The Journal of Molecular Diagnostics, Vol. 18, No. 4, July 2016
jmd.amjpathol.org
Hybridization-Induced Aggregation Technology for Practical Clinical Testing KRAS Mutation Detection in Lung and Colorectal Tumors Hillary S. Sloane,* James P. Landers,*yz and Kimberly A. Kellyx{ From the Departments of Chemistry,* Pathology,y Mechanical Engineering,z and Biomedical Engineering,x and the Robert M. Berne Cardiovascular Research Center,{ School of Medicine, University of Virginia, Charlottesville, Virginia Accepted for publication February 18, 2016. Address correspondence to Kimberly A. Kelly, Ph.D., Department of Biomedical Engineering, University of Virginia, Box 800759 Health System, Charlottesville, VA 22908. E-mail: kak3x@ virginia.edu.
KRAS mutations have emerged as powerful predictors of response to targeted therapies in the treatment of lung and colorectal cancers; thus, prospective KRAS genotyping is essential for appropriate treatment stratification. Conventional mutation testing technologies are not ideal for routine clinical screening, as they often involve complex, time-consuming processes and/or costly instrumentation. In response, we recently introduced a unique analytical strategy for revealing KRAS mutations, based on the allele-specific hybridization-induced aggregation (HIA) of oligonucleotide probe-conjugated microbeads. Using simple, inexpensive instrumentation, this approach allows for the detection of any common KRAS mutation in T, p.G12S), H2122 (nonesmall cell adenocarcinoma; c.34G>T, p.G12C), CaCo2 (colorectal carcinoma; wild-type), and H1975 (lung adenocarcinoma; wild-type), were purchased from the ATCC (Manassas, VA). cDNA was synthesized from cultured cells using the FastLane Cell cDNA Kit (Qiagen, Venlo, the Netherlands).
Tumor Samples Twenty frozen surgical resections (10 lung and 10 colon) were obtained from the Department of Pathology through Table 1
Oligonucleotide Sequences
Oligonucleotide PCR primers HIAMD forward HIAMD reverse Sequencing forward Sequencing reverse Targets KRAS wild type
KRAS mutant (c.34G>T)
Probes Discriminating Stabilizing
548
Sequence 50 -GACTGAATATAAACTTGTGGTAGTTGGA-30 50 -CATATTCGTCCACAAAATGATTCTG-30 50 -GAGAATTCATGACTGAATATAAACTTGT-30 50 -TCGAATTCCTCTATTGTTGGATCATATTCG-30 50 -GACTGAATATAAACTTGTGGTAGTTGGAGCTGGTGGCGTAGGCAAGAGTGCCTTGACGATACAGCTAATTCAGAATCATTTTGTGGACGAATATG-30 0 5 -GACTGAATATAAACTTGTGGTAGTTGGAGCTTGTGGCGTAGGCAAGAGTGCCTTGACGATACAGCTAATTCAGAATCATTTTGTGGACGAATATG-30 50 -CTACGCCTCCAGCTCTTTTTT[BiotinTEGwQ]-30 0 5 -[Biotin-TEG]TTTTTTCTGAATTAGCTGTATCGTCAAGGCACTC-30
Sequencing PCR reactions were composed of 1X MyTaq Reaction Buffer (Bioline Reagents Ltd., London, UK), 0.4 mmol/L primers (Table 1), 0.05 U/mL MyTaq HS DNA Polymerase (Bioline Reagents Ltd), and 10%/v template cDNA. A GeneAmp PCR System 2700 thermocycler (Applied Biosystems, Foster City, CA) was used with the following thermocycling conditions: 2 minutes at 95 C; followed by 45 cycles of 15 seconds at 95 C, 30 seconds at 57 C, and 30 seconds at 72 C; then finally 2 minutes at 72 C. PCR products were analyzed using Agilent 2100 DNA 1000 Series II kits and instrumentation (Agilent Technologies, Santa Clara, CA) to confirm amplification and estimate amplicon concentration. The products were then purified using the QIAquick PCR Purification Kit (Qiagen). Sequencing was performed by Eurofins Genomic (Huntsville, AL) using the reverse-sequencing PCR primer.
PCR for Generation of HIAMD Targets PCR reactions were composed of 1X MyTaq Reaction Buffer (Bioline Reagents Ltd), 0.4 mmol/L primers (sequences described previously26) (Table 1), 0.05 U/mL MyTaq HS DNA Polymerase (Bioline Reagents Ltd), and 10%/v template cDNA. A GeneAmp PCR System 2700 thermocycler (Applied Biosystems) was used with the following thermocycling conditions: 3 minutes at 95 C; followed by 50 cycles of 15 seconds at 95 C, 15 seconds at 58 C, and 5 seconds at 72 C; then finally 2 minutes at 72 C. The PCR products were analyzed using Agilent 2100 DNA 1000 Series II kits and instrumentation (Agilent Technologies) to confirm amplification. Before HIAMD analysis, the PCR product was denatured at 95 C for 2 minutes and snapcooled on ice.
jmd.amjpathol.org
-
The Journal of Molecular Diagnostics
KRAS Genotyping via Bead Aggregation Table 2
Patients’ Tumor Samples
Patient no.
Anatomic site
Pathologic diagnosis
Tumor differentiation
% Tumor by cellularity
Sequencing result
HIAMD result
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Colon Lung Colon Colon Lung Lung Lung Colon Colon Colon Colon Lung Lung Lung Colon Lung Colon Lung Colon Lung
Adenocarcinoma Adenocarcinoma Adenocarcinoma Adenocarcinoma Squamous cell carcinoma Adenocarcinoma Squamous cell carcinoma Adenocarcinoma Adenocarcinoma Adenocarcinoma Adenocarcinoma Squamous cell carcinoma Adenocarcinoma Adenocarcinoma Adenocarcinoma Adenocarcinoma Adenocarcinoma Adenocarcinoma Adenocarcinoma Adenocarcinoma
Moderate Well Well Poor Moderate Moderate Poor Poor Poor Moderate Moderate Moderate Moderate Moderate Moderate Moderate Moderate Poor Poor Poor
60 90 40 80 70 90 90 80 80 80 70 90 90 90 70 95 95 85 80 70
wt wt wt wt wt wt wt wt wt wt wt wt wt mut mut mut mut mut mut mut
wt wt wt wt wt wt wt wt wt wt wt wt wt mut mut mut mut mut mut mut
HIAMD Instrumentation and Mircowell Chip Instrumentation was developed in-house and described previously.27 Briefly, the setup consists of a vortex mixer (MS 3 Basic Vortex Mixer; IKA, Wilmington, NC) to hold the chip and provide gentle agitation, and a rotating magnet positioned above the chip to provide additional mixing of the probe-bound magnetic beads. Each chip (4 cm 4 cm 1.5 mm) was made of two layers of a plastic substrate [poly(methyl methacrylate)] (Astra Products, Baldwin, NY) and featured a 12-well circular array of 5-mm (diameter) circular wells. A detailed microdevice fabrication method was described previously.22
Bead Preparation Each set of probe-bound beads was prepared by immobilizing biotinylated probe oligonucleotides to Dynabeads MyOne Streptavidin C1 superparamagnetic beads (InvitroGen, Oslo, Norway). After conjugation and wash steps, the beads were brought up in 1X binding/washing buffer [5 mmol/L Tris-HCl (pH 7.5), 0.5 mmol/L EDTA, 1 mol/L NaCl] in a volume equivalent to the initial volume of stock beads, to maintain a concentration of approximately 7 to 10 109 beads/mL.
HIAMD Assay Each HIAMD reaction took place in a 5-mm (diameter) circular well in a total volume of 20 mL, composed of 10 mL of target sample (either PCR product or synthetic target sequence at a concentration of 1 1011 copies/mL, corresponding to approximately 50 ng of input DNA), 9 mL of hybridization buffer (50 mmol/L KCl, 2.5 mmol/L Tris, and
The Journal of Molecular Diagnostics
-
jmd.amjpathol.org
(34G>T) (34G>T) (34G>T) (35G>A) (34G>T) (35G>A) (34G>T)
75%/v formamide), and 1 mL of probe-bound beads (equal parts of the two probe-bound beads). (Note that lesser concentrations of target sample in the assay may produce unreliable results.22) The chip was placed on the HIAMD setup, using a rotating magnet speed of 2000 rpm and a vortexing speed of 130 rpm, for a total reaction time of 2 minutes. A digital image of each well was then obtained using a T1i Digital SLR camera with an MP-E 65-mm f/2.8 15 macro lens (Canon U.S.A., Inc., Lake Success, NY) and analyzed using a Kapur algorithm in Mathematica software version 8 (Wolfram, Champaign, IL), to derive a quantitative value (saturation) corresponding to the extent of aggregation. The extent of aggregation from each sample was normalized (as a percentage) to the aggregation from the synthetic wild-type sequence. The aggregation measured in a blank sample (the background) was subtracted from this normalized value to yield the final percentage-aggregation value.
Results KRAS Mutation Analysis in Lung and Colorectal Cancer Cell Lines To ensure the feasibility of using HIAMD technology for assessing KRAS mutation status in lung and colorectal cancers, we applied the method for the analysis of a panel of lung and colorectal cancer cell lines. For both lung and colorectal cancers, wild-type and mutant cell lines were tested. The percentage-aggregation values measured from KRAS mutant cell lines (SW-620 and H2122) were significantly less than the percentage aggregation from KRAS wild-type cell lines (CaCo2 and H1975, Figure 2). Both mutant cell lines displayed percentage-aggregation values
549
Sloane et al
Figure 2
KRAS mutation analysis of lung cancer and colorectal cancer cell lines. Using a PCR-derived target sequence, each cell line was analyzed using the HIAMD assay. The results indicate a significant decrease in the extent of aggregation produced by cell lines bearing a KRAS mutation (SW-620 and H2122) as compared to wild-type cell lines (CaCo2 and H1975). Data are expressed as means SD. n Z 3 per group.
near zero, indicating that the extent of bead aggregation induced by these samples was similar to that of blank (no DNA) samples; this finding provides evidence of the high selectivity of the assay. Importantly, the assay was useful for distinguishing mutations in both the SW-620 and H2122 cell lines, even though the mutations are present at different positions in the gene (c.35 and c.34, respectively).
Figure 3 Resolution of mutant (mut) DNA in a background of wild-type (WT) DNA. HIAMD was applied for the analysis of synthetic samples with mixed genotypes. All samples with mutant content (down to 25% mutant) were significantly distinguishable from a wild-type sample. Data are expressed as means SD. n Z 3 per group.
KRAS Mutation Analysis of Patients’ Samples
In cancer, the mutated cells are surrounded by stroma, which include cancer-associated fibroblasts, endothelium, and immune cellsdall of which harbor wild-type KRAS alleles. Therefore, we investigated the utility of HIAMD in detecting a KRAS mutation in a background of wild-type DNA. Synthetic mutant targets were diluted with synthetic wild-type targets to model mixed samples of the following compositions: 75% mutant, 50% mutant, and 25% mutant. These samples were assayed on the same multiwell chip along with 100% mutant, 0% mutant, and blank (no DNA) samples. All samples with mutant content were distinguishable from the wild-type sample, including the sample with only 25% mutant DNA (Figure 3). This detection limit is comparable to that of sequencing, which is reported to be in the range of 15% to 30%.17,19,28 In terms of tumor purity, if we consider a heterozygous genotype of the cancer (allele frequency of 0.5), the tumor purity would need to be at least 50% to have at least 25% mutant content. A threshold value of 36% aggregation was calculated based on the results of the 100% mutant sample, using the mean aggregation value minus three times the SD. Thus, any aggregation value T and two were 35G>A, representing 71.4% and 28.6%, respectively. Notably, all samples that appeared to have a KRAS mutation also contained wild-type DNA based on the sequencing electropherograms (Figure 4A). This finding was expected for patients’ tumor samples, and therefore, we highlight that our sample set, although limited in size, serves as a representative collection of typical clinical samples. HIAMD assays were performed, and percentage-aggregation values of all tumor DNA samples were calculated and plotted (Figure 4B). The x axis was set to cross at the set threshold value (36% aggregation) for ease of data interpretation. Samples with a wild-type KRAS genotype have aggregation values 36% and therefore lie above the x axis, whereas samples harboring a KRAS mutation result in aggregation values A
38
37
36
35
34
Patient 14 Genotype: 34G>T
70 65 60
Aggregation (%)
55 50 45 40 35
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
30 25 20 15 10
Figure 4
KRAS mutation analysis of patient tissue samples. A: Representative sequencing results of three patient samples. (Note: the reverse strand was sequenced). B: HIAMD results. Aggregation of 36% was used as the threshold value to distinguish between wild-type (36%) and mutant (T versus 35G>A).
Discussion Targeted therapies are a growing trend in basic and clinical cancer research, and for good reasondthe potential for improved treatment outcomes and cost-savings is tremendous. The effective implementation of a targeted therapeutic regimen requires a practical means of preemptive molecular characterization of the cancer. Here we have shown that HIAMD can be applied for KRAS mutation analysis of primary tumors, which is essential for predicting a patient’s sensitivity to epidermal growth factor receptoremediated therapies,29 as
The Journal of Molecular Diagnostics
-
jmd.amjpathol.org
well as population stratification in ongoing therapeutic development.14 Importantly, the analysis is performed in a manner that is both rapid and cost-effective. Although a larger study is warranted for validation, this work establishes the potential of applying the HIAMD technology for routine analysis of other, increasingly important genetic markers. The initial mutation analysis of cell lines described in this work shows proof-of-concept for KRAS mutation detection with the HIAMD technology in lung and colorectal cancers. However, the value of KRAS mutation analysis requires that the method be validated for use in patients’ tumor samples. Unlike samples from cell lines, tumor samples are complex and heterogeneous, and variably contaminated with nontumor content. Thus, the requirements of a mutation-testing method with respect to analytical performance are considerably more demanding and more difficult to achieve without great technical complexity. Ultimately, we show a 100% correlation between results derived from sequencing and HIAMD analysis for KRAS
551
Sloane et al mutation screening of 20 lung and colorectal tumors from patients. Furthermore, we show that a sample with only 25% KRAS mutant content can be detected in a background of wild-type DNA, which is consistent with the detection limit reported using the more costly and cumbersome sequencing method.19,28 Thus, we demonstrate that, using HIAMD technology, mutation detection can be performed rapidly and inexpensively without compromising the analytical performance. Undoubtedly, there are a number of technologies capable of revealing point mutations with higher sensitivity (eg, nextgeneration sequencing platforms30), which may be preferred for the analysis of very lowepurity samples. However, it must be emphasized that, although impressive, high sensitivity alone does not render a technology appropriate for routine clinical screening. Significant consideration must be given to the cost, complexity, and speed of the method. The inherent features of the HIAMD technology make it a natural fit for use in the routine diagnostic setting. Fundamentally, the method is simple. The instrumentation comprises nothing less than a commercial vortex mixer, a rotating magnet obtained from conventional laboratory equipment,27 a reusable plastic microwell chip, a camera, and a laptop computer. The simple instrumentation keeps the method very low in cost (approximately $2500 for all equipment, compared to approximately $100,000 for sequencing instrumentation18), and therefore decreases the financial burden on the patient and the health care provider. In addition, the simple analytical approach translates to simple technical operation, data analysis, and interpretation, therefore eliminating the need for highly skilled personnel. Furthermore, because the simple optical-analysis strategy requires only a camera and a laptop computer, the potential for using a cell phone for detection is high.31 This transition is an obvious next step and will only increase the suitability of this technology for the point of care. The HIAMD method uses samples processed according to current clinical protocols and is compatible with the clinical laboratory workflow for sample preparation, which includes nucleic acid extraction and PCR amplification. After amplification, HIAMD analysis for the detection of all common KRAS mutations (located in codons 12 and 13) is complete in
jmd.amjpathol.org
Hybridization-Induced Aggregation Technology for Practical Clinical Testing KRAS Mutation Detection in Lung and Colorectal Tumors Hillary S. Sloane,* James P. Landers,*yz and Kimberly A. Kellyx{ From the Departments of Chemistry,* Pathology,y Mechanical Engineering,z and Biomedical Engineering,x and the Robert M. Berne Cardiovascular Research Center,{ School of Medicine, University of Virginia, Charlottesville, Virginia Accepted for publication February 18, 2016. Address correspondence to Kimberly A. Kelly, Ph.D., Department of Biomedical Engineering, University of Virginia, Box 800759 Health System, Charlottesville, VA 22908. E-mail: kak3x@ virginia.edu.
KRAS mutations have emerged as powerful predictors of response to targeted therapies in the treatment of lung and colorectal cancers; thus, prospective KRAS genotyping is essential for appropriate treatment stratification. Conventional mutation testing technologies are not ideal for routine clinical screening, as they often involve complex, time-consuming processes and/or costly instrumentation. In response, we recently introduced a unique analytical strategy for revealing KRAS mutations, based on the allele-specific hybridization-induced aggregation (HIA) of oligonucleotide probe-conjugated microbeads. Using simple, inexpensive instrumentation, this approach allows for the detection of any common KRAS mutation in T, p.G12S), H2122 (nonesmall cell adenocarcinoma; c.34G>T, p.G12C), CaCo2 (colorectal carcinoma; wild-type), and H1975 (lung adenocarcinoma; wild-type), were purchased from the ATCC (Manassas, VA). cDNA was synthesized from cultured cells using the FastLane Cell cDNA Kit (Qiagen, Venlo, the Netherlands).
Tumor Samples Twenty frozen surgical resections (10 lung and 10 colon) were obtained from the Department of Pathology through Table 1
Oligonucleotide Sequences
Oligonucleotide PCR primers HIAMD forward HIAMD reverse Sequencing forward Sequencing reverse Targets KRAS wild type
KRAS mutant (c.34G>T)
Probes Discriminating Stabilizing
548
Sequence 50 -GACTGAATATAAACTTGTGGTAGTTGGA-30 50 -CATATTCGTCCACAAAATGATTCTG-30 50 -GAGAATTCATGACTGAATATAAACTTGT-30 50 -TCGAATTCCTCTATTGTTGGATCATATTCG-30 50 -GACTGAATATAAACTTGTGGTAGTTGGAGCTGGTGGCGTAGGCAAGAGTGCCTTGACGATACAGCTAATTCAGAATCATTTTGTGGACGAATATG-30 0 5 -GACTGAATATAAACTTGTGGTAGTTGGAGCTTGTGGCGTAGGCAAGAGTGCCTTGACGATACAGCTAATTCAGAATCATTTTGTGGACGAATATG-30 50 -CTACGCCTCCAGCTCTTTTTT[BiotinTEGwQ]-30 0 5 -[Biotin-TEG]TTTTTTCTGAATTAGCTGTATCGTCAAGGCACTC-30
Sequencing PCR reactions were composed of 1X MyTaq Reaction Buffer (Bioline Reagents Ltd., London, UK), 0.4 mmol/L primers (Table 1), 0.05 U/mL MyTaq HS DNA Polymerase (Bioline Reagents Ltd), and 10%/v template cDNA. A GeneAmp PCR System 2700 thermocycler (Applied Biosystems, Foster City, CA) was used with the following thermocycling conditions: 2 minutes at 95 C; followed by 45 cycles of 15 seconds at 95 C, 30 seconds at 57 C, and 30 seconds at 72 C; then finally 2 minutes at 72 C. PCR products were analyzed using Agilent 2100 DNA 1000 Series II kits and instrumentation (Agilent Technologies, Santa Clara, CA) to confirm amplification and estimate amplicon concentration. The products were then purified using the QIAquick PCR Purification Kit (Qiagen). Sequencing was performed by Eurofins Genomic (Huntsville, AL) using the reverse-sequencing PCR primer.
PCR for Generation of HIAMD Targets PCR reactions were composed of 1X MyTaq Reaction Buffer (Bioline Reagents Ltd), 0.4 mmol/L primers (sequences described previously26) (Table 1), 0.05 U/mL MyTaq HS DNA Polymerase (Bioline Reagents Ltd), and 10%/v template cDNA. A GeneAmp PCR System 2700 thermocycler (Applied Biosystems) was used with the following thermocycling conditions: 3 minutes at 95 C; followed by 50 cycles of 15 seconds at 95 C, 15 seconds at 58 C, and 5 seconds at 72 C; then finally 2 minutes at 72 C. The PCR products were analyzed using Agilent 2100 DNA 1000 Series II kits and instrumentation (Agilent Technologies) to confirm amplification. Before HIAMD analysis, the PCR product was denatured at 95 C for 2 minutes and snapcooled on ice.
jmd.amjpathol.org
-
The Journal of Molecular Diagnostics
KRAS Genotyping via Bead Aggregation Table 2
Patients’ Tumor Samples
Patient no.
Anatomic site
Pathologic diagnosis
Tumor differentiation
% Tumor by cellularity
Sequencing result
HIAMD result
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Colon Lung Colon Colon Lung Lung Lung Colon Colon Colon Colon Lung Lung Lung Colon Lung Colon Lung Colon Lung
Adenocarcinoma Adenocarcinoma Adenocarcinoma Adenocarcinoma Squamous cell carcinoma Adenocarcinoma Squamous cell carcinoma Adenocarcinoma Adenocarcinoma Adenocarcinoma Adenocarcinoma Squamous cell carcinoma Adenocarcinoma Adenocarcinoma Adenocarcinoma Adenocarcinoma Adenocarcinoma Adenocarcinoma Adenocarcinoma Adenocarcinoma
Moderate Well Well Poor Moderate Moderate Poor Poor Poor Moderate Moderate Moderate Moderate Moderate Moderate Moderate Moderate Poor Poor Poor
60 90 40 80 70 90 90 80 80 80 70 90 90 90 70 95 95 85 80 70
wt wt wt wt wt wt wt wt wt wt wt wt wt mut mut mut mut mut mut mut
wt wt wt wt wt wt wt wt wt wt wt wt wt mut mut mut mut mut mut mut
HIAMD Instrumentation and Mircowell Chip Instrumentation was developed in-house and described previously.27 Briefly, the setup consists of a vortex mixer (MS 3 Basic Vortex Mixer; IKA, Wilmington, NC) to hold the chip and provide gentle agitation, and a rotating magnet positioned above the chip to provide additional mixing of the probe-bound magnetic beads. Each chip (4 cm 4 cm 1.5 mm) was made of two layers of a plastic substrate [poly(methyl methacrylate)] (Astra Products, Baldwin, NY) and featured a 12-well circular array of 5-mm (diameter) circular wells. A detailed microdevice fabrication method was described previously.22
Bead Preparation Each set of probe-bound beads was prepared by immobilizing biotinylated probe oligonucleotides to Dynabeads MyOne Streptavidin C1 superparamagnetic beads (InvitroGen, Oslo, Norway). After conjugation and wash steps, the beads were brought up in 1X binding/washing buffer [5 mmol/L Tris-HCl (pH 7.5), 0.5 mmol/L EDTA, 1 mol/L NaCl] in a volume equivalent to the initial volume of stock beads, to maintain a concentration of approximately 7 to 10 109 beads/mL.
HIAMD Assay Each HIAMD reaction took place in a 5-mm (diameter) circular well in a total volume of 20 mL, composed of 10 mL of target sample (either PCR product or synthetic target sequence at a concentration of 1 1011 copies/mL, corresponding to approximately 50 ng of input DNA), 9 mL of hybridization buffer (50 mmol/L KCl, 2.5 mmol/L Tris, and
The Journal of Molecular Diagnostics
-
jmd.amjpathol.org
(34G>T) (34G>T) (34G>T) (35G>A) (34G>T) (35G>A) (34G>T)
75%/v formamide), and 1 mL of probe-bound beads (equal parts of the two probe-bound beads). (Note that lesser concentrations of target sample in the assay may produce unreliable results.22) The chip was placed on the HIAMD setup, using a rotating magnet speed of 2000 rpm and a vortexing speed of 130 rpm, for a total reaction time of 2 minutes. A digital image of each well was then obtained using a T1i Digital SLR camera with an MP-E 65-mm f/2.8 15 macro lens (Canon U.S.A., Inc., Lake Success, NY) and analyzed using a Kapur algorithm in Mathematica software version 8 (Wolfram, Champaign, IL), to derive a quantitative value (saturation) corresponding to the extent of aggregation. The extent of aggregation from each sample was normalized (as a percentage) to the aggregation from the synthetic wild-type sequence. The aggregation measured in a blank sample (the background) was subtracted from this normalized value to yield the final percentage-aggregation value.
Results KRAS Mutation Analysis in Lung and Colorectal Cancer Cell Lines To ensure the feasibility of using HIAMD technology for assessing KRAS mutation status in lung and colorectal cancers, we applied the method for the analysis of a panel of lung and colorectal cancer cell lines. For both lung and colorectal cancers, wild-type and mutant cell lines were tested. The percentage-aggregation values measured from KRAS mutant cell lines (SW-620 and H2122) were significantly less than the percentage aggregation from KRAS wild-type cell lines (CaCo2 and H1975, Figure 2). Both mutant cell lines displayed percentage-aggregation values
549
Sloane et al
Figure 2
KRAS mutation analysis of lung cancer and colorectal cancer cell lines. Using a PCR-derived target sequence, each cell line was analyzed using the HIAMD assay. The results indicate a significant decrease in the extent of aggregation produced by cell lines bearing a KRAS mutation (SW-620 and H2122) as compared to wild-type cell lines (CaCo2 and H1975). Data are expressed as means SD. n Z 3 per group.
near zero, indicating that the extent of bead aggregation induced by these samples was similar to that of blank (no DNA) samples; this finding provides evidence of the high selectivity of the assay. Importantly, the assay was useful for distinguishing mutations in both the SW-620 and H2122 cell lines, even though the mutations are present at different positions in the gene (c.35 and c.34, respectively).
Figure 3 Resolution of mutant (mut) DNA in a background of wild-type (WT) DNA. HIAMD was applied for the analysis of synthetic samples with mixed genotypes. All samples with mutant content (down to 25% mutant) were significantly distinguishable from a wild-type sample. Data are expressed as means SD. n Z 3 per group.
KRAS Mutation Analysis of Patients’ Samples
In cancer, the mutated cells are surrounded by stroma, which include cancer-associated fibroblasts, endothelium, and immune cellsdall of which harbor wild-type KRAS alleles. Therefore, we investigated the utility of HIAMD in detecting a KRAS mutation in a background of wild-type DNA. Synthetic mutant targets were diluted with synthetic wild-type targets to model mixed samples of the following compositions: 75% mutant, 50% mutant, and 25% mutant. These samples were assayed on the same multiwell chip along with 100% mutant, 0% mutant, and blank (no DNA) samples. All samples with mutant content were distinguishable from the wild-type sample, including the sample with only 25% mutant DNA (Figure 3). This detection limit is comparable to that of sequencing, which is reported to be in the range of 15% to 30%.17,19,28 In terms of tumor purity, if we consider a heterozygous genotype of the cancer (allele frequency of 0.5), the tumor purity would need to be at least 50% to have at least 25% mutant content. A threshold value of 36% aggregation was calculated based on the results of the 100% mutant sample, using the mean aggregation value minus three times the SD. Thus, any aggregation value T and two were 35G>A, representing 71.4% and 28.6%, respectively. Notably, all samples that appeared to have a KRAS mutation also contained wild-type DNA based on the sequencing electropherograms (Figure 4A). This finding was expected for patients’ tumor samples, and therefore, we highlight that our sample set, although limited in size, serves as a representative collection of typical clinical samples. HIAMD assays were performed, and percentage-aggregation values of all tumor DNA samples were calculated and plotted (Figure 4B). The x axis was set to cross at the set threshold value (36% aggregation) for ease of data interpretation. Samples with a wild-type KRAS genotype have aggregation values 36% and therefore lie above the x axis, whereas samples harboring a KRAS mutation result in aggregation values A
38
37
36
35
34
Patient 14 Genotype: 34G>T
70 65 60
Aggregation (%)
55 50 45 40 35
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
30 25 20 15 10
Figure 4
KRAS mutation analysis of patient tissue samples. A: Representative sequencing results of three patient samples. (Note: the reverse strand was sequenced). B: HIAMD results. Aggregation of 36% was used as the threshold value to distinguish between wild-type (36%) and mutant (T versus 35G>A).
Discussion Targeted therapies are a growing trend in basic and clinical cancer research, and for good reasondthe potential for improved treatment outcomes and cost-savings is tremendous. The effective implementation of a targeted therapeutic regimen requires a practical means of preemptive molecular characterization of the cancer. Here we have shown that HIAMD can be applied for KRAS mutation analysis of primary tumors, which is essential for predicting a patient’s sensitivity to epidermal growth factor receptoremediated therapies,29 as
The Journal of Molecular Diagnostics
-
jmd.amjpathol.org
well as population stratification in ongoing therapeutic development.14 Importantly, the analysis is performed in a manner that is both rapid and cost-effective. Although a larger study is warranted for validation, this work establishes the potential of applying the HIAMD technology for routine analysis of other, increasingly important genetic markers. The initial mutation analysis of cell lines described in this work shows proof-of-concept for KRAS mutation detection with the HIAMD technology in lung and colorectal cancers. However, the value of KRAS mutation analysis requires that the method be validated for use in patients’ tumor samples. Unlike samples from cell lines, tumor samples are complex and heterogeneous, and variably contaminated with nontumor content. Thus, the requirements of a mutation-testing method with respect to analytical performance are considerably more demanding and more difficult to achieve without great technical complexity. Ultimately, we show a 100% correlation between results derived from sequencing and HIAMD analysis for KRAS
551
Sloane et al mutation screening of 20 lung and colorectal tumors from patients. Furthermore, we show that a sample with only 25% KRAS mutant content can be detected in a background of wild-type DNA, which is consistent with the detection limit reported using the more costly and cumbersome sequencing method.19,28 Thus, we demonstrate that, using HIAMD technology, mutation detection can be performed rapidly and inexpensively without compromising the analytical performance. Undoubtedly, there are a number of technologies capable of revealing point mutations with higher sensitivity (eg, nextgeneration sequencing platforms30), which may be preferred for the analysis of very lowepurity samples. However, it must be emphasized that, although impressive, high sensitivity alone does not render a technology appropriate for routine clinical screening. Significant consideration must be given to the cost, complexity, and speed of the method. The inherent features of the HIAMD technology make it a natural fit for use in the routine diagnostic setting. Fundamentally, the method is simple. The instrumentation comprises nothing less than a commercial vortex mixer, a rotating magnet obtained from conventional laboratory equipment,27 a reusable plastic microwell chip, a camera, and a laptop computer. The simple instrumentation keeps the method very low in cost (approximately $2500 for all equipment, compared to approximately $100,000 for sequencing instrumentation18), and therefore decreases the financial burden on the patient and the health care provider. In addition, the simple analytical approach translates to simple technical operation, data analysis, and interpretation, therefore eliminating the need for highly skilled personnel. Furthermore, because the simple optical-analysis strategy requires only a camera and a laptop computer, the potential for using a cell phone for detection is high.31 This transition is an obvious next step and will only increase the suitability of this technology for the point of care. The HIAMD method uses samples processed according to current clinical protocols and is compatible with the clinical laboratory workflow for sample preparation, which includes nucleic acid extraction and PCR amplification. After amplification, HIAMD analysis for the detection of all common KRAS mutations (located in codons 12 and 13) is complete in
