University of Southern California Dissertations and Theses

Single cell gene expression analysis of KSHV infection in three-dimensional oral epithelial tissue cultures

(USC Thesis Other)

Single cell gene expression analysis of KSHV infection in three-dimensional oral epithelial tissue cultures

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content 1

Single Cell Gene Expression Analysis of KSHV Infection in Three-Dimensional Oral Epithelial
Tissue Cultures
Kyle L. Jung
Molecular Microbiology and Immunology
Master of Science
University of Southern California
December 2019
2

Abstract

Kaposi’s sarcoma-associated herpesvirus (KSHV) is the etiological agent of Kaposi’s
sarcoma (KS) and the development of oral KS is often the first indication of HIV infection in
patients. The oral cavity has been previously identified as the major site for shedding of
infectious KSHV virions and is the most likely site for KSHV transmission. To study KSHV
infection of the oral epithelium, we developed a three-dimensional (3D) oral tissue infection
model that mimics in vivo oral KSHV infection. Exposure of the basal layers in the 3D oral
tissues proved to be necessary for apical infection of KSHV. EpiOral tissues were harvested at
three, six, and thirteen days post KSHV infection for fluorescence microscopy, real-time PCR,
and single-cell RNA sequencing (scRNA seq). ScRNA seq showed that KSHV lytic cells were
present only on the more differentiated layers of epithelium, indicating that KSHV fails to
establish tight latency in the oral epithelia and spontaneously reactivates as the cells
differentiate, especially in the cornified oral epithelia. In addition, we are able to show
suppression of host mRNA expression by KSHV lytic reactivation, and a possible Rac1-
mediated autocrine and paracrine effect between KSHV lytic cells and non-infected cells. These
findings demonstrate the ability of our 3D oral KSHV infection model paired with single-cell RNA
sequencing to improve our understanding of oral KSHV infection and transmission.
3

Table of Contents
Introduction ...................................................................................................................... 4
Methods ........................................................................................................................... 6
Results ............................................................................................................................. 8
Discussion ...................................................................................................................... 14
References ..................................................................................................................... 16
Supplementary Figures .................................................................................................. 17

4

Introduction

Kaposi’s sarcoma (KS) is the most common cancer in HIV-infected untreated
individuals
7
and is often the first indication of HIV infection
3
. Kaposi sarcomas usually present as
multiple, pigmented, raised or flat, painless lesions on the epithelia and is caused by Kaposi’s
sarcoma-associated herpesvirus (KSHV)
2
. In this thesis, we will focus on the KSHV infection of
the oral cavity, which has been implicated in the transmission of KSHV. In HIV-positive patients
with KSHV co-infection, 20% of the patients will have the initial KS presentation in the mouth
and 70% of the patients will eventually develop oral KS
9
. While the exact mode of KSHV
transmission is not known, epidemiological data has shown an association between KSHV
transmission and men who have sex with men, with risk factors including deep kissing with an
infected partner and orogenetial contact
10
.
The risk factors for KSHV transmission match what is currently known about oral KSHV
infection. While KSHV establishes latency with very minimal lytic genes detected in skin KS,
KSHV infection of the oral epithelium leads to spontaneous reactivation and lytic viral
replication. This phenomenon was clearly demonstrated by Johnson et al. in 2005
5
, where they
infected a tonsil-derived oral keratinocyte monolayer with rKSHV.219, a recombinant virus that
expresses green fluorescent protein (GFP) during latency from the cellular EF-1a promoter and
red fluorescent protein (RFP) during lytic replication from the viral lytic PAN promoter. Then they
allowed the infected cells to differentiate into a multilayer. They show that differentiation of the
KSHV-infected epithelial cells occurred at the same time as activation of the KSHV lytic cycle,
indicating the importance of differentiation for KSHV replication. KSHV was unable to establish
latency in the oral tissues, forcing a spontaneous switch to lytic reactivation, resulting in
increased viral replication and shedding into the saliva. In clinical studies, real-time PCR was
used to detect the KSHV DNA in the saliva, where the titers were 100 to 1000-fold higher
compared to other anatomic sites
10
.
However, in vivo KSHV infections do not have complete exposure of the basal epithelial
layer when establishing infection. We wanted to better mimic in vivo infection by starting with a
fully-differentiated three dimensional tissue for infection. We used MatTek’s EpiOral 3D tissues,
Figure 1: (A) Diagram of MatTek EpiOral tissue. Figure is adapted from MatTek Corporation. The air-liquid interface exposes the
basal layers of the epithelia only to media while the suprabasal layers of the tissue are exposed to air. The differentiated layers consist
of basal, spinous, granular, and cornified layers (bottom to top). (B) Diagram of the 3D oral organoid tissue cultures setup for the
differentiation, infection and harvest stages.
A
B
5

which consist of normal human-derived primary oral keratinocytes that have been cultured to
form eight to eleven cell layers of fully-differentiated, non-cornified oral epithelia that develop a
cornified layer over time (Figure 1A). The tissues were cultured on specially-prepared inserts
and cultured in serum-free medium to attain levels of differentiation such as cytokeratin K13
expression in the apical layers and cytokeratin K14 expression in the basal layers of the tissue
6
.
We infected the fully-differentiated MatTek EpiOral tissue with rKSHV.219, using the GFP and
RFP expression as a guide for latent and lytic KSHV. The tissues are first lifted to air-liquid
interface four days before the infection to expose the suprabasal layers of the epithelium to air
while maintaining the basal level contact with the culture medium. The tissues are then infected
with rKSHV.219 for four hours, allowed to grow for three, six, and thirteen days post infection
(dpi), and harvested for fluorescence microscopy, real-time PCR, and single-cell RNA
sequencing (Figure 1B). The mock control for the study was treated with heat-inactivated
rKSHV.219 and harvested at three days post infection.
To define the transcriptional landscape during KSHV infection of our 3D oral epithelial
organoid tissue, we used the 10X Genomics Chromium system to profile both host and viral
gene transcriptome at a single-cell resolution. Bulk RNA sequencing is the standard method to
study gene expression and is good when studying a homogenous sample. In our KSHV EpiOral
infection model, there is a large variety of cell types including KSHV infected cells, KSHV non-
infected cells, and epithelial cells at varying stages of differentiation. In order to preserve the
heterogenous nature of our sample, the 10X system uses microfluidics to isolate individual cells
with a gel bead into micelles, also known as a gel bead in emulsion (GEM). Within each GEM,
the cells were lysed and mRNA molecules are bound to the gel bead reverse transcribed with
barcodes corresponding to each individual cell. The resulting cDNA is used to create standard
Illumina sequencing libraries and sequenced using an Illumina NovaSeq platform. Analysis was
done using a combination of 10X Genomics Cell Ranger and Seurat
1
analysis pipelines (Figure
2).

Figure 2. An overview of the workflow to generate scRNA-seq from KSHV-infected 3D oral epithelial organoid tissue. EpiOral cultures
are infected with KSHV.219 for 3,6, or 13 days and then harvested, enzymatically digested into a single cell suspension, and used for
Sc-RNA-seq. With the 10X Chromium platform, we are able to use about 5,000 cells per sample to generate a complete single-cell
sequencing library which can be used to analyze different transcriptome at a single-cell level. Figure is adapted from 10x Genomics
website.
6

Methods

KSHV Infection of Three-Dimensional Oral Epithelial Tissue

We used a commercial three-dimensional culture of primary oral epithelial cells from
MatTek Corporation, EpiOral, for this study. These cultures were generated from normal
human-derived primary oral keratinocytes. The primary keratinocytes were cultured by MatTek
to form multilayered, highly-differentiated tissues with a non-cornified, buccal phenotype that
could cornify during long-term culture. The tissues were cultured on specially-prepared collagen
membrane inserts, and cultured in serum-free medium with epidermal growth factors,
hydrocortisone, insulin, and other growth factors to induce buccal tissue differentiation. The
tissues were tested by MatTek to confirm differentiation within the tissues and a thickness of
eight to eleven layers of non-cornified tissue.
Four days before infection, the tissues were lifted to the air-liquid interface to expose the
top layer to air while the bottom is still in contact with the culture media. At the time of infection,
these tissues were lightly scored with a blunted insulin syringe needle, just enough to expose
the basal layer of cells, but not deep enough to puncture through the collagen membrane. 450
uL of rKSHV.219 virus (at 2x10^6 IU/mL in 293T cells) with 8 ug/mL of polybrene was added on
top of the tissue and incubated for four hours at 37C. The supernatant was then removed and
washed with PBS two times. Heat inactivated rKSHV.219 for the mock control was prepared by
incubating the rKSHV.219 virus at 100C for 15 minutes. The tissues were then incubated at 37C
for three, six, and thirteen days before harvesting for analysis. At these timepoints, the tissues
were harvested for immunofluorescence, real-time PCR, and single-cell RNA sequencing.

Immunofluorescence

For immunofluorescence analysis, we used Keyence BZ-X800 fluorescence microscope.
The images shown were taken from the top at 4X using the GFP, RFP, and brightfield imaging
with the same settings amongst the samples.

Harvesting of EpiOral Tissues

To harvest the tissues, the EpiOral tissues were washed with PBS and then separated
from the collagen membrane with forceps. The tissues were then placed into a 6-well plate with
50 uL of a 1:1 enzyme mix of collagenase/dispase (Millipore Sigma, Catalog #: 10269638001)
and trypsin/EDTA (Thermo-Fisher, Catalog #: 25200114). The tissues were further chopped up
into small pieces using a scalpel and then moved into a shaking incubator at 37C and 300rpm
for 30-45 minutes depending on the tissue size. The digested tissues were then filtered through
70um cell strainer (Falcon, Catalog #: 352350), and processed with the MACS Dead Cell
Removal Kit (Miltenyi Biotec, Catalog #: 130-090-101) to achieve a live cell percentage of 85%
or greater. From here, the samples were split for FACS and single cell RNAseq. The single cell
RNAseq sample was diluted to a concentration of 1,000 cells per uL in PBS with 0.04% BSA.

Real-Time PCR of EpiOral Tissues

For real-time PCR of the EpiOral tissues, RNA was harvested from the final single-cell
suspension using the Qiagen RNeasy mini kit (Qiagen, 74104). The resulting RNA was then
treated with DNase I (Sigma-Aldrich, Catalog #: AMPD1) to remove any contaminating DNA and
then used for reverse transcription with the Bio-Rad iScript cDNA synthesis kit (Bio-Rad,
Catalog #: 1708891). Real-time PCR was done using the Bio-Rad SsoAdvanced Universal
7

SYBR Green Supermix (Bio-Rad, Catalog #: 1725272). The primers that were used are listed
below:
LANA F: GAGTCTGGTGACGACTTGGAG R: AGGAAGGCCAGACTCTTCAAC
RTA F: TTGCCAAGTTTGTACAACTGCT R: ACCTTGCAAAGACCATTCAGAT
K8 F: GGTCTGTGAAACGGTCATTGA R: TCTATGTAGTCGCCTCTTGGA

10X Genomics Single Cell RNAseq Processing

For all scRNA seq processing, Rainin Pipet-Lite XLS pipettes and Universal fit tips were
used as recommended by the 10X Chromium Single Cell 3’ Reagent Kits v2 User Guide
Revision E. We also used this protocol for all steps in the processing.
After dead cell depletion with the MACS columns, the single cells were immediately
counted with a hemocytometer. The cells were diluted to 1,000 cells per uL and then aliquoted
for 10X processing. The cells were then mixed with RT reagent mix and nuclease-free water to
generate the final cell master mix. The cell master mix was then loaded onto the 10X Genomics
chromium chip in its specific well along with 40 uL of the thawed 10X gel beads and 135 uL of
partioning oil in their wells respectively. The chip was then run on the chromium controller with
the standard settings. An estimated 1,500 to 7,000 cells were recovered in the final gel beads
in emulsion (GEMs) and checked for opaqueness in the tip to confirm proper GEM formation.
GEMS were then incubated with the standard protocol in an Eppendorf MasterCycler
thermocycler.
Post GEM-RT cleanup and amplification of cDNA was done according to the protocol,
with the cDNA amplification using 8 amplification cycles. Amplified cDNA was sent for quality
control using the Agilent BioAnalyzer from the USC Molecular Genomics Core. Library
construction of the samples were done following the 10X protocol and after library construction,
the samples were again analyzed by the USC Molecular Genomics Core using the Agilent 2100
BioAnalyzer Platform for library QC.

Single-Cell Library Sequencing

For all of the single-cell libraries, we ran an initial low-depth sequencing run on the
Illumina MiSeq platform to calculate total number of cells in the library and determine the final
number of reads necessary. The large-scale sequencing run was done on the Illumina NovaSeq
platform with a total of around 2 billion reads (Supplementary Figure S1). Sequencing,
demultiplexing, and .bcl to .fastq file conversion was handled by the USC Molecular Genomics
Core.

Single Cell RNA Sequencing Analysis

Fastq files from the NovaSeq sequencing was then initially analyzed using the 10X
Genomics Cell Ranger software version 2.2.0. To analyze these samples, we created a novel
custom genome reference containing the human hg38 reference genome, KSHV JSC-1 Bac16
clone reference genome (GenBank: GQ994935.1), eGFP sequence, and mRFP sequence. This
genome was used to annotate transcripts from the rKSHV.219 virus used for infection.
After the initial analysis by Cell Ranger, we used the Seurat R toolkit for single cell
genomics v3.0 for further analysis. For additional computing power, we utilized the compute
nodes of the USC High Performance Computing Core. Pathway Analysis of the gene
expression was done using Qiagen’s Ingenuity Pathway Analysis software.
8

Results

rKSHV.219 Infection of the EpiOral Tissue

rKSHV.219 was initially applied to the apical surface of a raft culture 4 days after the
EpiOral tissues were lifted to the air-liquid interface. However, KSHV was unable to infect the
3D epithelial organoid culture. To achieve infection, we scored the top of the oral tissue in order
to expose the basal cell layer. This suggests KSHV cannot infect the intact apical surface of oral
keratinocytes, and KSHV infection is only possible when the basal layers are exposed due to a
cut or a wound.
Next, we performed fluorescence microscopy, real-time PCR, and scRNA-seq with the
KSHV-infected EpiOral tissues harvested at 3, 6, and 13 days post infection (Figure 3B).
Fluorescence microscopy of the rKSHV.219 infected tissues indicated successful KSHV
infection and demonstrated that KSHV undergoes spontaneous reactivation in our EpiOral
infection model. Quantitative-PCR (q-PCR) showed increased LANA, RTA, and K8 gene
Figure 3. (A) Schematic of rKSHV.219. GFP expression is under the control of the EF-1 promoter to indicate latent KSHV. RFP
is under the viral PAN promoter to indicate lytic KSHV. (B) Immunofluorescence of rKSHV.219 infected EpiOral tissues at 3, 6,
and 13 dpi. (C) Real-time PCR of LANA (latency), RTA (early lytic), and K8 (late lytic). KSHV copy number was also calculated
at each timepoint.
rKSHV.219
A
B
C
latent lytic
KSHV
Reactivation
GFP EF-1 Promoter
RFP
PAN Promoter
9

expression and viral DNA copy number, throughout the timepoints (Figure 3C). This confirms
our ability to establish a KSHV infection in the EpiOral tissue with an increasing KSHV lytic
population throughout the timepoints.

Single Cell RNA Sequencing

Using our single-cell transcriptomics data, we validated the ability of our 3D oral
epithelial organoid model to accurately replicate an oral KSHV infection. Using specific
differentiation markers such as CDH3 (Basal Layer), KRT6A (Spinous Layer), KRT10 (Granular
Layer), and SPRR2A (Cornified Layer), we annotated each cell into specific layers of the oral
epithelium (Figure 4C). We observed that the spatial clustering of mock tissue and 3dpi tissue
were very similar, indicating that there are very similar in gene expression, while the 6dpi
sample has its own distinct cluster (Figure 4A). This was also confirmed in the unsupervised
clustering (Figure 4B) where cells are clustered based on gene expression alone. For this
analysis, we excluded the 13dpi sample because the tissue was beginning to lose tissue
integrity and we were concerned about the host transcriptional changes irrelevant to the viral
infection. This analysis confirmed our ability to determine the epithelial layer of each cell in the
scRNA seq data, which we used later to investigate differences between latent and lytic KSHV
cells.

3dpi mock 6dpi
Basal
Granular
Cornified
Basal
Granular
Cornified
Cornified
Spinous
Granular
Granular
Basal
C
A
B
Figure 4. tSNE (T-distributed stochastic neighbor embedding) plots of scRNA-seq. Each dot represents one cell. (A) tSNE plot
of all cells grouped by sample. (B) tSNE plot of all cells grouped by unsupervised clustering. (C) Spatial clustering of the mock
(red), 3dpi (blue) and 6dpi (green) samples are projected into distinctly clusters that correspond to epithelial layers.

10

In the sequencing data, we took advantage of the GFP and RFP expression from
rKSHV.219 to specifically look at the gene expression of KSHV infected cells. GFP mRNA
expression was used to isolate all KSHV infected cells (Figure 5A) and RFP mRNA expression
was used to distinguish KSHV lytic from KSHV latent cells (Figure 5B and Figure 5C). We
detected clear expression of the majority of KSHV lytic genes while KSHV latency genes were
only detected in the latent KSHV cells. Although all of the seven latent genes should be
expressed during latency, the only latent genes detected in KSHV latent cells by 10x were
KSHV ORF71 or vFLIP. This is most likely due to the low abundance of these transcripts during
KSHV latency. Even though these genes were not picked up by the 10X platform, we did
confirm the presence of these transcripts in the real-time PCR data (Figure 3C).
A B C
Figure 5. Heatmaps of KSHV infected cells. Each column represents one cell and each row represents one viral gene. (A)
Heatmap of all KSHV infected cells. GFP expression from the rKSHV.219 virus was used to identify KSHV infected cells. (B)
Heatmap of KSHV latent cells. These cells have GFP expression but no RFP expression. (C) Heatmap of KSHV lytic cells. These
cells have GFP and RFP expression.
11

After identifying the viral gene expression in our sequencing data, we wanted to
determine which epithelial layers harbor the KSHV latent and lytic cells in our 3D EpiOral tissue.
A minimum of four gene expression markers for each of the basal, spinous, granular, and
cornified layers (Figure 6). We can clearly see that KSHV latent cells are not present in the
basal layer and the cornified layer. In addition, only KSHV lytic cells are the only cells present in
the cornified layer and not present in the granular layer. This suggests that KSHV initially infects
the basal layer of the epithelia, but fails to establish latency and reactivates as the basal
epithelial cell differentiates. Also, fully differentiated cornified epithelia are also unable to support
KSHV latency and leads to spontaneous reactivation in these cells as well. In the KSHV
infected cells, we investigated any differences in the overall gene expression between the latent
and lytic KSHV cells. KSHV Latency is characterized by minimal changes in the host gene
expression, while lytic KSHV is known to suppress host gene expression through KSHV SOX
mediated degradation of host mRNA
4
. In our single cell sequencing data, we compared all host
and viral gene expression between the KSHV latent and lytic cells (Figure 7). We can clearly
see an upregulation of KSHV viral genes in lytic cells, such as the KSHV lytic genes PAN and
K4. We can also detect a broad downregulation of host genes in the lytic cells compared to
latent cells. Surprisingly, we see the upregulation of four host genes: GOSR2, SPINK1, AVPL1,
and KPRP. Additional experiments are needed to confirm the significance of these upregulated
host genes and to show how these genes are possibly involved in lytic KSHV.
In our 3D EpiOral KSHV infection model, our interest was to see how the KSHV infected
cells may impact the surrounding non-infected cells. Using the GFP marker from rKSHV.219 to
PAN
ORF16
K4
mRFP
AVPI1
SPINK1
GOSR2
KPRP
EEF1A1
SAA1
SAA2
Host Genes
Viral Genes
Figure 7. Volcano plot of gene expression
comparing KSHV lytic cells against KSHV latent
cells. Each dot represents one gene. Non-
significant genes (p-value >0.05, fold-change
between -2 and 2) were filtered out before
generating the plot. Red dots indicate
upregulated genes and blue dots indicate down
regulated genes. The blue box highlights host
genes and the red box highlights viral genes.
Figure 6. Dot plot of KSHV latent and lytic cells. Gene expression markers for the epithelial layers are shown below the dots.
Each column of dots represents the data for one gene and each row represents either latent or lytic cells. The size of the dots
corresponds to the percent of cells expressing the gene and the color of the dots represent the average expression in of the gene
in either the latent or lytic cells.
12

identify all infected vs. non-infected EpiOral cells, we then compared the gene expression of the
non-infected cells between the mock, 3dpi, and 6dpi samples. The differential genes for each
group were then used for pathway analysis using Ingenuity Pathway Analysis (QIAGEN Inc.,
https://www.qiagenbioinformatics.com/products/ingenuitypathway-analysis) to understand how
they affect specific cellular pathways (Figure 8A). Integrin signaling, actin cytoskeleton
signaling, signaling by Rho family GTPases, and regulation of actin-based motility by rho all
showed upregulation in the mock and 3dpi samples and downregulation in the 6dpi sample. All
four pathways involve the activity of Rac1, which is known to stimulate NF-kB and AP-1
A B
D
C Canonical Pathways
Latent
Lytic
Activation Z-Score
-6 6
Figure 8. (A) Pathway analysis of non-infected EpiOral cells from Mock, 3dpi, and 6dpi samples. The canonical pathways in the red
box all include Rac1 in the pathway. (B) Violin plots of Rac1, CXCL8, and CXCL1 gene expression in non-infected EpiOral cells. Each
dot represents a cell and the colored violins behind represent how many cells are at that level of gene expression. (C) Pathway
analysis of KSHV latent and lytic cells. The canonical pathway in the red box includes Rac1 in the pathway. (D) Violin plots of Rac1,
CXCL8, and CXCL1 gene expression in KSHV latent and lytic cells.
Canonical Pathways Mock
3dpi
6dpi
Activation Z-Score
-2.5 2.5
13

signaling. NF-kB and AP-1 signaling leads to the production and secretion of chemokines such
as CXCL1 and CXCL8, which we see increased gene expression in the 6dpi sample compared
to the mock and 3dpi samples (Figure 8B). CXCL8 is known to promote immune cell infiltration,
which is characteristic of KS lesions. CXCL1 is a known agonist of KSHV ORF74, which is also
known to stimulate Rac1 activity in KSHV infected cells
8
. In the KSHV infected cells, we see that
the pathway for regulation of actin-based motility is again upregulated in KSHV latent cells and
downregulated in KSHV lytic cells (Figure 8C). Rac1 expression is decreased in KSHV lytic cells
compared to KSHV latent cells, while CXCL1 and CXCL8 increases with time after infection
(Figure 8D). The gene expression leads us to believe that there may be Rac1-mediated
signaling between the KSHV infected and non-infected cells in our infection model (Figure 9).
In our EpiOral KSHV infection model, KSHV initially tries to establish latent infection but
fails to do so in the basal layer of the EpiOral cells. These cells resort to a lytic KSHV infection
as the epithelial cell differentiates, resulting the release of CXCL1, CXCL8, and other
chemokines. CXCL1 and CXCL8 then promotes immune cell infiltration and upregulates Rac1
expression in the surrounding non-infected epithelial cells. The Rac1 activity in the uninfected
cells stimulates further release of CXCL1 and CXCL8. The secreted CXCL1 from both the non-
infected and KSHV lytic cells stimulate KSHV ORF74 in the KSHV lytic cells, demonstrating a
possible autocrine and paracrine signaling in between the KSHV lytic and non-infected EpiOral
cells.

Figure 9. Schematic of the possible autocrine
effect in KSHV lytic cells and the possible
paracrine effect between lytic KSHV cells and
non-infected cells.

14

Discussion

In our 3D EpiOral KSHV infection model, we are able to mimic in vivo oral KSHV
infection after exposing the basal layer of the tissue to rKSHV.219 virions. After initial infection,
the infected cells initially express GFP alone in its attempt to establish latency. These cells are
eventually unable to establish latency, resulting in spontaneous reactivation into RFP+ lytic
KSHV cells. This phenomenon was initially demonstrated in our immunofluorescence and real-
time PCR data, and again demonstrated later in our single-cell RNA sequencing data.
To overcome the loss of heterogeneity in traditional bulk RNA sequencing, we leveraged
the microfluidics technology of the 10X Genomics Chromium platform to maintain the
heterogeneity of the KSHV EpiOral infection model. We characterized the different epithelial
layers using gene expression makers and identified KSHV latent and lytic cells by the GFP and
RFP transcripts from rKSHV.219. Our data indicates that KSHV initially infects basal layer of the
oral epithelia, similar to human papilloma virus
11
. The epithelial layer gene expression markers
of the KSHV lytic cells indicated that KSHV lytic cells were not present in the basal layers,
supporting the previous paper that KSHV lytic cells are present in more differentiated oral
epithelial cells. The cornified cells of the most apical surface of the oral epithelia show
enrichment of lytic KSHV cells, which may strongly contribute to the high level of viral shedding
in oral KS detected in patients.
In terms of viral gene expression, we saw clear viral transcriptional patterns in the KSHV
latent and lytic cells. KSHV latent cells showed very low viral transcription, which proved to be
too low for the sensitivity of the 10X platform, but we were able to detect KSHV LANA
expression by real-time PCR. The only latent gene that was able to be detected by 10x was
KSHV vFLIP, which has been shown to activate NF-kB signaling and may also block autophagy,
possibly requiring higher expression during latency. KSHV lytic cells showed dramatic
upregulation of all KSHV viral genes, including the latent genes that were unable to be detected
in the KSHV latent cells.
On the other hand, we were able to see a dramatic downregulation of host mRNA in
KSHV lytic cells when looking at the host gene expression in KSHV infected cells. This clearly
matches the activity of KSHV SOX, a lytic protein which has been previously described as
promoting host mRNA turnover to suppress host gene expression. However, there were very
few host genes that were upregulated (GOSR2, SPINK1, AVPL1, and KPRP), which may be
interesting targets to study.
KSHV ORF74 encodes K-GPCR protein that has been previously shown to activate NF-
kB and AP-1 transcription factors through Rac1 activity. NF-kB and AP-1 mediated transcription
activates the transcription and secretion of CXCL1, CXCL8, and other cytokines
8
.CXCL8 is
involved in promoting immune cell infiltration, which is a hallmark of KS lesions. CXCL1 is a
chemokine may also act to promote immune cell infiltration, but also stimulates KSHV ORF74.
In the non-infected EpiOral cells, we believe that oral KSHV infection stimulates Rac1 gene
expression and activity. We believe that initial KSHV infection activates Rac1 gene expression
and activity, leading to CXCL1 and CXCL8 upregulation. CXCL1 secretion from the non-infected
cells stimulates ORF74 in the lytic KSHV cells. KSHV ORF74 activates Rac1 in the lytic KSHV
cells to promote CXCL1 and CXCL8 transcription and secretion. Secreted CXCL1 stimulates
both KSHV infected cells and continued KSHV infection stimulates Rac1 activity in the non-
infected cells (Figure 8E). This demonstrates a possible Rac-1/CXCL1 mediated autocrine
effect in KSHV lytic cells and paracrine effect between lytic KSHV cells and non-infected cells.
Further work will be needed to determine the exact mechanism of the Rac1 medicated autocrine
and paracrine effect.
Our ability to detect possible autocrine and paracrine effects further demonstrates the
ability of our infection model to mimic in vivo oral KSHV infection. The EpiOral KSHV infection
model is able to demonstrate an inability of KSHV to establish a latent infection in the oral
15

epithelium. Paired with single cell RNA sequencing, we are able to keep the heterogeneity
within the model by analyzing gene expression at a single cell resolution to discover novel
cellular interactions.

16

References
1. Butler, A., Hoffman, P., Smibert, P., Papalexi, E., & Satija, R. (2018). Integrating single-
cell transcriptomic data across different conditions, technologies, and species. Nature
Biotechnology. https://doi.org/10.1038/nbt.4096
2. Cesarman, E., Damania, B., Krown, S. E., Martin, J., Bower, M., & Whitby, D. (2019).
Kaposi sarcoma. Nature Reviews Disease Primers, 5(1), 9.
https://doi.org/10.1038/s41572-019-0060-9
3. Ganem, D. (1997). KSHV and Kaposi’s sarcoma: The end of the beginning? Cell.
https://doi.org/10.1016/S0092-8674(00)80398-0
4. Glaunsinger, B., & Ganem, D. (2004). Lytic KSHV infection inhibits host gene expression
by accelerating global mRNA turnover. Molecular Cell. https://doi.org/10.1016/S1097-
2765(04)00091-7
5. Johnson, A. S., Maronian, N., & Vieira, J. (2005). Activation of Kaposi’s Sarcoma-
Associated Herpesvirus Lytic Gene Expression during Epithelial Differentiation. Journal
of Virology. https://doi.org/10.1128/jvi.79.21.13769-13777.2005
6. Klausner, M., Ayehunie, S., Breyfogle, B. A., Wertz, P. W., Bacca, L., & Kubilus, J.
(2007). Organotypic human oral tissue models for toxicological studies. Toxicology in
Vitro. https://doi.org/10.1016/j.tiv.2007.01.024
7. Mesri, E. A., Cesarman, E., & Boshoff, C. (2010). Kaposi’s sarcoma and its associated
herpesvirus. Nature Reviews Cancer, 10(10), 707–719. https://doi.org/10.1038/nrc2888
8. Montaner, S., Sodhi, A., Servitja, J. M., Ramsdell, A. K., Barac, A., Sawai, E. T., &
Gutkind, J. S. (2004). The small GTPase Rac1 links the Kaposi sarcoma-associated
herpesvirus vGPCR to cytokine secretion and paracrine neoplasia. Blood.
https://doi.org/10.1182/blood-2003-12-4436
9. Pantanowitz, L., Khammissa, R. A. G., Lemmer, J., & Feller, L. (2013). Oral HIV-
associated Kaposi sarcoma. Journal of Oral Pathology and Medicine.
https://doi.org/10.1111/j.1600-0714.2012.01180.x
10. Pauk, J., Huang, M.-L., Brodie, S. J., Wald, A., Koelle, D. M., Schacker, T., … Corey, L.
(2002). Mucosal Shedding of Human Herpesvirus 8 in Men. New England Journal of
Medicine. https://doi.org/10.1056/nejm200011093431904
11. Schiller, J. T., Day, P. M., & Kines, R. C. (2010). Current understanding of the
mechanism of HPV infection. Gynecologic Oncology.
https://doi.org/10.1016/j.ygyno.2010.04.004

17

Supplementary Material
Supplementary Figure S1:

A B
C D
Figure S1. Cell Ranger output summaries for (A) Mock, (B) 3dpi, (C) 6dpi, and (D) 13dpi samples. Median reads per cell was above
80,000 and the median genes per cell was similar, ranging from 2,390 to 3,600. A custom reference genome containing human
hg38, KSHV JSC-1, eGFP, and mRFP was used for the annotation of the sequencing reads.

18

Supplementary Figure S2:

Figure S2. Quality control graphs from initial Seurat processing. (A) Scatterplot of mitochondrial read percentage (y-axis) vs number
of transcripts (x-axis). Each dot represents a cell. Cells with a mitochondrial percentage of over 10% was excluded from the analysis.
(B) Scatter plot of the number of genes (y-axis) vs number of transcripts (x-axis).
A
B

Abstract (if available)

Abstract Kaposi’s sarcoma-associated herpesvirus (KSHV) is the etiological agent of Kaposi’s sarcoma (KS) and the development of oral KS is often the first indication of HIV infection in patients. The oral cavity has been previously identified as the major site for shedding of infectious KSHV virions and is the most likely site for KSHV transmission. To study KSHV infection of the oral epithelium, we developed a three-dimensional (3D) oral tissue infection model that mimics in vivo oral KSHV infection. Exposure of the basal layers in the 3D oral tissues proved to be necessary for apical infection of KSHV. EpiOral tissues were harvested at three, six, and thirteen days post KSHV infection for fluorescence microscopy, real-time PCR, and single-cell RNA sequencing (scRNA seq). ScRNA seq showed that KSHV lytic cells were present only on the more differentiated layers of epithelium, indicating that KSHV fails to establish tight latency in the oral epithelia and spontaneously reactivates as the cells differentiate, especially in the cornified oral epithelia. In addition, we are able to show suppression of host mRNA expression by KSHV lytic reactivation, and a possible Rac1- mediated autocrine and paracrine effect between KSHV lytic cells and non-infected cells. These findings demonstrate the ability of our 3D oral KSHV infection model paired with single-cell RNA sequencing to improve our understanding of oral KSHV infection and transmission.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Epigenomic analyses of Kaposi's sarcoma-associated herpesvirus

PDF

Kaposi's sarcoma associated herpes-virus induces cellular proxi expression to modulate host gene expression that benefits viral infection and oncogenesis

PDF

Kaposi's sarcoma‐associated herpesvirus‐Bcl-2 mediated regulation of NM23-H2 is required for KSHV lytic replication

PDF

N-glycosylation on E protein and temperature regulate Zika virus infection in Aedes aegypti and Aedes albopictus mosquito cell

PDF

c-JUN mediated alteration of SLC2A2 expression in hepatoma cell line HepG2

PDF

Identification and characterization of adult stem cells in the oral cavity

Asset Metadata

Creator Jung, Kyle Lawrence (author)

Core Title Single cell gene expression analysis of KSHV infection in three-dimensional oral epithelial tissue cultures

School Keck School of Medicine

Degree Master of Science

Degree Program Molecular Microbiology and Immunology

Publication Date 12/17/2019

Defense Date 07/25/2019

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

Tag 3D culture,Kaposi's sarcoma associated herpesvirus,OAI-PMH Harvest,oral infection,single cell sequencing

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Jung, Jae U. (committee chair), Eoh, Hyungjin (committee member), Machida, Keigo (committee member)

Creator Email kylejung@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c89-256030

Unique identifier UC11673330

Identifier etd-JungKyleLa-8089.pdf (filename),usctheses-c89-256030 (legacy record id)

Legacy Identifier etd-JungKyleLa-8089.pdf

Dmrecord 256030

Document Type Thesis

Rights Jung, Kyle Lawrence

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 a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA