater leukemogenic potential.
The specific approaches to test this hypothesis were:
(i) in vivo studies with a defined mixture of rFeLVs
either with or without FeLV-A helper virus, (ii)
comparison of nucleotide sequences for parental rFeLVs
to those of in vivo-propagated rFeLVs, and (iii)
ultimately comparison of sequences for those in vivo-
derived rFeLVs found in the above study to sequences for
rFeLVs isolated from other experimentally-induced and
naturally-occurring thymic tumor DNA, in order to
determine whether a common pattern of selection existed
for certain rFeLV species. The results presented in
143
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
this work support the hypothesis that certain rFeLV
species do evolve and predominate in vivo and may, by
extension, represent species with greater leukemogenic
potential.
In Vivo Selection of Specific rFeLV Species
The use of a mixture of recombinants with defined
sequences for the in vivo study presented here allowed
us to analyze "if and when" over the timecourse of
infection a specific subpopulation of rFeLVs would
emerge as a dominant species during an in vivo selection
process. Previous characterization of the parental
rFeLV mixture used in this study showed that the in
vitro-generated recombinants contained 3' crossover
junctions within a limited region in the middle of the
env gene, spanning at most 270 nucleotides (Sheets et
al., 1992). By combining previous results with those
from the current work, it is clear that while a variety
of crossover sites existed in the initial rFeLV mixture
(sites Â to G), only a few of these crossover sites were
observed in rFeLVs recovered from the three
experimentally-induced tumor-bearing cats, namely
recombinants with site D to site G crossover junctions.
Since the samples used in the first phase of this study
were obtained from necropsy tissues of these three cats,
144
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
the question of when in the course of infection the in
vivo selection for certain recombinant species occurred
could not be addressed. During the study's second
phase y subsequent analysis of serum-derived RT-PCR
products from one tumor-bearing cat over its 44-week
infection timecourse, allowed us to assess the pattern
of in vivo selection, at least in this particular cat.
Results indicated there was initial selection against
those recombinant species with upstream crossover
junctions at sites A to C (species with lesser amounts
of endogenously-derived env sequence) followed by a
gradual selection for the site G recombinant species
(which contains the greater amount of endogenously-
derived sequence).
The third phase of this work compared the above
results with the profile of recombinant species observed
in other tumor-bearing cats to determine whether a
similar pattern might emerge. Two cats with
experimentally-induced thymic tumor following infection
with FeLV-A alone were shown to harbor rFeLVs. Analysis
indicated these recombinant viruses presumably arose de
novo. DNA from one naturally-occurring thymic tumor was
included in the third phase. Results suggest there is
indeed a pattern in which recombinant species with
relatively greater amounts of endogenously-derived
145
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
sequence do predominate. Clones from the four tumor
DNAs indicate 75% (9 of 12) contain recombination
junctions at sites F to G. A comparison with the 3'
junction sites for the known FeLV-B isolates shows that
2 of 3 isolates have site E junctions and one isolate
has a 3' recombination junction >G. While it is unclear
why recombinant species with 3' junctions at sites E to
G may be preferred in vivo, examination of published
amino acid sequences (Kumar et al., 1989) indicate that
recombination sites E to G reside just downstream of
variable region V (vrV) in the middle of the SU.
Perhaps retention of this vrV provides some selective
advantage, however, since the three-dimensional
structure of the SU is currently unknown, this
observation is clearly open to speculation.
In Vivo Nucleotide Sequence Conversion
As described above, the use of a defined mixture of
in vitro-generated rFeLVs not only allowed us to monitor
the in vivo selection of certain recombinant species
within a larger population, but it also allowed us to
observe the in vivo evolution of variants harboring
mutations within the SU portion of the env gene, and to
determine whether FeLV—A helper virus may have an effect
upon evolution of substitutions within the rFeLV
146
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
population. The issue of potential FeLV-A helper
function addressed the fact that all natural FeLV-B
infections are found in the presence of FeLV-A.
Utilizing rFeLV clones isolated from serum-derived RT-
PCR products obtained from one co-infected cat (4746-5)
at various timepoints over its 44-week infection
timecourse, it was initially observed that virtually all
85 clones contained a set of four point mutations in the
mid-region of the env, an area encompassing variable
region V (vrV) and part of a conserved region just 5' of
vrV. Results indicated this set of point mutations was
strongly selected since all four mutations were present
at the first timepoint at which rFeLVs were detected in
the sera ( 8 weeks pi ). Three of the four point
mutations resulted in amino acid changes.
Interestingly, it was observed that this set of
mutations resulted in complete seguence identity to
three molecularly cloned FeLV-B isolates. Intrigued by
this, the analysis was expanded to include the entire SU
portion encoded by the env gene. In the third phase of
this study, multiple clones were examined from three
experimentally-induced thymic LSAs, and results compared
to findings from a naturally-arising thymic tumor. It
was startling to find a total of 19 point mutations
consistently observed in all four cats. This set of 19
147
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
mutations was scattered throughout the entire SU domain,
in both variable and conserved regions. Amazingly, the
presence of these 19 mutations resulted in complete
seguence identity to the cloned FeLV-B isolates.
Conclusions
Retroviruses are unique in their ability to
generate considerable genetic diversity, a subject which
has been extensively reviewed (Coffin, 1992; Katz and
Skalka, 1990). However, this genetic diversity is
balanced by selection processes which promote
proliferation of certain variants containing
advantageous mutations. As a result, these selective
forces actually drive the evolution of retroviral
variants. In this regard, the work described here
demonstrates how strongly in vivo selection processes
were exerted upon a mixture of rFeLVs very early during
infection so that variants harboring a specific set of
point mutations in the env gene quickly emerged. This
strong early selection process was followed by
additional selection for a subset of these variants
containing relatively greater amounts of endogenously-
derived env sequence, allowing this subset to gradually
predominate, emd by extension, potentially implicating
such recombinant species as proximal leukemogens.
148
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Future Directions
In this study, 12 of 19 point mutations observed in
the env gene were strictly conserved in all clones
derived from the four cats analyzed, and as mentioned
earlier, these 12 mutations were also conserved in the
cloned FeLV-B isolates. In future studies it would be
useful to generate some variant rFeLVs by site-directed
mutagenesis to determine the various roles for these
individual changes in FeLV-B replication efficiency.
For example, at position l in Fig. 10, an A-K3 transition
mutation resulted in an Isol91Val amino acid
substitution. Interestingly, murine MCF viruses and all
FeLV isolates, regardless of subgroup, express Vall91 at
this position (Elder and Mullins, 1983), while the two
cloned enFeLVs express Ilel91 (Kumar et al., 1989).
Moreover, the Asn adjacent to Vall91 has been identified
as a glycosylation site in Friend-MCF virus (F-MCFV)
(Geyer et al., 1990). Thus, while Val is a conservative
substitution for lie, perhaps glycosylation may be more
efficient in the presence of a Val residue rather than
an lie residue. Unfortunately, since the tertiary
structure of the SU glycoprotein for FeLVs is currently
unknown, we are left to considerable speculation as to
which of the 19 positions we may wish to initially
study. In that regeurd, a number of reports have
149
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
provided amino acid sequence comparisons for FeLVs and
MCF viruses describing considerable sequence
conservation between these distantly related
retroviruses (Elder and Mullins, 1983; Mullins and
Hoover, 1990; Kayman et al., 1991; Battini et al.,
1992). Additionally, a recent report identified the
cysteine residues involved in disulfide bond formation
for Friend-MCF virus and presented a two-dimensional
model for the N-terminal portion of the SU for this
polytropic virus (Linder et al.^ 1994). Using this
model as the basis for a first approximation, a cartoon
of the possible two-dimensional structure for the N-
terminal SU of polytropic FeLV-B/GA strain is presented
in Fig. 17. A visual inspection of this cartoon
suggests that perhaps the disulfide-bonded structural
element A, which includes a portion of variable region
II (vril), may be an interesting target for mutagenesis,
particularly in light of the fact that host cell
receptor determinants for FeLVs are located within the
N-terminal 25% of the SU domain (reviewed in Roy-Burman,
1996) . Other regions which represent potentially
interesting targets for mutagenesis include the
disulfide-bonded structural element B, which also
contains vrIII, and structural element C, which contains
a portion of vrIV.
150
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Fig. 17 Proposed Model of the N—terminal Half of the
FeLV-B/GA Mature SU Glycoprotein. This representation
is based upon the MCF model as published by Linder et
al., 1994. Glycosylation sites at amino acid positions
10 and 25 are as reported by Geyer and Geyer, 1991.
Amino acid residues in gray are conserved between FeLV-
B/GA and Friend-MCF sequence while residues with white
background are not conserved between F-MCF and FeLV-
B/GA. Residues highlighted by a black background are
within FeLV variable regions as described by Kumar et
al., 1989. The viral neutralizing epitope, MGPNL, at
position 213 to 217, is highlighted with white letters
in a black background for reference. The disulfide-
bonded structural elements A, B, and C are designated by
dashed boxes.
151
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
1
o O
Ii
« u .
II
d)
152
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
While the in vivo evolution of the set of point
mutations described in this work is very interesting,
and most likely essential for efficient replication of
exogenous FeLV-B isolates, of greater potential interest
with regard to leukemogenesis is the gradual
predominance of certain rFeLV species over the
timecourse of in vivo infection. We observed a
preference for recombinant species harboring relatively
greater amounts of enFeLV-derived sequence such that
rFeLVs containing 3' crossover junctions at sites E
through G appeared to be preferred; this was noted in
the exogenous FeLV-B isolates as well as the in vivo-
derived clones analyzed in this study. For this reason
I think it would be useful to focus on vrV, which
resides just 5' of recombination site E. Some questions
to consider are: (i) would an ecotropic FeLV-A variant
containing vrV derived from FeLV-B demonstrate more
leukemogenic potential? (ii) does vrV serve to mask the
previously characterized immunodominant MGPNL
determinant? (iii) would the FeLV-A variant harboring
FeLV-B-derived vrV evade in vitro neutralization as
described in Chapter 5; and would such an FeLV-A variant
manifest an altered in vivo phenotype? It is possible
that rFeLVs containing more enFeLV sequence may not
potentiate leukemogenesis, but merely have an in vivo
153
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
proliferative advantage independent of disease
modulation. However, I believe this is an area which
deserves further study and hope to read about such
studies performed by our lab in the future.
154
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
REFERENCES
ATHAS, 6.B., B. CHOI, S. PRABHIJ, P.A. LOBEUÆ-RICH, and
I*.S. LEVY. (1995). Genetic determinants of feline
leukemia virus—induced multicentric lymphomas. Virology
214, 431-438.
BATTINI, J-L., J.N. HEARD, AND O. DANOS. (1992).
Receptor choice determinants in the envelope
glycoproteins of amphotropic, xenotropic, and polytropic
murine leukemia viruses. J. Virol. 66, 1468-1475.
BECHTEL, M.K., M.R. STALLCOP, R.M. BEDGOOD, J.L. COREY,
R. PANDEY, and P. ROY-BURNAN. (1994). Abnormal
processing of a recombinant feline leukemia virus
envelope polyprotein and its interference with subgroup
C virus infection. Virology 202, 329-338.
BERRY, B.T., A.K. GHOSH, D.V. KOMAR, D.A. SPODICK, and
P. ROY—BORMAN. (1988). Structure and function of
endogenous feline leukemia virus long terminal repeats
and adjoining regions. J. Virol. 62, 3631-3641.
BISHOP, J.M. and H. VARMOS. Functions and origins of
retroviral transforming genes. In Weiss, RA, Teich, NM,
Varmus, HE, Coffin, JM (eds.) SNA tumor viruses. (Second
Edn. ) Cold Springs Harbor Laboratory Press: Cold Springs
Harbor, NY, 1984, pp.999-1108.
BOSCH, M.P., B.G. DEVI, L.H. SOE, B. PERBAL, M.A. BALODA
and P. ROY-BORMAN. (1983). Characterization of the
expression of cellular retrovirus genes and oncogenes in
feline cells. Hematol. Oncol. 1, 61-75.
CHAKRABARTI, R., P.M. HOFMAN, R. PANDEY, L.E. MATHES,
and P. ROY-BURMAN. (1994). Recombination between feline
exogenous and endogenous retroviral sequences generates
tropism for cerebral endothelial cells. Am. J. Pathol.
144, 348-358.
COFFIN, J.M. (1992). Genetic Diversity and Evolution of
retroviruses. Current Topics Microbiol. Immunol. 176,
143-164.
DOMINGO, E., E. MARTINEZ-SALAS, F. SOBRINO, J.C. de la
TORRE, A. PORTELA et al. (1985). The quasispecies nature
of viral RNA genome populations: a review. Gene 40, 1-8.
155
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
DOMINGO. E., M.G. MATED, C. ESCARMIS, E. MARTINEZ-SALAS,
D. ANDREU, N. VERDAGUER, and I. FITA. (1996). Molecular
evolution of aphthoviruses. Virus Genes 11, 197-207.
DONAHUE, P.R., E.A. HOOVER, G.A. BELTZ, N. RIEDEL, V.M.
HIRSCH, M., J. OVERBAUŒ, and J.I. MDLLINS. (1988).
Strong sequence conservation among horizontally
transmissible, minimally pathogenic feline leukemia
viruses. J. Virol, 62, 722-731.
ELDER, J.H., J.S. MC GEE, M. MDNSON, R.A. HOUGHTEN, W.
KLOETZER, J.L. BXTTLE, and C.K. GRANT. (1987).
Localization of neutralizing regions of the envelope
gene of feline leukemia virus by using anti-synthetic
peptide antibodies. J. Virol. 61, 8-15.
ELDER, J.H., and J.I. MDLLINS. (1983). Nucleotide
sequence of the envelope gene of Gardner-Amstein feline
leukemia virus B reveals unique sequence homologies with
a murine mink cell focus-forming virus. J. Virol. 46,
871-880.
ELDER, J.H. and J.I. MDLLINS. Nucleotide sequence of
the envelope genes and LTR of the subgroup B, Rickard
strain of feline leukemia virus, p.1005-1010. In Weiss
R., Teich, N., Varmus, H., Coffin, J. (eds.), RNA tumor
viruses. 1985. Cold Springs Harbor Laboratory, New York.
ESSEX, M.M. , 6. KLEIN, S.P. SNYDER, and J.B. HARROLD.
(1971). Feline sarcoma virus (FeSV)—induced tumors:
correlation between humoral antibody and tumor
regression. Nature (London) 233, 195-197.
FAN, H. Retroviruses and their role in cancer. In. Levy
JÀ (ed.) The retroviridae. Plenum Press. NY, 1994, pp
313-362.
FISCHINGER, P.J., C.S. ELEVENS, and S. NCNIDRA. (1974).
Simple, quantitative assay for both xenotropic murine
leukemia and endotropic feline leukemia viruses. J.
Virol. 14, 177-179.
GEYER, R., J. DABROWSKI, D. LINDER, H.H. SCHOTT, AND S.
STIRM. (1990). Oligosaccharides at individual
glycosylation sites in glycoprotein 71 of friend murine
leukemia virus. Eur. J. Biochem. 187, 95—110.
156
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
GEYER, H. and R. GEYER. (1991). Isolation of the
glycopeptides of the envelope glycoprotein from a
polytropic murine retrovirus. GBF Monogr. 15, 39-48.
GRANT, C.K., B.J. ERNISSE, O. JARRETT, and F.R. JONES.
(1983). Feline leukemia virus envelope gp70 of
subgroups B and C defined by monoclonal antibodies with
cytotoxic and neutralizing functions. J. Inamnol, 131,
3042-3048.
HARDY, W.D. Jr. Biology of feline retroviruses. In R.C.
Gallo and F.Wong-Staal (ed.). Retrovirus biology and
human diseases (1990). Marcel Dekker, NY, p.33-85.
HARDY, W.D., Jr. (1993). Feline oncoretroviruses, p.l09-
180. In "The Retroviridae". J.A. Levy (Ed.) Vol 2,
Plenum Press, New York.
HARDY, W.D. Jr, P.W. HESS, E.G. MacEWEN, A.J.____
MCCLELLAND, E.E. ZUCKERMAM, M. ESSEX, S.H. COTTER, and
O. JARRETT. (1976). Biology of feline leukemia virus in
the natural environment. Cancer Res. 36, 582-588.
HOOVER, E.A. and J.I. MULLINS. (1991). Feline leukemia
virus infection and diseases. J. Am. Vet. Med. Assoc.
199, 1287-97.
HOOVER, E.A., J.L. ROJKO, and R.6. OLSEN. (1980).
Factors influencing host resistance to feline leukemia
virus. In, “Feline Leukemia" (R. G. Olsen, Ed.), pp.
69-76, CRC Press, Boca Raton, Florida.
HUEBNER, R.j. and G.J. TODARO. (1969). Oncogenes of RNA
tumor viruses as determinants of cancer. PNAS. 64, 1087-
1094.
JARRETT, O. , W.D. HARDY, JR., M.C. GOLDER, and D. HAY.
(1978). The frequency of occurrence of feline leukemia
virus subgroups in cats. Int. J. Cancer 21, 334-337.
JARRETT, O., H.M. LAIRD, AND D. HAY. (1969). Growth of
feline leukemia virus in human cells. Nature. 224, 1208-
1209.
JARRETT, o ., and P.H. RUSSELL. (1978). Differential
growth and transmission of feline leukaemia viruses
subgroups A and B. Int. J. Cancer 21, 466-472.
157
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
JARRETT, W.F.H. , E.M. CRAWFORD, W.B. MARTIN, and F.A.
DAVXE. (1964). A virus-like particle associated with
leukemia (lymphosarcoma). Natiire. 202, 567-569.
KATZ, R.A and A.M. SKALKA. (1990). Generation of
diversity in retroviruses. Anna. Rev. Genet. 24, 409-45.
KAWAKAMI, T.G., G.H. THEILEN, D.L. DDNGWORTH, R.J. MDNN,
and S-G. BEALL. (1967). "C'-type viral particles in
plasma of cats with feline leukemia. Science. 158, 1049-
1050.
KAYMAN, S.C., R. KOPEUIAN, S. PROJAN, D.M. KINNEY, AND
A. PINTER. (1991). Mutational analysis of N-linked
glycosylation sites of friend murine leukemia virus
envelope protein. J. Virol. 65, 5323-5332.
KOMAR, D.V., B.T. BERRY, and P. ROY-BURMAN. (1989).
Nucleotide sequence and distinctive characteristics of
the env gene of endogenous feline leukemia provirus. J.
Virol. 63, 2379-2384.
LAVIGNON, M. and L. EVANS. (1996). A multistep process
of leukemogenesis in moloney murine leukemia virus-
infected mice that is modulated by retroviral
pseudotyping and interference. J. Virol. 70, 3852-3862.
LEVY, L.S., P.A. LOBELLE-RICH, J. OVERBAUGH, J.L.
ABKOWITZ, R. FULTON, and P. ROY-BURMAN. (1993).
Coincident involvement of tlvi—2, c-myc, and novel env
genes in natural and experimental lymphosarcomas induced
by feline leukemia virus. Virology 196, 892-895.
LINDER, M., V. WENZEL, D. LINDER, and S. STIRM. (1994).
Structural elements in glycoprotein 70 from polycopie
friend mink cell focus-inducing virus and glycoprotein
71 from ecotropic friend murine leukemia virus. J.
Virol. 68, 5133-5141.
MATHES, L.E., R. PANDEY, R. CHAKRABARTI, F.M. HOFMAN,
K.A. HAYES, P. STROMBERG, and P. ROY-BURMAN. (1994).
Pathogenicity of a subgroup C feline leukemia virus
(FeLV) is augmented when administered in association
with certain FeLV recombinants. Virology 198, 185-195.
MC DOUGALL, A., A. TERRY, T. TZAVARAS, J. ROJKO, C.
CHENEY, and J.C. NEIL. (1994). Defective endogenous
viruses are expressed in feline lymphoid cells: evidence
158
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
for a role in natural resistance to subgroup B feline
leukemia viruses. J. Virol, 68, 2151-2160.
MDLLINS, J.I. and E.A. HOOVER. (1990). Molecular aspects
of feline leukemia virus pathogenesis, p.87-116. In
Gallo, R. and Wong-Staal, F. (eds.). Retrovirus biology
and human disease. Marcel Dekker, New York.
NEIL, J.C. , R. FDLTON, M. RIGBY, and M. STEWART.
(1991). Feline leukemia virus: generation of pathogenic
and oncogenic variants. Current Topics in Microbiol.
Immunol. 171, 68-92
NICOLAISEN-STROUSS, K. , H.P.M. KOMAR, T. FITTING, C.K.
GRANT, and J.H. ELDER. (1987). Natural feline leukemia
variant escapes neutralization by a monoclonal antibody
via an amino acid change outside the antibody-binding
epitope. J. Virol. 61, 3410-3415.
NIMAN, H.L., J.R. STEPHENSON, M B. GARDNER, and P. ROY-
BORMAN. (1977). RD-114 and feline leukemia virus genome
expression in natural lymphomas of domestic cats. Nature
(London). 266, 357-360.
NONBERG, J.H., M E. WILLIAMS, and M.A. INNIS. (1984).
Nucleotide sequences of the envelope genes of two
isolates of feline leukemia virus subgroup B. J. of
Virol. 49, 629-632.
OVERBAUGH, J., E.A. HOOVER, J.I. MDLLINS, D.P.W. BORNS,
L. RODENSEY, S.L. QUACKENBOSH, V. STALLARD, and P.R.
DONAHOE. (1992). Structure and pathogenicity of
individual variants within an immuno-deficiency disease-
inducing isolate of FeLV. Virology 188, 558-569.
OVERBAUGH, J., N. REIDEL, E.A. HOOVER, and J.I. MDLLINS.
(1988). Transduction of endogenous envelope genes by
feline leukemia virus in vitro. Nature (London) 332,
731-734.
PANDEY, R. , A.K. ŒOSH, D.V. KOMAR, B.A. BACHMAN, D.
SHIBATA, and P. ROY-BORMAN. (1991). Recombination
between feline leukemia virus sub^oup B or C and
endogenous env elements alters the in vitro biological
activities of the viruses. J. Virol. 65, 6495-6508.
PANDEY, R., M.K. BECHTEL, Y. SO, A.K. GHOSH, K.A. HAYES,
L. E. MATHES, and P. ROY-BORMAN. (1995). Feline leukemia
159
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
virus variants in experimentally induced thymic
lymphosarcomas. Virology 214, 584-592
QUACKENBUSH, S.L., P.R. DONAHUE, G.A. DEAN, N.H. MYLES,
C.D. ACKLEY, M.D. COOPER, J.I. MDLLINS, and E.A. HOOVER.
(1990). Lymphocyte subset alterations and viral
determinants by the feline leukemia virus FeLV-FAIDS.
J. Virol. 64, 5465-5474.
REIDEL, N., E.A. HOOVER, P.H. GASPER, M.O. NICOLSON, and
J.I. MULLINS. (1986). Molecular analysis and
pathogenesis of the feline aplastic anemia retrovirus,
FeLV-C-Sarma. J. Virol. 60, 242-260.
ROHM, J.L., M.L. LINENBER6ER, E.A. HOOVER, and J.
OVERBAUGH. (1994). Evolution of feline leukemia virus
variant genomes with insertions, deletions, and
defective envelope genes in infected cats with tumors.
J. Virol. 68, 2458-2467.
ROJKO, J.L. , E.A. HOOVER, S. KRAKOWKA, R.G. OLSEN, and
L.E. MATHES. (1979). Influence of adrenal
corticosteroids on the susceptibility of the cats to
feline leukemia virus. Cancer Res. 39, 3789-3791.
ROUS, P. (1911). A sarcoma of the fowl transmissible by
an agent separable from the tumor cells. J. Exp. Med.
13, 397-411.
ROY-BURMAN, P. (1996). Endogenous env Elements: Partners
in Generation of Pathogenic Feline Leukemia viruses.
Virus Genes 11, 147-161.
ROY-BURMAN, P., M.P. BUSCH, S. RASHEED, M.B. GARDNER,
and M.M.C. LAI. (1980). Oncodevelopmental gene
expression in feline leukemia, p.361-372. In W.D. Hardy,
Jr., M. Essex, and A.J. McClelland (ed.). Feline
leukemia virus. Elsevier/North-Holland, New York.
RUSSELL, P.H. AND O. JARRETT. (1978). The occurrence of
feline leukemia virus neutralizing antibodies in cats.
Int. J. Cancer 22, 351-357.
SARNA, P.S., and T. LOG. (1971). Viral interference in
feline leukemia-sarcoma complex. Virology 44, 352-358.
SARNA, P.S., and T. LOG. (1973). Subgroup classification
of feline leukemia and sarcoma viruses by viral
160
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
interference and neutralization tests. Virology 54,
160-169.
SARMA, P.S., T. LOG, S. SKDNTZ, S. KRISHNAN, and K.
BURKLEY. (1978). Experimental horizontal transmission of
feline leukemia viruses of subgroups A, B, and C. J,
Natl. Cancer Inst. 60, 871-874.
SHEETS, R.L. , R. PANDEY, N.-C. JEN, and P. ROY-BDRMAN.
(1993). Recombinant feline leukemia virus genes
detected in naturally occurring feline lymphosarcomas.
J. Virol. 67, 3118-3125.
SHEETS, R.L., R. PANDEY, V. ELEMENT, C.K. GRANT, and P.
ROY-BDRMAN. (1992) Biologically selected recombinants
between feline leukemia virus (FeLV) subgroup A and an
endogenous FeLV element. Virology 190, 849-855.
SOE, L.H., B.G. DEVI, J.I. MDLLINS, euld P. ROY-BDRMAN.
(1983). Molecular cloning and characterization of
endogenous feline leukemia virus sequences from a cat
genomic library. J. Virol. 46, 829-840.
SOE, L.H., R.N. SHIMIZn, J.R. LANDOLPH, and P. ROY-
BDRMAN. (1985). Molecular analysis of several classes of
endogenous feline leukemia virus elements. J. Virol.
46, 829-840.
STEWART, M.A., M. WARNOCK, A. WHEELER, N. WILKIE, J.I.
MDLLINS, D.E. ONIONS, and J.C. NEIL. (1986). Nucleotide
sequences of a feline leukemia virus subgroup A envelope
gene and long terminal repeat and evidence for the
recombinational origin of subgroup B viruses. J. Virol.
58, 825-834.
TSATSANIS, C., R. FDLTON, K. NISHIGAKI, H. TSUJIMOTO, L.
LEVY, A. TERRY, D. SPANDIDOS, D. ONIONS, and J.C. l^IL.
(1994). Genetic determinants of feline leukemia virus-
induced lymphoid tumors: patterns of proviral insertion
and gene rearrangement. J. Virol. 68, 8296-8303.
TZAVARAS, T., M. STEWART, A. MC DOUGALL, R. FDLTON, N.
TESTA, D.E. ONIONS, and J.C. NEIL. (1990). Molecular
cloning and characterization of a defective recombinant
feline leukemia virus associated with myeloid leukemia.
J. Gen. Virol. 71, 343-354.
161
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
IMAGE EVALUATION
TEST TARGET (Q A -3 )
'<■
" %
Y.
y
%
%
%
1.0
l.l
J2-2
Là
■M
1 . 8
1 .2 5 1 . 4 1 . 6
1 5 0 m m
V
V
7
%
/APPLIED A IIVMGE . Inc
j a g 1653 East Main Street
Rochester. NY 14609 USA
Phone: 716/4820300
— Fax: 716/288-5989
0 1 99 3, AppOed Im a ge , In c.. AU R ig h ts R esenrad
‘ 4fS
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
INFORMATION TO USERS
This manuscript has been reproduced from the microfilm master. UMI
films the text directly from the original or copy submitted. Thus, some
thesis and dissertation copies are in ^ew riter free, while others may be
from any ^ e of computer printer.
The quality of this reproduction is dependent upon the quality of the
copy submitted. Broken or indistinct print, colored or poor quality
illustrations and photographs, print bleedthrough, substandard margins,
and improper alignment can adversely afreet reproduction.
In the unlikely event that the author did not send UMI a complete
manuscript and there are missing pages, these will be noted. Also, if
unauthorized copyright material had to be removed, a note will indicate
the deletion.
Oversize materials (e.g., maps, drawings, charts) are reproduced by
sectioning the original, beginning at the upper left-hand comer and
continuing from left to right in equal sections vrith small overlaps. Each
original is also photographed in one exposure and is included in reduced
form at the back of the book.
Photographs included in the original manuscript have been reproduced
xerogr%q)hically in this copy. Higher quality 6” x 9” black and white
photographic prints are available for any photographs or illustrations
appearing in this copy for an additional charge. Contact UMI directly to
order.
UMI
A Bell & Howell Infoimatioii Conqiany
300 North Zed) Road, Ann Arbor MI 48106-1346 USA
313/761-4700 800/521-0600
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
UMI Number: 9816084
UMI Microform 9816084
Copyright 1998, by UMI Company. All rights reserved.
This microform edition is protected against unauthorized
copying under Title 17, United States Code.
UMI
300 North Zeeb Road
Ann Arbor, MI 48103
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Moloney murine leukemia virus (MoMuLV) envelope protein: Interaction with receptor and engineering for targeted gene transfer

PDF

Pattern and mechanism of troponin C-Fast gene expression during tongue myogenesis

PDF

The murine leukemia virus envelope protein: Engineering for vector targeting and analysis of interactions with viral core proteins

PDF

Identification of the first symbiotic determinants of Vibrio fischeri, the light organ symbiont of Euprymna scolopes

PDF

The role of host proteins in retroviral cDNA integration

PDF

Experimental and computational analysis of microsatellite repeat instability in mismatch-repair-deficient mice

PDF

Isolation and characterization of mouse GRIP1, a novel transcriptional coactivator of steroid receptors

PDF

The role of DNA methylation and tumor suppressor genes in the pathogenesis of prostate cancer and leukemias

PDF

Identification and characterization of cis-elements that regulate tropomodulin gene expression during muscle differentiation

PDF

The role of the human steroid 5-alpha-reductase type II gene in the molecular genetic predisposition to prostate cancer

PDF

Receptor interaction determinants in nonecotropic MuLV envelope proteins

PDF

Studies of Drosophila sex-lethal protein interactions and cloning of a novel interactor, SIN

PDF

Demethylation and reactivation of growth regulatory genes by 5-Aza-2'-deoxycytidine in human neoplasia

PDF

In vivo and in vitro analysis of the epigenetic and genetic alterations in the progression of human bladder cancer

PDF

Expression and characterization of recombinant collagen targeting TGF-beta and application in osteogenesis

PDF

Transcriptional regulation in osteoblasts

PDF

Functional analyses of SpMsx and SpOct in sea urchin embryogenesis

PDF

Regulation of Ras-related guanosine triphosphatases by guanine nucleotide exchange factors

PDF

Studies on the relationship between the gibbon ape leukemia virus (GaLV) Env glycoprotein and HIV-1-based lentiviral vectors: Implications in enveloped vector assembly

PDF

Kinetic studies of the interaction between U1A protein and U1 hairpin II RNA

Asset Metadata

Creator Bechtel, Marta Keith (author)

Core Title In vivo evolution of recombinant feline leukemia viruses (FeLV) toward FeLV subgroup B viruses in thymic lymphomagenesis

School Graduate School

Degree Doctor of Philosophy

Degree Program Biochemistry and Molecular Biology

Degree Conferral Date 1997-08

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag agriculture, animal pathology,biology, genetics,biology, microbiology,biology, molecular,OAI-PMH Harvest

Language English

Contributor Digitized by ProQuest (provenance)

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c17-299616

Unique identifier UC11353069

Identifier 9816084.pdf (filename),usctheses-c17-299616 (legacy record id)

Legacy Identifier 9816084.pdf

Dmrecord 299616

Document Type Dissertation

Rights Bechtel, Marta Keith

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the au...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus, Los Angeles, California 90089, USA