WO2011020247A1

WO2011020247A1 - Replica barcode selection assay

Info

Publication number: WO2011020247A1
Application number: PCT/CN2009/073398
Authority: WO
Inventors: T. Lahn Bruce
Original assignee: Cyagen Biosciences (Guangzhou) Inc.
Priority date: 2009-08-21
Filing date: 2009-08-21
Publication date: 2011-02-24
Also published as: CN102203281B; CN102203281A

Abstract

The RBS assay is intended for isolating cells that may be predisposed to have a certain response to a treatment regime, but the assay accomplishes this without actually subjecting the isolated cells to that treatment. The assay relies on the assumption that, if a cell has a predisposition to show a certain response to a treatment due to its intrinsic genetic or epigenetic makeup, then this predisposition is likely inherited by its daughter cells after this cell divides. The RBS assay follows the following scheme: First, a unique genetic identifier, or "barcode", is inserted into each cell of the starting population. Next, these cells are allowed to proliferate such that each uniquely barcoded cell is amplified into multiple daughter cells bearing the same barcode. Cells are then split into two pools: a treatment pool and a reserve pool. The treatment pool is subject to the treatment to identify cells showing a positive response. Barcodes from these positive cells are then read. Finally, for each barcode identified this way, cells in the reserve pool bearing this same barcode are retrieved. These cells should be the sisters of the positive cells in the treatment pool by virtue of having the same barcode. Cells retrieved in this manner can be studied prospectively to address whether they have a predisposition to respond to the treatment, and if so, the genetic or epigenetic basis thereof.

Description

REPLICA BARCODE SELECTION ASSAY

BACKGROUND

[0001] The replica barcode selection (RBS) assay is designed to isolate cells that may be predisposed to have a certain response to a treatment regime, but do so without actually subjecting the cells to the treatment.

[0002] J_{n ce}u biology research, a common practice involves subjecting a population of cells to a treatment regime in order to select for the small subset of cells that show a particular response. The nature of the selection can vary widely from study to study.

Sometimes, it is based on the ability of a subset of cells to survive the treatment. An example is treating cultured cancer cells with chemotherapeutic drugs to select for resistant cells as a way of studying chemo resistance and cancer relapse. Another example is transplanting tumor cells into live hosts to obtain secondary tumors stemming from a small number of transplanted cells that have survived and proliferated. Alternatively, the selection scheme is not based on viability but rather on the ability of a subset of cells to acquire a distinct phenotype. For example, it was recently discovered that the introduction of several defined genes into somatic cells can reprogram a subset of the cells into the so- called induced pluripotent stem cells (iPSCs) that closely resemble embryonic stem cells (ESCs). As another example, studies of lineage differentiation frequently involve treating stem cells with conditions that induce a subset of them to differentiate into specific lineages.

[0003] I_n virtually all cases, it is of interest to understand why only a subset of cells

should display the desired response to the treatment. In particular, it is important to discriminate between two possibilities. One is that all the cells have the same inherent potential to show the response, but due to stochastic reasons such as the asynchronous cell cycle phase during treatment or the unevenness of drug exposure, only a fraction of cells end up displaying the response. The other possibility is that from the start, the population of treated cells is heterogeneous in their inherent potential to respond to the treatment due to genetic or epigenetic differences, resulting in only a subset of cells showing the response. In the latter situation, it would be of great interest to study what

genetic/epigenetic differences separate the responding cells from the non-responding cells. [0004] Take the example of the reprogramming of fibroblasts into iPSCs by defined genes.

The fact that only a small fraction of fibroblasts transduced with the genes would be reprogrammed into iPSCs could be due entirely to stochastic aspects of the experiment that vary from cell to cell, such as the point of cell cycle during transgene introduction, the copy number of transgenes, and the transgene integration sites. Alternatively, it could be due, at least in part, to the fact that some fibroblasts are inherently more predisposed than others to undergo reprogramming. If the latter is true, understanding the molecular basis of this predisposition could shed important light on the reprogramming process and might also lead to improve reprogramming protocols.

[0005] Also take the example of secondary tumor formation by transplantation of tumor cells. The fact that only a small fraction of transplanted cells give rise to secondary tumors could be due to stochasticity of the process or some cells in the original population having greater tumorigenic potential (i.e., the cancer stem cell hypothesis). If the latter is correct, then it is important to understand how the subset of cells with greater tumorigenic potential might differ from the rest of the cells.

[0006] Xo differentiate between stochasticity and predisposition, and to further understand the mechanism of any predisposition that might exist, it is useful to isolate the subset of cells that can display the desired response to the treatment regime. However, one may wish to do so prior to subjecting the cells to the treatment, given that the treatment is likely to profoundly alter the original properties of the cells.

L0007J Herein lies a paradox: the treatment has untoward effects on cells and therefore not desirable, but without the treatment, how would one know which cells would show the response? This is analogous to the "peanut farmer's dilemma" - a peanut farmer in China wishes to select the tastiest peanuts from his harvest to seed next year's crop, but the very act of tasting the peanuts renders them unusable as seeds. The RBS assay is designed to circumvent the paradox.

SUMMARY

The RBS assay is intended for isolating cells that may be predisposed to having a certain response to a treatment regime, but the assay accomplishes this without actually subjecting the isolated cells to that treatment. It relies on a key assumption: if a cell has a predisposition to have a certain response to a treatment due to its intrinsic genetic or epigenetic properties, then this predisposition is likely inherited by its daughter cells when this cell divides. Although this is not necessarily true under all circumstances, it is a reasonable assumption in many situations given our current knowledge of how genetic/epigenetic mechanisms influence cellular phenotype.

[0009] fhe general scheme of the RBS assay follows the steps depicted in Figure 1. In

Step 1 , a starting population of cells is modified in a manner that accords each cell with a distinct genetic identifier, or "barcode". In Step 2, these cells are allowed to proliferate such that each cell with its unique barcode is amplified into multiple daughter cells carrying that same barcode. In Step 3, cells are split into two pools: a treatment pool and a reserve pool. In Step 4, the treatment pool is subject to the treatment regime in order to identify positive cells - i.e., cells showing the desired response to the treatment. In Step 5, barcodes from these positive cells are read. Finally, in Step 6, cells in the reserve pool bearing the same barcodes are retrieved in a manner that does not involve significant alterations of the cells. These cells should be the sisters of the positive cells in the treatment pool, namely, they have descended from the same parental cell in the starting population.

[0010] The actual implementation of the RBS assay involves the delivery of two

categories of vectors: one can be called the barcode vector and the other the knockdown vector (details of these vectors are described in DETAILED DESCRIPTION). The barcode vector exists as a library whereby individual copies of the vector within this library share the same sequence except for a barcode region - a short stretch of sequence that is randomized such that the chance of two copies of the vector having the same sequence within this barcode region is exceedingly small. The barcoding of cells is achieved by introducing the vector library into the cells. The barcode region is embedded within a selectable reporter gene. After vector introduction, different cells should be distinct from each other by virtue of possessing different barcodes. These cells are divided into a treatment pool and a reserve pools, with the treatment pool subjected to the treatment.

[0011] Cells in the treatment pool that respond positively to the treatment can have their barcodes retrieved by PCR- amplification across the barcode region followed by sequencing of PCR product. For each barcode identified this way, a knockdown vector expressing a small hairpin RNA (shRNA) targeting this specific barcode can be constructed. This knockdown vector can be introduced into the reserve pool. In cells containing the barcode that is targeted by the shRNA, expression of the selectable reporter (within which the barcode resides) would be suppressed. This in turn is exploited to isolate these cells. For example, the selectable reporter could be a visual one such as the green fluorescent protein (GFP). In this case, a reduction in fluorescence from the knockdown can be used to select these cells by fluorescence-activate cell sorting (FACS). The selectable reporter could also be a negative drug- selectable gene such as thymidine kinase (TK), which metabolizes the drug ganciclovir into a cytotoxic substance. In this case, cells experiencing a knockdown of TK expression can be selected by virtue of their increased viability under ganciclovir treatment.

BRIEF DESCRIPTION OF THE DRAWINGS

[0012] Figure 1. Schema of the replica barcode selection assay. The 6 major steps of the assay are described in detail in the text. Each grey oval represents a cell while the black bar within it represents the cell's genome. The colored segment within the black bar represents the barcode sequence embedded in the genome, with different colors denoting different barcodes. The number of cells or barcodes is only figurative.

[0013] Figure 2. The 3-component plasmid system for the rapid construction of lentiviral gene expression vectors. The 3 types of plasmids, pPromoter, pReporter and pDestination are depicted. They can recombine using the Gateway recombination technology into a single vector, pFinal (recombination events are indicated by the crisscrosses). Recombination occurs at att sites, which are depicted. Only att sites of the same color can recombine with each other. Features are not drawn precisely to scale.

[0014] Figure 3. Diagram of the barcode vector and the shRNA target region. (A) Key features of the barcode vector. Two potential insertion sites for the shRNA target region are indicated. LTR: long terminal repeat, which marks the ends of the lentiviral genome. Features are not drawn precisely to scale. (B) Sequence of the shRNA target region. "N" denotes random sequence. Nucleotide sequences are drawn in different colors to reflect different features.

[0015] Figure 4. Procedure for producing a library of the barcode vector. The 4 major steps of the procedure are detailed in the text. Nucleotide sequences are drawn in different colors to reflect different features.

[0016] Figure 5. Diagram of the knockdown vector. Features are not drawn precisely to scale. (A) Repression of shRNA expression in the absence of doxycycline. (B) Derepression of shRNA expression by doxycycline.

[0017] Figure 6. Design of shRNA sequence. (A) Forward and reverse shRNA oligos.

They are annealed with each other to be ligated into the empty knockdown vector. (B) Hairpin structure of the shRNA product expressed from knockdown vector.

DETAILED DESCRIPTION

[0018] As described in SUMMARY, two categories of vectors are used in the RBS assay.

One is the barcode vector, which is intended to barcode cells. The other is the knockdown vector, which is used to express the shRNA that targets a barcode of interest in order to achieve knockdown of the selectable reporter containing that barcode. The vectors can be delivered into cells by a number of methods. One example is lentivirus-based gene delivery, which is described here.

[0019] Barcode vector

L0020J A _map _of Jj₁₆ barcode vector is shown in Figure 3 A. It is modified from the pFinal lentiviral vector depicted in Figure 2 by inserting a shRNA target region (which contains the barcode) into either the 5'-UTR or the 3'-UTR of the selectable reporter gene (see details below). The shRNA target sequence is commonly set at 19 bases, but other sequence lengths can be accommodated too.

[0021] As shown in Figure 3B, the barcode sequence within the 19-mer shRNA target region is 10 bases long. A 10-mer barcode region would in theory have a coding capacity of 4¹⁰, or about one million different barcodes. Other barcode sequence lengths can be used too. An example of an actual sequence of the target region is shown in Figure 3B.

LUUzzJ ^rp^_{e se}j_ectai₃2_{e re}p_Orter gene can be a visual marker (an example is EmGFP, which is a brighter variant of conventional GFP) or a negative drug-selectable marker (an example is deltaTK, which is an improved version of conventional TK). The reporter can be driven by a promoter appropriate for the cells under investigation (an example is the human EFlA promoter, which is a commonly used strong constitutive promoters).

[00231 The pFinal plasmid is modified to allow the insertion of the shRNA target region

(potential insertion sites are indicated in Figure 3A). Two unique restriction sites (for example, BamHI and Sail) are engineered into either the 5'-UTR or the 3'-UTR of the selectable reporter. The plasmid containing these restriction sites, which can be referred to as the empty barcode vector, could allow the shRNA target region to be ligated between the two restriction sites.

[0024] xhe shRNA target region, which contains the random barcode sequences, is made and inserted into the empty barcode vector following the steps depicted in Figure 4. In Step 1 , three oligonucleotides are synthesized to be used in a PCR reaction. One of them, named the barcode oligo, serves as template in the PCR, while the other two serve as primers. The barcode oligo contains the shRNA target region, flanked by the two aforementioned restriction sites (BamHI and Sail), and further flanked by extra sequences to provide sufficiently long priming sites for the PCR primers. Crucially, the target region contains the barcode at its core that is composed of completely random sequence. This random portion of the barcode oligo can be generated during the synthesis of the oligo simply by supplying all 4 nucleotides when synthesizing each of these bases. In Step 2, the pool of single-stranded barcode oligo fragments is converted into a pool of double- stranded molecules by PCR amplification. In Step 3, the PCR product is digested with BamHI and Sail, and treated with phosphatase to remove overhanging phosphates (this prevents their concatamerization in the subsequent ligation step), and purified by the polyacrylamide gel electrophoresis (PAGE) method to isolate the large internal fragment. In Step 4, this internal fragment, which contains the random barcode, is ligated into the empty barcode vector that has also been cut with BamHI and Sail. Then the ligation product is then transformed into E.coli to create a library of the final barcode vector, with every clone in the library bearing a distinct barcode.

[0025] Knockdown vector

[0026] The knockdown vector can be constructed by modifying an inducible shRNA

lentiviral vector. As shown in Figure 5, the vector has a TRE/U6 promoter that drives the expression of the shRNA. The TRE/U6 promoter is normally silenced by the

transcriptional repressor tTS expressed from the same vector driven by the constitutive CMV promote (Figure 5A). However, in the presence of doxycycline, tTS binds to doxycycline, and the resulting conformational change makes it unable to repress the TRE- U6 promoter, which in turn leads to the expression of the shRNA (Figure 5B). [0027] The empty version of the knockdown vector (i.e., without the shRNA sequence inserted) contains two unique restrictions sites (for example, Xhol and Hindlll), into which the shRNA sequence can be ligated. The final knockdown vector is constructed following steps depicted in Figure 6. First, two oligonucleotides (hereon referred to as the "shRNA oligos") are designed based on the sequence of the shRNA target region including the barcode sequence of interest, one for the forward strand and the other for the reverse strand (Figure 6A). These two oligos are complementary to each other and contains segments corresponding to the sense and antisense sequences of the 19-mer target region separated by a 9-base hairpin loop. Once annealed, the double- stranded DNA contains single-stranded overhangs at the left and right ends that are compatible with the overhangs produced by Xhol and Hindlll digestion, respectively. After the two shRNA oligos are annealed, they are ligated into the empty knockdown vector that has also been cut by Xhol and Hindlll, followed by transformation into E.coli and confirmation of clones by sequencing. This generates the final knockdown vector depicted in Figure 5. Expression of the shRNA sequence from this vector should produce the desired shRNA (Figure 6B).

[0028] Procedure for the RBS assay

[0029] The procedure for the RBS assay is described below by way of an example. When mammalian cells are treated with the drug 6-thioguanine (6-TG), only a small fraction of cells that carry loss-of-function mutations in the X-linked Hprt gene would survive. The RBS assay should be able to retrieve these predisposing, Hprt-deficient cells, and do so without actually subjecting them to 6-TG treatment.

ΓQQ3QI Cultured mouse fibroblasts are transduced with the barcode lentivirus library and selected for transduced cells by hygromycin treatment (the lentivirus carries hygromycin resistance gene). The viral titer is adjusted such that the great majority of infected cells would have taken in only one viral particle per cell. This is important for two reasons. First, the selectable reporter carried on the lentiviral vector would be expressed more evenly from cell to cell if most cells carry only one copy of the viral vector. This would enhance the ability to find cells with knockdown of reporter expression in subsequent experiments. Second, if a cell is infected by multiple viral particles, the cell would end up bearing more than one barcode, which would confound subsequent analysis. It is worth noting that although each lentiviral particle packages two copies the RNA-based viral genome and that these two copies may have different barcodes, only one of the two copies integrate into the host genome upon infection.

L003 IJ Transduced cells are propagated further before divided into a treatment pool and a reserve pool. The treatment pool is treated with 6-TG to select for resistant cells. A few dozen surviving clones can be individually picked and expanded, and the barcode of each clone can be retrieved by PCR amplification across the barcode region followed by sequencing of PCR product.

[0032] Among the barcodes identified in this manner, a set of corresponding knockdown lentiviral vectors can be created to target them. The reserve pool can be expanded and divided further into several subpools. Each subpool is superinfected with one of the knockdown viral species and the cells are treated with doxycycline to turn on shRNA expression. These cells then can be selected for those that display reduced levels of reporter expression. These would contain true knockdown cells in which knockdown of reporter expression has occurred as well as "contaminating" non-knockdown cells that happen to express the reporter at low levels for reasons unrelated to the shRNA knockdown. The ability to inducibly turn on and off shRNA expression by doxycycline affords a convenient tool to remove some of the contaminating cells. When the EmGFP is the selectable reporter used in the barcode vector, after superinfection and treatment with doxycycline to turn on shRNA expression, cells with low levels of EmGFP are isolated by FACS, which should include knockdown cells as well as non-knockdown cells that happen to have low EmGFP expression. Doxycycline is then removed from the culture, which should lead to the repression of shRNA expression (see Figure 5). As a result, expression of EmGFP in knockdown cells would revert to the normal level, resulting in increased fluorescence, whereas the low level of EmGFP expression in non-knockdown cells would remain low. Cells with strong EmGFP expression after removal of doxycycline can be sorted to enrich for true knockdown cells.

LUUJ JJ Cells that come through the above selection scheme are grown at clonal density. A few dozen clones are individually picked and expanded, and their barcodes retrieved. When the RBS assay is working as designed, there should be a significant enrichment for clones with the targeted barcode. The clones containing the targeted barcode are subjected to 6-TG treatment. They should be resistant because, by virtue of having the targeted barcode, they should have the same resistance-causing mutations as the resistant cells in the treatment pool that also bear that barcode.

Claims

CLAIM:

1. A method consisting of the "replica barcode selection" (RBS) assay that is

intended to isolate cells that may be predisposed to have a certain response to a treatment regime, but do so without actually subjecting the cells to the treatment, the method comprising:

(a) modifying the starting population of cells by a manner that accords each cell with a unique genetic identifier, or "barcode".

(b) proliferating these cells such that each cell carrying its unique barcode is amplified into multiple daughter cells carrying that same barcode.

(c) splitting cells into two pools: a treatment pool and a reserve pool.

(d) subjecting the treatment pool to the treatment regime to identify positive cells - i.e., cells displaying the desired response.

(e) reading the barcodes of the positive cells.

(f) retrieving cells in the reserve pool bearing the same barcodes.

2. The method of claim 1, wherein the barcoding of cells is achieved by introducing DNA from a barcode vector library into cells such that each cell contains a unique barcode.

3. The method of claim 2, wherein the barcode vector exists as a library whereby individual copies of the vector within this library share the same sequence except for a randomized barcode region.

4. The method of claim 1, wherein barcodes from the positive cells in the treatment pool are read and have their barcodes retrieved by PCR- amplification across the barcode region followed by sequencing of PCR product.

5. The method of claim 1, wherein a knockdown vector expressing a small hairpin RNA (shRNA) targeting the specific barcode is introduced into the reserve pool to retrieve cells with that barcode by virtue of knocking down the expression of the selectable reporter gene (which contains the shRNA target region) present in the barcode vector.