Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
Filters applied. Clear all
. 2010 Nov;84(22):11771-80.
doi: 10.1128/JVI.01355-10. Epub 2010 Sep 15.

High-throughput, Sensitive Quantification of Repopulating Hematopoietic Stem Cell Clones

Affiliations
Free PMC article

High-throughput, Sensitive Quantification of Repopulating Hematopoietic Stem Cell Clones

Sanggu Kim et al. J Virol. .
Free PMC article

Abstract

Retroviral vector-mediated gene therapy has been successfully used to correct genetic diseases. However, a number of studies have shown a subsequent risk of cancer development or aberrant clonal growths due to vector insertion near or within proto-oncogenes. Recent advances in the sequencing technology enable high-throughput clonality analysis via vector integration site (VIS) sequencing, which is particularly useful for studying complex polyclonal hematopoietic progenitor/stem cell (HPSC) repopulation. However, clonal repopulation analysis using the current methods is typically semiquantitative. Here, we present a novel system and standards for accurate clonality analysis using 454 pyrosequencing. We developed a bidirectional VIS PCR method to improve VIS detection by concurrently analyzing both the 5' and the 3' vector-host junctions and optimized the conditions for the quantitative VIS sequencing. The assay was validated by quantifying the relative frequencies of hundreds of repopulating HPSC clones in a nonhuman primate. The reliability and sensitivity of the assay were assessed using clone-specific real-time PCR. The majority of tested clones showed a strong correlation between the two methods. This assay permits high-throughput and sensitive assessment of clonal populations and hence will be useful for a broad range of gene therapy, stem cell, and cancer research applications.

Figures

FIG. 1.
FIG. 1.
Quantitative sequencing via pyrosequencing-based, bidirectional vector integration site analysis. (a) Sequencing efficiency for different lengths of DNA. The sequencing efficiency dropped for DNA of >500 bp but was relatively stable for DNA of ≤500 bp when 15 control DNAs (lengths ranging from 125 to 1,680 bp) were tested by the 454 pyrosequencing. (b and c) Analysis of the serially diluted control VIS DNA showed linear correspondence of the relative sequence frequencies to the dilution factor for DNA of <500 bp. A mixture of four control VIS DNAs was serially diluted and mixed with the VIS DNA from acutely infected cells (b). Panel B shows a 2% agarose gel run after the second PCR. Lanes 1 to 7, 2-fold serial dilutions of the control DNAs; lane N, water; lane M, 2-log DNA marker; dark arrows, four control VIS DNA; empty arrows, internal control (vector, LTR circles from acute infection). The relative sequence counts of the control VIS DNAs of <500 bp are in linear correlation with the expected values after 2-fold dilution (c). (d) The schematic view of the assay. The bidirectional VIS analysis enabled concurrent analysis of both the left and the right vector-host DNA junctions. Amplified junction DNA was then subjected to 454 pyrosequencing, followed by controlled sequence analysis. (e) Percentage of random vector integrants that can generate VIS DNA of <500 bp. The length of VIS DNA was determined based on the location of the nearest TaqαI site. With the one-directional approach, about 22.6% of random VIS generated VIS DNA of <500 bp. With the bidirectional approach, 40.3% could generate VIS DNA of <500 bp in either one of two directions (upstream or downstream from the vector).
FIG. 2.
FIG. 2.
Bidirectional VIS analysis. (a) Schematic diagram outlining the major steps of the two approaches (experiment 1 and experiment 2) for the bidirectional VIS amplification. See Materials and Methods for a detailed description of the experimental procedures. Oblique arrows indicate nuclease digestion of single-stranded DNA. Bent arrows indicate TaqαI digestion. Ovals and stars (*) represent streptavidin beads and 5′ end biotins, respectively. DNA linkers for experiment 2 are denoted by double purple lines. (b) QIAEXEL capillary electrophoresis analysis. Bidirectional PCR analysis of acutely infected CD34+ cells (Acute Infn) and peripheral blood repopulating cells collected at 5, 50, 88, and 108 months (mos) posttransplant. Left junctions and right junctions were concurrently analyzed by either INV-PCR or LM-PCR during experiment 1 or experiment 2. A 100-bp marker (Mkr) is included in each panel. Arrowheads indicate DNA bands resulting from LTR circles after acute infection.
FIG. 3.
FIG. 3.
Sequence frequency distribution for different lengths of VIS DNAs. The length of VIS DNA was calculated based on the distance of the nearest TaqαI site from the junction added to 50 bp of vector DNA. Individual VIS were distributed primarily from 25 to 450 bp.
FIG. 4.
FIG. 4.
Sequence frequency analysis for unique VIS of <450 bp. (a) Both the amount of sample DNA and the number of available VIS sequences influence detection of unique VIS. The number of unique VIS and the number of VIS sequences from the left (red) and the right (blue) junctions generated by experiment 1 or experiment 2 as well as the amount of genomic DNA used in the analysis were displayed at each time point. (b) Reproducibility of the assay. The relative frequencies (percentage of total sequences for VIS DNA of <500 bp) for individual sites from the two experiments were plotted on logarithmic scales. Reproducibility was tested using a Pearson product-moment correlation coefficient (r). Green diagonal lines indicate complete frequency match (r = 1). Higher correlation was observed when >5 μg of genomic DNA and >2,000 VIS sequences were used for the analysis (PBC from 5 months and 108 months, respectively).
FIG. 5.
FIG. 5.
Determination of relative frequencies for individual clones. (a) Calculation of clonal frequencies. The relative frequencies of VIS DNA of <500 bp from the left (x) and the right (y) junction were combined as described and represented as quantifiable vector integrants (QVIs). (b) The relation between the fraction of QVI in total vector (vertical axis) and expected overestimation (n-fold) of individual QVI frequency (horizontal axis). When 40.3% of vector integrants are QVIs, individual QVI frequencies are 2.56-fold overestimated. (c) Clonal frequency changes at four time points. The adjusted frequencies for individual QVIs were displayed according to the following color scheme: white to black to red, representing 0% to 0.1% to 3.1%. The frequency change of the top 15 highest-frequency QVIs was magnified on the right side. Ten of them were unambiguously mapped onto the rhesus genome (genomic locations are indicated on the right). Among those, seven were further tested with clone-specific real-time PCR. ND, nondetermined. (d) Clone-specific real-time PCR. The relative frequencies of seven QVIs were confirmed by clone-specific real-time PCR. Dark bars denote the adjusted relative frequencies determined by quantitative VIS sequencing. The values obtained by clone-specific real-time PCR (percentage of total vector copies) at the right junction and at the left junction of the vector integrants are shown with yellow and green bars, respectively.

Similar articles

See all similar articles

Cited by 11 articles

See all "Cited by" articles

Publication types

Feedback