Journal Publication

This article was published with modifications in Epidemiology 2002;13:256-254

“The findings and conclusions in this report are those of the author(s) and do not necessarily represent the views of the funding agency."

---

DNA Banking for Epidemiologic Studies: a Review of Current Practices

by
Steinberg, Karen¹, Beck, Jeanne², Nickerson, Deborah³, Garcia-Closas, Montserrat⁴, Gallagher, Margaret¹, Caggana, Michele⁵, Reid, Yvonne⁶, Cosentino, Mark⁷, Ji, Jay⁷, Johnson, Delene⁸, Hayes, Richard B⁴, Earley, Marie¹, Lorey, Fred⁹, Hannon, Harry¹, Khoury, Muin J¹⁰, Sampson, Eric¹.

From the:
¹National Center for Environmental Health, Centers for Disease Control and Prevention, Atlanta, GA.
²Coriell Institute for Medical Research, 403 Haddon Avenue, Camden, NJ.
³ University of Washington, Seattle, WA.
⁴Division of Cancer Epidemiology and Genetics, National Cancer Institute, National Institutes of Health, Bethesda, MD.
⁵Division of Genetic Disorders, Wadsworth Center,  New York State Department of Health, Albany, NY.
⁶Cell Biology Program, American Type Culture Collection, 10801 University Blvd., Manassas, VA.
⁷Biotech Research Laboratory, Boston Biomedical, Incorporated (BBI), Biotech Research Laboratory, 217 Perry Parkway, Gaithersburg, MD.
⁸Blood Centers of the Pacific, Research and Scientific Services, San Francisco, CA.
⁹Genetic Disease Branch, California Department of Health, Berkeley, CA.
¹⁰Office of Genetics and Disease Prevention, National Center for Environmental Health, Centers for Disease Control and Prevention, Atlanta, GA.

---

---

Running head:  DNA Banking for Epidemiologic Studies

Key words:  biological specimen bank, cryopreservation, DNA, lymphocyte transformarion, blood splots.

Document word count: 3,208

Correspondence to: Karen K. Steinberg, Molecular Biology Branch, MS F-24, Division of Environmental Health Laboratory Sciences, National Center for Environmental Health, Centers for Disease Control and Prevention, 4770 Buford Highway, N.E., Atlanta, GA 30341-3724, (770)488-4955(phone), (770)488-4005(fax), kks1@cdc.gov.

Abstract

To study genetic risk factors for common diseases, researchers have begun collecting DNA specimens in large epidemiologic studies and surveys.  However, little information is available to guide researchers in selecting the most appropriate specimens.  In an effort to gather the best information for the selection of specimens for these studies, we convened a meeting of scientists engaged in DNA-banking for large epidemiologic studies.   In this discussion,  we review the information presented at that meeting in the context of recent published information.  Factors to be considered in choosing the appropriate specimens for  epidemiologic studies include quality and quantity of DNA, convenience of collection and storage, cost, and the ability to accommodate future needs for genotyping.  We focus on four types of specimens that are stored in these banks: 1) whole blood preserved as dried blood spots, 2) whole blood from which genomic DNA is isolated, 3) immortalized lymphocytes from whole blood or separated lymphocytes, prepared immediately or subsequent to cryopreservation, and 4) buccal epithelial cells.   Each of the specimens discussed is useful for epidemiologic studies according to specific needs which we enumerate in our conclusions.

As the Human Genome Project provides the foundation for understanding the genetic basis of common disease,1 population-based genetic studies will provide the information needed for the practical application of genetic risk factors to public health practice.   To this end, researchers have begun collecting specimens for molecular analysis in epidemiologic studies and surveys (Table 1).2-15   Here we address factors to be considered in choosing appropriate specimens for epidemiologic studies, including convenience of collection and storage, quantity and quality of DNA, and the ability to accommodate future needs for genotyping.  We focus on four types of specimens that are stored in these banks: 1) whole blood preserved as dried blood spots, 2) whole blood from which genomic DNA is isolated, 3) whole blood from which lymphocytes are isolated and immediately transformed or cryopreserved for later EBV-transformation, and 4) buccal epithelial cells.

Blood Spots

During the past 10 - 15 years, dried blood spots  have been used in state newborn screening programs to identify an increasing number of disorders.17-20   As a result, dried blood spots from these programs comprise at least two population-based repositories.21,22 Further, blood spots from military personnel are stored to serve as biologic “dog tags” for identification purposes.22a

Blood spots are a stable, inexpensive source of DNA, useful for genotyping polymorphisms for association studies.24   Those collected in newborn screening programs can serve as samples from which to determine population gene frequencies.  However, use of these specimens in any way other than anonymously is problematic because the specimens may not have been collected with adequate informed consent to perform genetic studies.21,23  Blood spots can be collected without a phlebotomist and safely transported by regular mail.  Blood spots yield enough DNA to genotype multiple gene variants 25,26 (Table 2).

The New York State Department of Health (NYSDOH), has developed a method for simultaneously genotyping polymorphisms related to hereditary hemochromatosis, sickle cell disease, methylene tetrahydrofolate reductase (MTHFR) deficiency, and medium chain acyl CoA dehydrogenase (MCAD) deficiency. Genotyping is performed by  allele-specific oligonucleotide hybridization in a 96-well format using 1 mm punches that yield 0.8-FL of whole blood.   Up to 40 1-mm punches have been obtained from 50-FL  blood spots.    By combining  4 markers per spot, a minimum of 800 markers could be genotyped from one 250-FL aliquot of blood, that is five 50-FL  blood spots obtained in 1 specimen.25

To ensure the quality of blood spots for newborn screening, CDC evaluates the chromatographic properties of filter paper, the variation of blood volume among different lots of filter paper, and the effects of differences in hematocrit on blood volume.  Two brands of filter paper are approved by the Food and Drug Administration (FDA):  Schleicher and Schuell Grade 903 and Whatman BFC 180.   The NCCLS, formerly the National Committee for Clinical Laboratory Standards, has published guidelines about how blood should be collected on filter paper for newborn screening programs.36

Whole Blood from Which Genomic DNA is Purified

Many large epidemiologic studies with a genetic component include whole blood specimens for obtaining genomic DNA. This approach provides high quality DNA in amounts sufficient for current applications, including genome scans using single nucleotide polymorphisms (SNPs), microsatellite repeat polymorphisms, and polymorphism identification (using methods such as single strand conformational polymorphisms and restriction fragment length polymorhpisms), as well as for genotyping loci (using methods such as allele specific oligonucleotides or sequencing).

The National Heart, Lung, and Blood Institute (NHLBI) has published guidelines for obtaining specimens,37 and we have discussed some of these issues in a previous review.24   Blood is most often collected using ethylenediaminetetraacetic acid (EDTA), although anticoagulants including heparin and acid citrate dextrose (ACD) have also been used.  Cells can be stored in whole blood, either anticoagulated or as a clot, or in buffy coats.4,5,11-13,38-42

Buffy coats have been obtained in large epidemiologic studies for isolation of DNA and establishment of transformed cell lines.  However, obtaining consistent DNA yields from buffy coats requires careful technique and is time-consuming.  Because of ease of specimen handling and storage, DNA is often isolated from fresh whole blood or blood stored at -80EC in large epidemiologic studies.11-13 DNA can also be extracted from blood clots, and improved commercial methods now allow yields that are similar to those from anticoagulated whole blood (PurgeneTM Minneapolis, MN 55441).  The use of clotted blood allows the investigator to obtain serum for other analyses including environmental toxicants and, at the same time, obtain DNA from clots.

Polypropylene rather than glass containers should be used to store frozen blood, and blood should be divided into aliquots to prevent freeze-thaw cycles. Purification methods include use of 1) enzymes (including proteinase K and RNAse),43  2) organic solvents or organic solvents with enzymes,44 3) salt percipitation,45-47 and 4) resins or affinity gels which are also the basis for many commercial kits.48,49

Several large epidemiologic studies collect anticoagulated whole blood from which to extract DNA or for storage and later extraction (Table 1). The National Health and Nutrition Examination Survey that began in April 1999, (NHANES99), is a  continuous survey that collects specimens from approximately 5,000 people in 15 U.S. locations each year.13   Because of limited resources, lymphocytes are not being transformed as was done for NHANES III,24 but instead approximately 250 - 700 Fg of DNA per participant is isolated from two 10 mL tubes with EDTA following the Puregene (D-50K) kit protocol (Gentra Systems, Inc., Minneapolis, MN).

Transformed Fresh or Cryopreserved Lymphocytes

EBV-transformed lymphocytes provide an unlimited source of high-quality genomic DNA for genotyping large numbers of polymorphisms requiring microgram quantities of DNA as well as cells that may prove to be useful for functional studies.  Traditionally, blood specimens have been sent to a central laboratory within 48 hours of collection for separation and EBV-transformation of lymphocytes.  However, an appealing alternative would be to cryopreserve initially and transform lymphocytes later.  This option would allow the investigator to transform selected subsets of specimens for nested case-control studies, but it presupposes that lymphocytes can be cryopreserved for long durations, and then transformed.

Although two published articles have reported viability of B-lymphocytes after cryopreservation and long-term storage,50,51 we found only one published report of a systematic study of the effects of long-term cryopreservation on transformation success rates.52  In most cases, cells are cryopreserved in RPMI with 30 percent fetal bovine serum (FBS), DMSO is added to a final concentration of 6 - 10 percent, and cells are frozen in a programmable freezer and stored in liquid nitrogen (vapor or liquid phase).  Some scientists have suggested that cryopreserved cells may have a higher rate of transformation than fresh cells because of treatment with DMSO prior to EBV-exposure.51

Investigators conducting the Multi-Center AIDS Cohort Study (MACS) published data on the  viability and ability to EBV-transform specimens stored for up to 12 years with a 90 percent transformation success rate.50  The authors emphasized the importance of cell separation and storage within 6 hours of collection and use of  a controlled-rate freezer when cryopreserving cells - procedures which may not be possible in many epidemiological studies.

The Cell Biology Program of the American Type Culture Collection (ATCC, Manassas, VA) reported their experiences with a project of the National Cancer Institute during the past 10 years in which specimens were obtained for study from people who either have cancer or who are at high risk for cancer.52  A typical specimen comprised 20 - 30 mL of whole blood from which DNA is immediately extracted and lymphocytes separated on a density gradient for EBV-transformation (for which ATCC reports a 97.5% success rate). Lymphocytes have been successfully transformed from 250 - 500 FL of  fresh or cryopreserved whole blood, and all attempts to EBV-transform lymphocytes from 245 cryopreserved whole blood specimens which had 10% DMSO and frozen in a programmable freezer were successful. Whether the whole blood was fresh or cryopreserved, culture conditions included irradiated MRC-5 cells as a feeder layer with exposure to EBV. Based on this information, ATCC now cryopreserves 2 mL of whole blood for later EBV-transformation.

NCI's Cooperative Family Registry for Breast Cancer Studies (CFRBCS)11 includes cell lines and a DNA repository, some of which is maintained by the Coriell Institute for Medical Research (Camden, NJ).  Repository specimens include blood and buccal cells.  For each blood specimen, a tube of blood is stored at -80EC for direct DNA isolation; another is used for the isolation and cryopreservation of lymphocytes for future transformation or nucleic acid preparation; and a third tube is used for preparation of plasma and blood spots for identity checks.  When venipuncture is not possible, DNA is prepared from buccal cells obtained using mouthwash.  Cryopreserved cells are held in the liquid phase of liquid nitrogen54 as four aliquots of about 10 million cells each.

Attempts to transform freshly isolated lymphocytes from CFRBCS participants yields a success rate of 93 percent and for cryopreserved lymphocytes a 90 percent success rate.52 In addition, the data suggest that the length of time lymphocytes are held cryopreserved has little or no effect on the time to transformation, an observation noted also by Penno et al.55  Coriell receives specimens 1 to 6 days after collection with no significant difference in time required to transformation over the 6 days whether the lymphocytes are fresh or cryopreserved.  Successful lymphocyte transformations are obtained up to 10 days after collection in ACD.

The National Marrow Donor Program (NMDP, Minneapolis, MN) was established in 1986 to provide donors for bone marrow transplantation.56 The NMDP has established a repository, managed by Blood Centers of the Pacific (BCP, San Francisco, CA), comprising serum, lymphocytes, and EBV-transformed lymphocytes.  Blood is collected in ACD and is usually received less than 48 hours after collection, although international specimens can take up to 5 days to arrive.  BCP has transformed 95 percent of donor specimens on the first attempt which can be augmented with a second try.

The Biotech Research Laboratories of Boston Biomedical, Inc. (BBI, Rockville, MD) works in the area of infectious disease, particularly HIV and other retroviruses.  BBI processes specimens for NCI, NHLBI, and FDA repositories in three ways: Ficoll separation, nucleic acid extraction, and EBV-transformation of lymphocytes. BBI is involved in studies to determine optimal specimen handling for these repositories.

As an index of cell viability, BBI  measures lymphocyte apoptosis using Hoechst 33342 dye.57   To determine whether cryopreserved lymphocytes should be shipped in a dry shipper or on dry ice, particularly for situations in which samples are delayed for several days using this system, lymphocytes were isolated using Ficoll-density gradient centrifugation, frozen in a controlled-rate freezer with 7.5 percent DMSO, or placed into liquid nitrogen for two weeks.  Aliquots were then transferred to -70EC or left in liquid nitrogen for 4 days.  All aliquots were then placed into liquid nitrogen.  The rate of apoptosis for cells that had remained in the vapor phase of liquid nitrogen was 6.8 percent and for cells moved from liquid nitrogen to -70EC and back to liquid nitrogen was 54 percent.

Transformation rates were virtually identical in cells that were stored at -70EC and those stored in liquid nitrogen for up to 2 months, although lymphocytes stored at -70EC took longer to transform.   For the present, BBI recommends that specimens be shipped in dry shippers to maintain viability.

NCI reported on a cohort study within a program that evaluates strategies for early detection of cancer.   Blood specimens collected at each of six annual exams are used to search for early markers and etiologic factors for cancer.  Specimens include plasma, buffy coats, and whole blood (in ACD) for cryopreservation. Approximately 95 percent of the specimens arrive  within 24 hours, and virtually all within 48 hours.   After receipt by NCI, the whole blood is divided into 10 to 12 aliquots in 1.8-mL vials with 10 percent DMSO, frozen in a rate-controlled freezer, and placed into the vapor phase of liquid nitrogen.

In collaboration with ATCC, NCI is evaluating EBV-transformation after long-term storage of lymphocytes.  NCI finds that lymphocytes from blood collected with ACD maintain viability longer at room temperature than blood collected with heparin.  However, in all cases, viability drops by 3 days after collection, an experience similar to that of CDC.24

Despite the virtually unlimited supply of DNA furnished by transformed cell lines, this type of specimen has drawbacks, not the least of which is the high cost.  Another difficulty is the absence of serum or whole blood which could be used to measure compounds that may be important study variables.

Buccal Cells

Buccal cells can be obtained for DNA isolation using cytobrushes, swabs, or oral lavage.23,53,57 Buccal cells are being collected in several large epidemiologic studies as the primary specimen or as a supplement to whole blood specimens (Table 1).  Although there are few systematic studies that compare the various methods of collecting buccal cells in terms of their yield of human DNA (hDNA), a growing consensus suggests that the use of mouthwash gives a greater yield and higher quality hDNA (excluding bacterial contamination) than other methods of collection.32-35 The use of alcohol- containing mouthwash has been proposed as the optimal collection medium to prevent bacterial growth on swabs, due to the bacteriostatic properties of the alcohol. Alternatively, some have proposed the use of cytobrushes to exfoliate buccal cells, followed by expectoration of fluid which is spotted onto Guthrie cards treated with bactericidal and nuclease inhibiting compounds.59

One published study reported the quantity, quality, and stability of hDNA collected with mouthwash.35 Specimens yielded a median quantity of at least 32 Fg DNA (2 - 194) if specimens were held for up to 5 days, with yields declining to 21 Fg(5 - 56) at 30 days.  PCR success rates were greater than 94%, and high molecular weight DNA (>23 kb) was found in all but one of the 24 specimens.  Yields were greater when specimens were collected before brushing teeth.

Another recent publication reported hDNA yields in a comparison of cytobrush and mouthwash collection methods as well as DNA extraction methods.34 Median yields for cytobrushes were between 1 and 2 Fg compared with yields between 16 and 27 Fg for mouthwash.  Although PCR success rates were similar between the methods, mouthwash specimens were superior to cytobrushes for obtaining high molecular weight DNA.

ATCC is  developing methods for use of buccal cells as a cost-efficient, noninvasive source of DNA and is shifting its emphasis from DNA derived from whole blood and EBV-transformed lymphocytes to greater use of buccal cells.  Oral rinses are collected on filter-paper cards59 and mailed to a central laboratory.  DNA is then extracted and its quality evaluated by electrophoresis.   However, quantities of  DNA collected on these cards have not been sufficient for spectrophotometric detection and fragment sizes from only 536 base pairs (bp) to 989 bp have been amplified.

Methods such as the mouthwash method that require participant to expectorate are not an option for infants and small children. Because many studies report lower yields using cytobrushes, particularly in children and infants, methods for optimizing yield from buccal specimens obtained with cytobrushes are needed.  Primer extension preamplification (PEP) is one method for whole genome amplification (WGA).60,61  Zheng et al.62 reported improved DNA yields using PEP with WGA, which allowed about 900 PCR reactions per cytobrush with results verified with blood or marrow specimens.

To determine the quality of buccal cell DNA as well as the feasibility of whole-genome amplification, Coriell developed a protocol for obtaining buccal cells from registry participants who declined to have blood drawn.  After collection with mouthwash, DNA is prepared with QiagenTM kits, quantified with SYBR GreenTM, and characterized by pulse field electrophoresis.  DNA 20 to 40 kb in length was obtained and was suitable for PCR-based assays.  The DNA supported long PCR using a 5.6 kb segment of the calmodulin gene.

Because of the limited amount of DNA obtained from buccal specimens (0.2 - 6.0 Fg), Coriell used the PEP method of Zhang et al. 61 to determine whether the DNA could serve as a template for WGA. Using mouthwash DNA as a template, DNA can be increased 500- to 1,000-fold using PEP with sizes ranging from 500 to 4,000 bp.  However, a portion of the amplification is attributed to bacterial DNA. PEP appears to be suitable for PCR-based assays such as the amelogenin allelic discrimination assay, genotyping with di-, tri- and tetra-nucleotide markers, as well as detection of the tri-nucleotide repeats in the androgen receptor and polymorphisms in mitochondrial DNA. Although PEP is a method for whole genome amplification yielding sufficient DNA for extensive genetic analysis, it is not routinely done or routinely successful.  First, the fidelity of repeat sequences is problematic because telomere repeats appear to shorten, and amplification of trinucleotide repeats in Huntington's disease is difficult. Others have also noted the need for validation of nucleotide repeats.62 Second, the product size is limited which may preclude its use for long PCR.

Sources of variation in yield from buccal cells include the method of collection, the DNA extraction method (which may vary with the type of column used), and the unsupervised collection of specimens in the homes of participants and subsequent transport to central laboratories.

Specimen Type	DNA yield	Advantages	Disadvantages
Blood spots	12 - 42 ng/FL (adults)17 43 - 78 ng/FL (neonates) 17 1/4 inch punch from 75 FL volume yields about 12 FL of blood 27	-small sample size -ease of sample collection -ease of shipping (regular mail) -stability and low cost storage -offers a source for study of exogenous or endogenous compounds other than DNA -genotyping generally requires 10 ng/genotype, and with current technology as little as 2.5 ng per SNP so that scores to hundreds of genotypes could be obtained from one blood spot.	-low DNA yield: may not be suitable for whole genome amplification. -non-renewable - smaller amplicons
Whole blood (anticoagulated or blood clots) Buffy coat	100 - 400 Fg/10 mL 28-31a-d ~200 Fg/mL 28-31a-d	-relatively low-cost storage (-80E C) -yields large quantities of high-quality, genomic DNA -offers a source for study of exogenous or endogenous compounds other than DNA	-invasive sample collection -shipping (special requirements) -non-renewable
Transformed lymphocytes	106 cells = 6 Fg 28-31a-d 1 - 2 x106 cells = 5 - 10 Fg	-renewable source of DNA -yields large quantities of high-quality, genomic DNA	-labor intensive preparation -high-cost storage (liquid nitrogen-and periodic reculture) -does not offer a source for study of exogenous or endogenous compounds other than DNA or RNA
Buccal cells	49.7 Fg mean; 0.2 - 134 Fg range (mouthwash- total DNA) 32 12 - 60 Fg range(mouthwash - total DNA)33 ~16 - 30 Fg median; 1 - 290 Fg range (mouthwash- hDNA)34 32 Fg median; 4 - 196 Fg range (mouthwash- hDNA)35 1 - 1.6 Fg/2 cytobrushes median; 6 ng - 13 Fg range (hDNA)33 1 - 2 Fg/ cytobrush (total DNA) 31 1 - 2 Fg/ swab (total DNA) 31	-non-invasive collection -ease of sample collection (allows participant to collect and mail specimen). -genotyping generally requires 10 ng/genotype, and with current technology as little as 2.5 ng per SNP so that many thousands of genotypes could theoretically be obtained from a buccal cell specimen.	-low DNA yield: not in general use for whole genome amplification; -highly variable yield -does not offer a source for study of exogenous or endogenous compounds other than DNA or RNA - bacterial contamination must be addressed