University of Southern California Dissertations and Theses

The effect of cytomegalovirus on gene expression of pediatric acute lymphoblastic leukemia

(USC Thesis Other)

The effect of cytomegalovirus on gene expression of pediatric acute lymphoblastic leukemia

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content Copyright 2021 Rachel Elizabeth Gallant
The Effect of Cytomegalovirus on Gene Expression of Pediatric Acute Lymphoblastic Leukemia
by
Rachel Elizabeth Gallant
A Thesis Presented to the
FACULTY OF THE USC KECK SCHOOL OF MEDICINE UNIVERSITY OF
SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
MASTER OF SCIENCE
(CLINICAL, BIOMEDICAL AND TRANSLATIONAL INVESTIGATIONS)
May 2021
ii

Table of Contents
List of Tables ............................................................................................................................... iii
List of Figures ............................................................................................................................ iv
Abstract ....................................................................................................................................... v
Chapter 1. The Association of CMV and Acute Lymphoblastic Leukemia and Its Effect on
Gene Expression of Leukemic Lymphoblasts ......................................................................... 1
ABSTRACT .............................................................................................................................. 1
INTRODUCTION ...................................................................................................................... 3
METHODS ................................................................................................................................ 7
RESULTS ................................................................................................................................. 9
DISCUSSION .......................................................................................................................... 14
References ................................................................................................................................. 20

iii
List of Tables
Table 1. Demographic and Clinical Characteristics and CMV Status ......................................... 10
Table 2. ALL Subtype and CMV Status ...................................................................................... 12
Table 3. Mutational Signature and CMV Status ......................................................................... 12
Table 4. Genes upregulated in CMV-positive cases .................................................................. 14
Table 5. Genes downregulated in CMV-positive cases .............................................................. 14

iv
List of Figures
Figure 1. Alignment Methods ....................................................................................................... 8

v
Abstract
Acute lymphoblastic leukemia (ALL) is the most common pediatric malignancy and
accounts for approximately thirty percent of all cancers in children. With advances in risk
stratification and treatment, the overall survival for ALL is now relatively high, but children with
ALL are still subject to the toxicities of chemotherapy and resultant late effects of therapy having
an impact not only in childhood, but into adulthood as well
1
. Therefore, it is imperative to
understand the etiology of childhood ALL in order to develop novel targets for treatment and
prevention with the goal of sparing children the acute and long-term toxicities of chemotherapy.
Recent evidence implicates Cytomegalovirus (CMV) as a risk factor in the development
of childhood ALL in that children with CMV infection early in life have higher rates of childhood
ALL than their CMV negative counterparts
2,3
. Still unknown is the mechanism by which CMV
confers a higher risk of ALL, and whether it affects a certain population of ALL patients such as
specific ethnicities, age groups, or cytogenetic subtypes.
In this study, we attempted to elucidate the relationship of CMV and ALL by evaluating
how CMV influences gene expression of precursor B-cell ALL lymphoblasts. CMV was detected
in diagnostic bone marrow samples of ALL patients by alignment of RNA-sequencing reads with
the CMV genome. Then analysis was performed to identify differentially expressed genes in
CMV-positive and CMV-negative cases.
Further investigation of the relationship between CMV and ALL is crucial, as it could
have significant impact on treatment and prevention of ALL representing a specific actionable
target.
1
Chapter 1. The Association of CMV and Acute Lymphoblastic Leukemia and Its Effect on
Gene Expression of Leukemic Lymphoblasts

ABSTRACT
Background: Advances in treatment have led to improved overall survival of patients with acute
lymphoblastic leukemia (ALL), the most common pediatric cancer. However, many patients still
sustain acute and chronic toxicities from treatment making the study of leukemia etiology
crucial to develop new therapies and identify targets for prevention. Patterns of infection have
long been considered a potential etiology for leukemia and recently cytomegalovirus (CMV) has
been identified as a risk factor for childhood ALL development.
Methods: RNA-sequencing data from diagnostic bone marrow samples of ALL patients was
obtained from Therapeutically Applicable Research to Generate Effective Treatments
(TARGET) database. This data was aligned to the human genome (hg38) as well as the CMV
genome (Merlin strain) to nominate CMV-positive cases of ALL. Differential gene expression
analysis was performed between CMV-positive and CMV-negative cases.
Results: CMV sequences were detected in the diagnostic bone marrow samples of 23/128 cases
(18%). CMV-positive cases were found to be associated with high risk according to the National
Cancer Institute risk stratification criteria (age ≥ 10 years and/or white blood cell count at
diagnosis ≥ 50K/mm
3
), but no association with ALL subtype or clinical outcome was
demonstrated in this analysis. Eleven genes were found to be statistically significantly
differentially expressed between CMV-positive and CMV-negative cases.
2
Conclusion: CMV infection may contribute to higher risk disease in children with ALL. The
impact of CMV on ALL development requires further elucidation and remains an important area
of research. CMV, if found to be a driver of ALL development, would be a modifiable risk factor
of ALL development and a potential target for leukemia prevention.

3
INTRODUCTION
Acute lymphoblastic leukemia (ALL) is the most common pediatric malignancy and
accounts for approximately thirty percent of all cancers in children.
1
With advances in treatment
and risk stratification, the overall survival for ALL is now relatively high and the majority of
patients achieve a remission without recurrence of disease.
1
Though survival has increased,
children with ALL are still subject to the toxicities of chemotherapy and resultant late effects of
therapy such as cognitive delays, stroke, neuropathy, cardiomyopathy, endocrinopathies, and
even secondary malignancies.
1
Medical management of these disorders may require lifelong
medications, blood tests or imaging studies for surveillance, and services such as physical or
occupational therapy. Sequelae such as these may impact school performance, social
interactions, interpersonal relationships, and employment opportunities having an effect not only
in childhood, but into adulthood as well. Therefore, it is imperative to understand the etiology of
childhood ALL in order to develop novel targeted treatments and inform prevention efforts with
the goal of sparing children the acute and long-term toxicities of chemotherapy.
Immune Dysregulation and ALL
Infection is theorized to contribute to the development of childhood leukemia presumably
through dysregulation of the immune system at an early age.
4–7
Many epidemiologic studies
investigating the association of childhood infections and development of ALL have been
performed and show variable associations. Those using record-based infection history tend to
show null or positive correlation between early life infection and ALL development.
4
Chang et al
performed a record-based case-control study and found that frequency of infections severe
enough to come to medical attention was associated with a higher rate of ALL development.
4

These types of studies are subject to misclassification as they rely on medical record or registry
data extraction. Alternatively, survey-based data, which may be complicated by recall bias, tend
to show a negative correlation between early life infections and ALL development.
4
For this
reason a surrogate for infection exposure, such as daycare attendance, has been used and is
4
shown to have a negative correlation with ALL suggesting that more frequent infections
common in children attending daycare are associated with lower ALL risk.
6

Mel Greaves proposes a potential mechanism to clarify these conflicting associations;
exposure to infectious agents early in life primes the immune system to respond appropriately to
such stimuli.
5
Children who lack early exposure to infections lack proper priming of their immune
system compared to those who are exposed, and when they do develop infection later in
childhood they have a dysregulated immune response with more severe symptoms.
5
This
immune dysregulation may allow pre-leukemic clones to expand and increase risk of developing
ALL later in childhood.
5,7
This concept has been demonstrated in a mouse model as well; mice
harboring a PAX5 deletion were born in a specific pathogen-free environment, and those with
delayed exposure to infection developed pre-B ALL, but those unexposed to infection did not.
8

Chang et al found that low levels of IL-10, an anti-inflammatory cytokine, measured in neonatal
blood spots was associated with higher risk of ALL development.
9
Additionally, Nielsen et al
showed an association between higher levels of arginase-2 in neonatal blood spots and
childhood ALL development.
10
Arginase-2 has an immunosuppressive effect which could alter
the immune response early in life affecting the risk of ALL.
10
Together, these findings support
the hypothesis of immune dysregulation contributing to the development of childhood ALL.
CMV and ALL
Though there is a great deal of evidence exploring the association between infection and
leukemia, a particular infectious agent has not been implicated. However, recent epidemiologic
and laboratory evidence suggests cytomegalovirus (CMV), a human herpes virus, may be an
important infectious agent contributing to the development of childhood ALL.
2,3
Francis et al
isolated RNA from diagnostic bone marrow aspirates of children with newly diagnosed ALL and
AML and performed total RNA sequencing to detect the presence of CMV RNA and found CMV
to be present in ALL samples much more frequently than AML samples (OR=18, 95% CI: 2.04-
159, p=0.003). The same authors screened neonatal blood spots of 268 ALL cases and 270
5
cancer-free controls for both CMV and EBV and found that children with CMV detected at birth
were 3.71 times as likely to develop ALL in childhood than those without CMV (p = 0.0016), and
this effect was more pronounced among Hispanic children (OR=5.90, 95% CI: 1.89-26, p =
0.006). There was no difference in the presence of EBV in cases compared to controls,
suggesting that CMV is uniquely associated with ALL risk.
2
A Swedish population-based cohort
study also found CMV infections to be associated with ALL development. Wiemels et al
identified children with hematologic malignancies and assessed health registry data for
documented history of CMV infections occurring at least six months prior to diagnosis of
leukemia and found that children with a history of CMV infection had over ten-fold higher rates
of hematologic malignancies than their CMV negative counterparts.
3
Furthermore, the authors
identified children in whom CMV was diagnosed at birth or who had a known maternal CMV
infection at birth and found significantly higher rates of hematologic malignancies in those with
either a maternal or personal history of CMV at the time of birth.
3
These two studies
demonstrate a novel association of CMV and childhood ALL that inspires further investigative
efforts, as the involvement of CMV for the first time identifies a specific target for prevention of
ALL. Still unknown is the mechanism by which CMV confers a higher risk of developing ALL.
Furthermore, it is unclear whether CMV affects a certain population of ALL patients such as
specific ethnicities, age groups, or cytogenetic subtypes.
CMV Biology
Cytomegalovirus is part of the Herpesviridae family, and is a common infection in the
general population, infecting 30-60% of pre-pubescent children and up to 90% of the population
by the age of 50 years.
11
Primary CMV infection can go unnoticed with mild clinical symptoms,
however even immune competent hosts cannot eradicate the virus. Ultimately CMV establishes
latent infection in hematopoietic cells such as monocytes and CD34+ cells.
12
CMV has a
complex relationship with the host immune system, and produces many viral proteins to
counteract host innate and adaptive immunity. CMV interferes with cytokine activity, NK cell
6
function, T cell response, chemokine activity, and prevents apoptosis,
13
and it is through these
mechanisms that CMV is able to survive in a latent state in host cells including hematopoietic
cells.
11
Furthermore, CMV has been shown to alter the host immune response to other stimuli.
Furman et al showed that those with a history of CMV have a more exaggerated immune
response to the influenza vaccine characterized by elevation of Th1 and Th2 cytokines and a
higher response of CD8+ T cells to IL-6.
14
We suspect that it is through this immune
dysregulation that children with CMV infection early in life are at a higher risk of developing
childhood ALL.
Though the overall survival in ALL approaches ninety percent, there are subsets of
patients that are considered high risk based on their cytogenetic features with much lower
survival rates. Relapse disease is leading cause of ALL-related mortality. Though typically
relapse is associated with higher risk disease, even some patients with favorable subtypes
typically associated with good prognosis, such as high hyperdiploidy and ETV6-RUNX1 fusion,
will relapse. Therefore, there remains room for improvement in ALL therapy and understanding
of the etiology. We aim to explain the role of CMV as a contributor to ALL development, an area
of research which has not yet been explored as it could have a significant impact on treatment
and/or prevention of ALL cases in the future. In this study, we aim to explore how CMV impacts
the gene expression of leukemia lymphoblasts at the time of diagnosis, by detecting CMV in
diagnostic ALL bone marrow samples using RNA sequencing (RNAseq) data and comparing
gene expression of CMV-positive to CMV-negative samples. Identifying differences in gene
expression based on CMV status may provide insight into potential etiologic mechanisms and
identify therapeutic targets that could augment ALL therapy for those found to be CMV-positive
at the time of diagnosis.

7
METHODS
RNA sequencing data were obtained from the Therapeutically Applicable Research to
Generate Effective Treatments (TARGET) database.
15
Patients were selected based on a
diagnosis of ALL and availability of both RNA and DNA sequencing data. Of the 178 cases
meeting this criteria, 119 samples were sequenced in 2013, 50 sequenced in 2016, and 9
samples that were sequenced in both 2013 and 2016. Though we did not intentionally select for
B-cell or T-cell ALL, all of the patient samples meeting the above criteria were from patients with
B-ALL. Quality of the RNA sequences were evaluated using Phred scores.
RNA sequencing reads were aligned to the human genome (hg38) using Spliced
Transcripts Alignment to a Reference (STAR) software. Any reads not mapping to the human
genome were then aligned to the CMV reference genome (Merlin strain) using Bowtie. Any
remaining unmapped reads were aligned to the RefSeq viral database using blastx (Figure 1).
Upon evaluation of these samples, there was concern for batch effect given that all samples
sequenced in 2016 had at least one read aligning to the CMV genome. Of the nine patient
samples sequenced at both time points, two had concordant results with at least one read
aligning to the CMV genome in both 2013 and 2016. However, the other seven showed
discordant CMV status with at least one read aligning to CMV when sequenced in 2016, but no
reads aligning to CMV when sequenced in 2013. Because of this discrepancy, and the unlikely
event that all 59 bone marrow samples sequenced in 2016 were truly CMV-positive we
suspected a batch effect, and therefore restricted our analysis to include only those patient
samples that underwent RNA-sequencing in 2013.

8

Figure 1. Alignment Methods
Demographic, clinical, and outcome information was obtained from the TARGET
database. ALL cytogenetic subtype was obtained from a prior study performed by Gu et al in
which they evaluated gene expression profiles of ALL samples including those in our dataset.
16

In their study, RNA sequencing data was analyzed using hierarchical clustering, t-distributed
stochastic neighbor embedding (tSNE), and predictive modeling.
16
Mutational signature data
were obtained from Ma et al; they performed genome-wide analysis to identify mutational
signatures for various pediatric cancers using DNA sequencing data from the TARGET
database including the ALL cases in our study.
17

Categorical data were analyzed using Chi-square test. Fisher’s exact test was used to
compare categorical variables when data were sparse. Means were compared using student t-
test, and medians compared using Wilcoxon Rank Sum test. Survival outcomes were compared
using Cox Proportional Hazards test. Differential gene expression was performed using the R
package DESeq2. Statistical analysis was performed using RStudio 1.2.1335.
9
RESULTS
After aligning RNAseq reads to the CMV genome, we found that of the patient samples
sequenced in 2013, 23 of 128 samples (18%) had at least one read aligning with the CMV
genome and therefore were considered CMV positive. There were 105 samples with no reads
aligning to the CMV genome and these were considered CMV-negative. Among the CMV-
positive samples, the number of reads aligning to the CMV reference genome ranged from 1 to
25 reads (mean: 3 ± 5.8, median: 1, IQR: 1.0, 1.5).
Clinical Features
Patient demographic and clinical features were compared between the CMV-positive
and CMV-negative groups. We found no significant differences in gender, ethnicity, WBC count
at diagnosis, or CNS disease status at diagnosis. Since age is used as a criterion for risk
stratification in ALL treatment, age at diagnosis was treated as a categorical variable (age >1
and <10 years, age ≥10 years) and CMV-positive patients were found to be older than their
CMV-negative (p = 0.03). As such, we found that within the CMV-positive group there was a
higher proportion of patients categorized in the National Cancer Institute (NCI) High Risk Group
than in the CMV-negative group (OR 3.56, 95% CI: 1.07, 15.27). The NCI Risk Stratification
criteria take into account age and white blood cell (WBC) count at diagnosis; those 10 years and
older and patients of any age with a WBC count greater than or equal to 50K/mm
3
fall into the
high risk disease group. High risk ALL requires more intensive treatment and portends a poorer
prognosis. Though NCI High Risk disease was more common in the CMV-positive group, we did
not find a statistically significant difference in overall or event free survival between the groups,
and relapse was the most common event in both groups (Table 1).

10
Table 1. Demographic and Clinical Characteristics and CMV Status in 128 ALL Patients

ALL Subtype, Mutational Signature, and CMV status
The TARGET database provided limited subtype information for each patient and for
many of the patients in our cohort subtype information was not recorded. Therefore, we
obtained more detailed information on ALL subtype from Gu et al including primary subtype and
fusion protein and/or secondary genomic lesions if applicable.
16
We compared the frequencies

CMV-positive
(N=23)
CMV-negative
(N=105)
p-value
Gender n(%)
Male
Female

15 (65.2%)
8 (34.8%)

51 (48.6%)
54 (51.4%)
0.15
Race n(%)
White
African American
Asian
Unknown (Not recorded)

17 (73.9%)
4 (17.4%)
0
2 (8.7%)

72 (68.6%)
11 (10.5%)
3 (2.9%)
19 (18.1%)
0.55
Ethnicity n(%)
Hispanic
Non-Hispanic
Unknown

4 (17.4%)
18 (78.3%)
1 (4.3%)

23 (21.9%)
66 (62.9%)
7 (6.7%)
0.66
Age at Diagnosis
Age >1 year and <10 years
Age 10 years or older
Range (years)

10 (43.5%)
13 (56.5%)
1.5 – 20.6

71 (67.6%)
34 (32.4%)
1.2 – 30.0

0.03
CNS Status at Diagnosis n(%)
CNS 1
CNS 2
CNS 3

17 (73.9%)
6 (26.1%)
0

86 (81.9%)
16 (15.2%)
3 (2.9%)
0.35
WBC Count at Diagnosis
Mean (K/mm
3
± SD)
Median (K/mm
3
(1
st
Quartile, 3
rd
Quartile))
WBC <50K/mm
3

WBC 50K/mm
3
or higher

99.2 ± 236.6
23.5 (8.4, 82.3)
14 (60.9%)
9 (39.1%)

62.4 ± 84.9
24.6 (12.5, 77.9)
65 (61.9%)
40 (38.1%)

0.45
0.64
0.93
Year of Diagnosis 2001 – 2008 2004 – 2010

Year of Last Follow Up 2005 – 2016 2005 – 2016

NCI Risk Stratification n(%)
Standard Risk
High Risk

4 (17.4%)
19 (82.6%)

45 (42.9%)
60 (57.1%)
OR (95% CI)
3.56 (1.07, 15.27)
Event Free Survival
Mean in years ± SD
Median in years (1
st
Quartile, 3
rd
Quartile)

4.5 ± 3.8
2.7 (1.8, 7.6)

3.3 ± 3.2
2.2 (1.2, 3.6)
HR (95% CI)
1.12 (0.63, 1.98), p=0.69
Event n(%)
Relapse
Death
Secondary Neoplasm
Censored

16 (69.6%)
-
1 (4.3%)
-

81 (77.1%)
3 (2.9%)
1 (1.0%)
1 (1.0%)
OR (95% CI)
0.68 (0.23, 2.19)
Overall Survival
Mean in years ± SD
Median in years (1
st
Quartile, 3
rd
Quartile)

6.06 ± 3.42
5.6 (3.6, 9.2)

4.8 ± 3.6
3.2 (1.8, 8.4)
HR (95% CI)
1.21 (0.63, 2.34) p=0.57
11
of each subtype between the CMV-positive and CMV-negative groups and did not find any
particular subtype to be enriched in the CMV-positive group (Table 2). Mutational signatures
were obtained from Ma et al as described above.
17
We were particularly interested in the
APOBEC (COSMIC S-2 and S-13) mutational signatures as these are associated with cancers
with a known infectious etiology (specifically cervical and head/neck cancers). Therefore, we
looked for an association between CMV status and the APOBEC signature, but did not find
CMV-positive cases to be associated with the APOBEC signature in this dataset. Upon analysis
of all mutational signatures we found that the COSMIC S-8 signature was present in a higher
proportion of CMV-positive patients compared to CMV-negative with an odds ratio of 2.60
(Table 3). COSMIC S-8 is characterized by a weak strand bias for C>A substitutions, but the
etiology of this signature is unknown. It has been described in a variety of cancers including
bone, CNS, pancreatic, ovarian, prostate, lung, bladder, breast, and stomach cancers.
Heretofore, the COSMIC S-8 mutational signature has not been reported in hematologic
malignancies.

12
Table 2. ALL Subtype and CMV Status
ALL Subtype CMV-positive
(N=23)
CMV-negative
(N=105)
OR (95% CI) p-value
BCR-ABL1 0 5 (4.8%)

ETV6-RUNX1 2 (8.7%) 10 (9.5%) 0.91 (0, 4.00) 1.00
ETV6-RUNX1-like 0 4 (3.8%)

Hyperdiploid 5 (21.7%) 18 (17.1%) 1.34 (0.51, 4.41) 0.47
DUX4 0 1 (0.9%)

iAMP21 1 (4.3%) 4 (3.8%) 1.15 (0, 8.17) 1.00
MEF2D 0 3 (2.9%)

MLL-Rearranged 1 (4.3%) 2 (1.9%) 2.34 (0, 18.90) 0.45
PAX5alt 1 (4.3%) 8 (7.6%) 0.55 (0, 3.64) 1.00
PAX5 P80R 0 0

Ph-like (CRLF2) 5 (21.7%) 15 (14.3%) 1.67 (0.56, 5.01) 0.37
HLF 1 (4.3%) 2 (1.9%) 2.34 (0, 18.90) 0.45
CRLF2 (Non-Ph-Like) 0 0

TCF3-PBX1 0 16 (15.2%)

ZNF384 4 (17.4%) 6 (5.7%) 3.47 (0.96, 12.72) 0.08
ZNF384-like 1 (4.3%) 0

Other 0 3 (2.9%)

NA 2 (8.7%) 8 (7.6%)

Table 3. Mutational Signature and CMV Status
Mutational Signature CMV-positive
(n=23)
CMV-negative
(n=105)
OR (95% CI) p-value
COSMIC S-1; age 23 (100%) 105 (100%) NA

COSMIC S-2; APOBEC 3 (13.0%) 14 (13.3%) 0.98 (0.28, 3.50) 1.00
COSMIC S-3; HR deficiency 3 (13.0%) 26 (24.8%) 0.46 (0.13, 1.57) 0.28
COSMIC S-5; age 23 (100%) 101 (96.2%) NA

COSMIC S-7; UV light 1 (4.3%) 4 (3.8%) 1.15 (0, 8.17) 1.00
COSMIC S-8; unknown 10 (43.5%) 24 (22.9%) 2.60 (1.03, 6.56) 0.04
COSMIC S-13; APOBEC 1 (4.3%) 6 (5.7%) 0.75 (0, 5.09) 1.00
COSMIC S-18; ROS 8 (34.8%) 45 (42.9%) 0.71 (0.28, 1.79) 0.48
COSMIC S-26; MSI 0 2 (1.9%) NA

T10; unknown 4 (17.4%) 19 (18.1%) 0.95 (0.30, 3.00) 1.00
T11; unknown 3 (13.0%) 10 (9.5%) 1.43 (0.39, 5.30) 0.70

13
Differential Gene Expression
Differential gene expression was performed using DESeq2 R package. Eleven genes
were found to be statistically significantly differentially expressed between the CMV-positive and
CMV-negative groups. Most of these genes are pseudogenes or miRNA genes. Ten of the 11
top differentially expressed genes in our study are pseudogenes or miRNA genes (Table 4,
Table 5). For pseudogenes with a known parent gene, we explored the role of the parent gene
in leukemia with the knowledge that the associated pseudogene may play a role in its
expression or activity. NPM1P17 (nucleophosmin 1 pseudogene 17) and RPS27AP14
(ribosomal protein S27a pseudogene 14) were found to have significantly lower expression in
CMV-positive cases compared to CMV-negative cases. MIR504, a microRNA, is also
downregulated in CMV-positive cases. TCL1B, a gene though to play a role in B-cell and T-cell
malignancies, is upregulated in CMV-positive compared to CMV-negative cases. Nine additional
genes approached significance for differential expression (Table 4 and Table 5). Of particular
interest is the upregulation of CXCL10, a proinflammatory chemokine, in CMV-positive cases.
CXCL10 has been associated with both CMV and B-ALL and therefore could provide a potential
mechanistic link between CMV infection and development of childhood ALL.

14
Table 4. Genes upregulated in CMV-positive cases
Gene Log Fold
Change
p Adjusted
TCL1B TCL1 Family AKT Coactivator B 2.91 3.18e
-02

MIR6724-4 MicroRNA 6724-4 18.41 3.32e
-02

IL1RN Interleukin 1 Receptor Antagonist 1.86 5.41e
-02

FTH1P7 Ferritin Heavy Chain 1 Pseudogene 7 1.34 5.53e
-02

PTPRS Protein Tyrosine Phosphatase Receptor Type S 1.74 7.84e
-02

CXCL10 C-X-C Motif Chemokine Ligand 10 2.56 7.96e
-02

GTSF1 Gametocyte Specific Factor 1 2.26 9.08e
-02

CSPG4 Chondroitin Sulfate Proteoglycan 4 1.84 9.08e
-02

Table 5. Genes downregulated in CMV-positive cases
Gene Log Fold
Change
p Adjusted
NPM1P17 Nucleophosmin 1 pseudogene 17 -29.53 2.14e
-08

MIR504 MicroRNA 504 -24.11 8.71e
-05

PNMA6B PNMA family member 6B -22.40 6.85e
-04

AC112492.2 Pseudogene -21.78 1.20e
-03

MAPK8P1P2 Mitogen-Activated Protein Kinase 8 Interacting Protein 1
Pseudogene
-4.43 1.31e
-03

AC108457.1 RNA gene -3.73 8.76e
-03

AC005829.2 Pseudogene -3.36 1.07e
-02

AL024493.1 Pseudogene -19.51 1.14e
-02

RPS27AP14 RPS27A Pseudogene 14 -18.00 4.98e
-02

ARL17B ADP Ribosylation Factor-Like GTPase 17B -1.56 5.41e
-02

RPL7AP28 Ribosomal Protein L7a Pseudogene 28 -4.85 5.53e
-02

AC018521.7 RNA gene -2.71 5.53e
-02

DISCUSSION
In this study we show that CMV RNA sequences can be detected in diagnostic leukemia
bone marrow samples. An interesting clinical finding in our analysis is the association of NCI
High Risk stratification and presence of CMV in the bone marrow. Interestingly, despite this
association there was not a significant difference in clinical outcomes between CMV-positive
15
and CMV-negative cases. This is likely due to the fact that the TARGET database is enriched for
patients with relapse disease. Typically, relapse would be expected in only 10-20% of children
with ALL.
1
However, in this cohort relapse occurred in 74% of CMV-positive cases and 77% of
CMV-negative cases which is much higher than expected for B-ALL regardless of NCI risk
stratification. Because of this relapse-enriched cohort the outcomes are likely not an accurate
depiction of the effect of CMV on clinical outcome for the overall population.
We also found it interesting that this study showed no association between CMV status
and ethnicity as Francis et al showed that history of congenital CMV infection was more
strongly predictive of ALL development in those of Hispanic ethnicity.
2
Furthermore, both CMV
infection and childhood ALL are more common in Hispanics.
18–20
Therefore, we suspected CMV
would be more commonly detected in the diagnostic bone marrows of Hispanic patients
compared to non-Hispanic. Only 27 patients in this cohort were identified as Hispanic, and this
small sample size may limit our ability to detect a significant difference in CMV status. The
association between Hispanic ethnicity and CMV seropositivity and pediatric ALL remains an
intriguing question worthy of further exploration.
No specific ALL subtype or mutational signature with a proposed etiology was found to
be enriched in CMV-positive cases, though we did find COSMIC-S8 signature to be associated
with CMV-positive status. Pediatric ALL has a great deal of genetic diversity with many different
subtypes; therefore it is possible that our sample size limited our ability to detect associations
between CMV status and ALL subtype or mutational signature. Repeating this analysis on a
larger sample size could reveal associations not seen in our study.
16
Many of the genes found to be significantly differentially expressed between CMV-
positive and CMV-negative groups are pseudogenes or miRNA genes. Pseudogenes, though
previously thought be non-functional components of the genome, have been shown to
contribute to regulation often of expression of their parent gene or in some cases of different
targets.
21,22
Pseudogenes can mimic or interfere with their parent genes and their function at the
DNA, RNA, or protein level.
21
The precise functions of many pseudogenes have still not been
characterized.
22
However, many pseudogenes play a role in regulation of gene expression of
their parent gene or in some cases other targets independent of the parent gene. Our study
showed that NPM1P17 (nucleophosmin 1 pseudogene 17) has significantly lower expression in
CMV-positive cases compared to CMV-negative cases. Though the function of NPM1P17 has
not been described, it’s parent gene NPM1 is involved in DNA-damage repair, ribosome
biogenesis, and apoptosis, and has been reported to be downregulated in acute myeloid
leukemia (AML) and T-cell ALL.
23
It is also a known fusion partner in acute promyelocytic
leukemia (APL), myelodysplastic syndrome (MDS), and non-Hodgkin lymphoma (NHL).
23

MIR504, also downregulated in CMV-positive cases, is a miRNA that has been shown to act on
multiple different targets affecting tumorigenesis of various cancers including glioma, oral
squamous cell carcinoma, and non-small cell lung cancer, but has not been shown to have an
effect on the development of hematologic malignancies.
24–27
RPS27AP14, is also
downregulated in CMV-positive cases. It’s precise function is not known, but it’s parent gene
RPS27A has been shown to contribute to cell cycle progression, apoptosis inhibition, and to be
overexpressed in both acute leukemias and chronic myeloid leukemia (CML).
28
TCL1B is
upregulated in CMV-positive compared to CMV-negative cases. TCL1 family proteins bind and
stabilize AKT promoting cell proliferation, growth, and survival.
29
TCL1B is normally expressed
during embryogenesis and in immature T and B lymphocytes, and its expression has been
detected in fetal liver, spleen, thymus, bone marrow, and peripheral blood lymphocytes.
29
The
17
TCL1 family of proteins has been described to play a role in oncogenesis of both B-cell and T-
cell hematologic malignancies, though not B-cell ALL specifically.
29
Though these genes could
be related to leukemia development, they do not seem to have a relation to CMV and therefore
are less likely to help explain the association between CMV infection and ALL.
CXCL10 expression was upregulated in CMV-positive cases compared to CMV-negative
cases. It is a proinflammatory chemokine typically secreted in response to IFN-gamma. The
CXCL10 receptor (CXCR3) is highly expressed on B-ALL cells, and exposure to CXCL10 has
been shown to promote migration and invasion of B-lymphoblasts.
30
Furthermore, Lee et al
showed that when monocytes are co-cultured with B-cell ALL cells, the monocytes exhibit
significant upregulation of CXCL10 which contributes to the increased migration and invasion
capabilities of B-ALL cells suggesting that CXCL10 may play a role in leukemia development.
30

Interestingly, CMV has been shown to play a role in CXCL10 production. When prenatal CMV
infection occurs, it is the chorioaminion that is infected by CMV.
31
Putri et al infected human
amniotic epithelial cells with CMV and found that there was an increase in transcription of
chemokines in response to the infection, namely CCL5, CXCL8, and CXCL10. They
demonstrated similar results in a guinea pig model.
31
Scott et al measured cytokines from
amniotic fluid of eight patients with congenital CMV infection as well as 12 women with primary
CMV infection during pregnancy and compared these to cytokine profiles of the amniotic fluid
and sera of uninfected women. They found that CXCL10, among other cytokines, was
significantly elevated in amniotic fluid of congenital CMV patients. Furthermore, in maternal sera
CXCL10 was significantly higher in CMV-infected women compared to non-infected controls.
32

CXCL10 has been shown to be produced by other tissues, including microglia, in response to
CMV.
33
These studies together suggest that CMV promotes production of CXCL10 which in turn
may play a role in B-ALL development through its effect on migration and invasion abilities of
precursor B-cells. The increased expression of CXCL10 in CMV-positive B-ALL lymphoblasts
18
compared to CMV-negative cases, suggesting that CXCL10 could play a role in the association
of CMV infection early in life and development of B-ALL. The possibility of CMV driving ALL
development through immune dysregulation is a question that certainly warrants further
exploration to confirm the association between CMV at the time of leukemia diagnosis and
CXCL10 expression and further delineate the mechanism.
Limitations
As RNAseq data was obtained retrospectively from a database, the quality of the RNA,
sequencing depth, and processing of the sequencing data were set before we accessed the data
for analysis. Though we are reasonably assured of the quality of the RNA based on evaluation of
Phred scores, these factors can still impact the completeness of the reads and proper alignment
of the reads to the reference genome. Additionally, of the CMV-positive samples the number of
reads aligning to the CMV reference genome are variable and some are low with read counts
ranging from 1-25. Those with low CMV-aligned read counts raise the question of whether
these represent true presence of CMV or if it is artifact. This risk is reduced somewhat by first
aligning to the human genome and using only the unmapped reads to align to CMV as we did in
this study, but this does not completely ameliorate the risk of false positives. Finally, our
sample size was too small to accurately identify associations between CMV status and ALL
subtype or to stratify results by subtype given the wide genetic variability of ALL. Stratification
of gene expression results by subtype will be important in future studies as gene expression
profiles differ by ALL subtype.
Conclusion
19
We found CMV-positive B-ALL cases to be associated with high risk disease per NCI risk
stratification criteria, though we did not identify any other clinical characteristics to be
associated with CMV status in this study. Through differential gene expression analysis we
found CXCL10 to be upregulated in CMV-positive ALL bone marrow samples and proposed a
possible mechanism for development of ALL driven by CMV infection early in life mediated by
CXCL10. Given the limitations of the current study, future studies will be necessary to confirm
and expand on these findings. This remains a very exciting area of research with significant
clinical implications. If CMV is shown to be a trigger or driver of ALL development, then
screening and prompt treatment of CMV with antiviral agents could be used to augment ALL
therapy or even to prevent a subset of cases of ALL. Furthermore, there are ongoing efforts to
develop a vaccine for CMV as it has other devastating sequelae such as hearing loss when
acquired congenitally.
34
Such a vaccine could play a role in prevention of a subset of cases of
ALL by preventing CMV infection, ultimately sparing children some of the toxic side effects and
psychological trauma of chemotherapy. CMV represents a unique, potential active target for
ALL prevention.
20
References
1. Orkin S, Fisher D, Ginsburg D, Look A, Lux S, Nathan D. Nathan and Oski’s Hematology and
Oncology of Infancy and Childhood. Vol 2. 8th ed. Elsevier; 2015.
2. Francis SS, Wallace AD, Wendt GA, et al. In utero cytomegalovirus infection and
development of childhood acute lymphoblastic leukemia. Blood. 2017;129(12):1680-1684.
doi:10.1182/blood-2016-07-723148
3. Wiemels JL, Talbäck M, Francis S, Feychting M. Early Infection with Cytomegalovirus and
Risk of Childhood Hematologic Malignancies. Cancer Epidemiol Prev Biomark.
2019;28(6):1024-1027. doi:10.1158/1055-9965.EPI-19-0044
4. Chang JS, Tsai C-R, Tsai Y-W, Wiemels JL. Medically diagnosed infections and risk of
childhood leukaemia: a population-based case–control study. Int J Epidemiol.
2012;41(4):1050-1059. doi:10.1093/ije/dys113
5. Greaves M. Infection, immune responses and the aetiology of childhood leukaemia. Nat
Rev Cancer. 2006;6(3):193-203. doi:10.1038/nrc1816
6. Rudant J, Lightfoot T, Urayama KY, et al. Childhood Acute Lymphoblastic Leukemia and
Indicators of Early Immune Stimulation: A Childhood Leukemia International Consortium
Study. Am J Epidemiol. 2015;181(8):549-562. doi:10.1093/aje/kwu298
7. Wiemels JL, Cazzaniga G, Daniotti M, et al. Prenatal origin of acute lymphoblastic
leukaemia in children. The Lancet. 1999;354(9189):1499-1503. doi:10.1016/S0140-
6736(99)09403-9
8. Hauer J, Martín-Lorenzo A, Sánchez-García I. Infection causes childhood leukemia. Aging.
2015;7(9):607-608.
9. Chang JS, Zhou M, Buffler PA, Chokkalingam AP, Metayer C, Wiemels JL. Profound Deficit
of IL10 at Birth in Children Who Develop Childhood Acute Lymphoblastic Leukemia. Cancer
Epidemiol Prev Biomark. 2011;20(8):1736-1740. doi:10.1158/1055-9965.EPI-11-0162
10. Nielsen AB, Zhou M, de Smith AJ, et al. Increased neonatal level of arginase 2 in cases of
childhood acute lymphoblastic leukemia implicates immunosuppression in the etiology.
Haematologica. 2019;104(11):e514-e516. doi:10.3324/haematol.2019.216465
11. Picarda G, Benedict CA. Cytomegalovirus: Shape-Shifting the Immune System. J Immunol.
2018;200(12):3881-3889. doi:10.4049/jimmunol.1800171
12. La Rosa C, Diamond DJ. The immune response to human CMV. Future Virol. 2012;7(3):279-
293. doi:10.2217/fvl.12.8
21
13. Miller-Kittrell M, Sparer TE. Feeling manipulated: cytomegalovirus immune manipulation.
Virol J. 2009;6(1):4. doi:10.1186/1743-422X-6-4
14. Furman D, Jojic V, Sharma S, et al. Cytomegalovirus infection improves immune responses
to influenza. Sci Transl Med. 2015;7(281):281ra43. doi:10.1126/scitranslmed.aaa2293
15. GenomeOC. Overview. Office of Cancer Genomics. Published February 4, 2013. Accessed
March 2, 2021. https://ocg.cancer.gov/programs/target/overview
16. Gu Z, Churchman ML, Roberts KG, et al. PAX5-driven subtypes of B-progenitor acute
lymphoblastic leukemia. Nat Genet. 2019;51(2):296-307. doi:10.1038/s41588-018-0315-5
17. Ma X, Liu Y, Liu Y, et al. Pan-cancer genome and transcriptome analyses of 1,699 paediatric
leukaemias and solid tumours. Nature. 2018;555(7696):371-376.
doi:10.1038/nature25795
18. Bate SL, Dollard SC, Cannon MJ. Cytomegalovirus seroprevalence in the United States: the
national health and nutrition examination surveys, 1988-2004. Clin Infect Dis Off Publ
Infect Dis Soc Am. 2010;50(11):1439-1447. doi:10.1086/652438
19. Giddings BM, Whitehead TP, Metayer C, Miller MD. Childhood Leukemia Incidence in
California: High and Rising in the Hispanic Population. Cancer. 2016;122(18):2867-2875.
doi:10.1002/cncr.30129
20. Siegel DA, Henley SJ, Li J, Pollack LA, Van Dyne EA, White A. Rates and Trends of Pediatric
Acute Lymphoblastic Leukemia — United States, 2001–2014. MMWR Morb Mortal Wkly
Rep. 2017;66(36):950-954. doi:10.15585/mmwr.mm6636a3
21. Xiao-Jie L, Ai-Mei G, Li-Juan J, Jiang X. Pseudogene in cancer: real functions and promising
signature. J Med Genet. 2015;52(1):17-24. doi:10.1136/jmedgenet-2014-102785
22. Hu X, Yang L, Mo Y-Y. Role of Pseudogenes in Tumorigenesis. Cancers. 2018;10(8).
doi:10.3390/cancers10080256
23. Brodská B, Šašinková M, Kuželová K. Nucleophosmin in leukemia: Consequences of anchor
loss. Int J Biochem Cell Biol. 2019;111:52-62. doi:10.1016/j.biocel.2019.04.007
24. Cui R, Guan Y, Sun C, et al. A tumor-suppressive microRNA, miR-504, inhibits cell
proliferation and promotes apoptosis by targeting FOXP1 in human glioma. Cancer Lett.
2016;374(1):1-11. doi:10.1016/j.canlet.2016.01.051
25. Hu W, Chan CS, Wu R, et al. Negative Regulation of Tumor Suppressor p53 by microRNA
miR-504. Mol Cell. 2010;38(5):689-699. doi:10.1016/j.molcel.2010.05.027
26. Wang X, Chang K, Gao J, et al. MicroRNA-504 functions as a tumor suppressor in oral
squamous cell carcinoma through inhibiting cell proliferation, migration and invasion by
22
targeting CDK6. Int J Biochem Cell Biol. 2020;119:105663.
doi:10.1016/j.biocel.2019.105663
27. Ye M-F, Zhang J-G, Guo T-X, Pan X-J. MiR-504 inhibits cell proliferation and invasion by
targeting LOXL2 in non small cell lung cancer. Biomed Pharmacother Biomedecine
Pharmacother. 2018;97:1289-1295. doi:10.1016/j.biopha.2017.11.005
28. Wang H, Yu J, Zhang L, et al. RPS27a promotes proliferation, regulates cell cycle
progression and inhibits apoptosis of leukemia cells. Biochem Biophys Res Commun.
2014;446(4):1204-1210. doi:10.1016/j.bbrc.2014.03.086
29. Teitell MA. The TCL1 family of oncoproteins: co-activators of transformation. Nat Rev
Cancer. 2005;5(8):640-648. doi:10.1038/nrc1672
30. Lee Y, Chittezhath M, André V, et al. Protumoral role of monocytes in human B-cell
precursor acute lymphoblastic leukemia: involvement of the chemokine CXCL10. Blood.
2012;119(1):227-237. doi:10.1182/blood-2011-06-357442
31. Putri DS, Berkebile ZW, Mustafa HJ, et al. Cytomegalovirus infection elicits a conserved
chemokine response from human and guinea pig amnion cells. Virology. 2020;548:93-100.
doi:10.1016/j.virol.2020.06.005
32. Scott GM, Chow SSW, Craig ME, et al. Cytomegalovirus Infection During Pregnancy With
Maternofetal Transmission Induces a Proinflammatory Cytokine Bias in Placenta and
Amniotic Fluid. J Infect Dis. 2012;205(8):1305-1310. doi:10.1093/infdis/jis186
33. Cheeran MC-J, Hu S, Sheng WS, Peterson PK, Lokensgard JR. CXCL10 Production from
Cytomegalovirus-Stimulated Microglia Is Regulated by both Human and Viral Interleukin-
10. J Virol. 2003;77(8):4502-4515. doi:10.1128/JVI.77.8.4502-4515.2003
34. Naing ZW, Scott GM, Shand A, et al. Congenital cytomegalovirus infection in pregnancy: a
review of prevalence, clinical features, diagnosis and prevention. Aust N Z J Obstet
Gynaecol. 2016;56(1):9-18. doi:https://doi.org/10.1111/ajo.12408

Abstract (if available)

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

The role of survivin in drug resistant pediatric acute lymphoblastic leukemia

PDF

Understanding acute lymphoblastic leukemia in different ethnic groups in the United States

PDF

Randomized clinical trial generalizability and outcomes for children and adolescents with high-risk acute lymphoblastic leukemia

PDF

Genetic epidemiological approaches in the study of risk factors for hematologic malignancies

PDF

The influence of DNA repair genes and prenatal tobacco exposure on childhood acute lymphoblastic leukemia risk: a gene-environment interaction study

PDF

Clinical outcomes of allogeneic hematopoietic stem cell transplant in acute lymphoblastic leukemia patients: a quality improvement project and systematic review meta-analysis

PDF

Integrin mediated cellular adhesion may alter the cytokine profile in acute lymphoblastic leukemia

PDF

Genetic and environmental risk factors for childhood cancer

PDF

Induction therapy in relapse adult acute lymphoblastic leukemia

PDF

Role of integrin α4 in drug resistant acute lymphoblastic leukemia

PDF

Ancestral/Ethnic variation in the epidemiology and genetic predisposition of early-onset hematologic cancers

PDF

The role of adipocyte-derived free-fatty acids in acute lymphoblastic leukemia

PDF

Induction of hypersignaling as a therapeutic approach for treatment of BCR-ABL1 positive Acute Lymphoblastic Leukemia (ALL) cells

PDF

Genes and hormonal factors involved in the development or recurrence of breast cancer

PDF

The interplay between tobacco exposure and polygenic risk score for growth on birthweight and childhood acute lymphoblastic leukemia

PDF

Competency based education in low resource settings: design and implementation of a novel surgical training program

PDF

Exploring the interplay of birth order and birth weight on leukemia risk

PDF

Exploring the effects of CXCR4 inhibition on circulating tumor cell populations in metastatic prostate cancer

PDF

The role of adipocytes in acute lymphoblastic leukemia cell migration and survival against daunorubicin

PDF

The environmental and genetic determinants of cleft lip and palate in the global setting

Asset Metadata

Creator Gallant, Rachel Elizabeth (author)

Core Title The effect of cytomegalovirus on gene expression of pediatric acute lymphoblastic leukemia

School Keck School of Medicine

Degree Master of Science

Degree Program Clinical, Biomedical and Translational Investigations

Publication Date 03/30/2021

Defense Date 03/23/2021

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

Tag acute lymphoblastic leukemia,All,cmv,Cytomegalovirus,gene expression,OAI-PMH Harvest,pediatric leukemia,RNA sequencing

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Wiemels, Joseph (committee chair), Bhojwani, Deepa (committee member), DeSmith, Adam (committee member), McKean-Cowdin, Roberta (committee member)

Creator Email rgallant@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c89-433738

Unique identifier UC11668234

Identifier etd-GallantRac-9374.pdf (filename),usctheses-c89-433738 (legacy record id)

Legacy Identifier etd-GallantRac-9374.pdf

Dmrecord 433738

Document Type Thesis

Rights Gallant, Rachel Elizabeth

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