methodological protocolsmethodological for improvement for highest the sensitivity has to led various Dilution due sampling to pool and need the is not feasible due to intensive the workload. tion screening. for eligibility and refinement donaof - blood repeat donors, enhanced donor questioning ures. include These careful donorselection, introductionthe preventive of several meas- ing as aconsequence past the few decades of has (HIV) progressivelyvirus dur decreased (HCV) andC virus human immunodeficiency riskThe transfusion-transmittedof hepatitis INT Keywords routinely donation screenblood for viremic donors increase safety the to of products. further blood incombination extraction magnetic bead-based ual to with used real-time PCRdetection can be and 24for were HIV) positive confirmed as serological by and NAT Conclusion methods. Results usingperformed : Out real-time PCR. of 20.808samples screened, 53samples (29for HCV for analysis the of anti-HCV/HIV antibodies. and Detection amplification RNAviral of was January 2010to September 2010,and minipools of were them submitted to testing. ELISAwas used polymerase chain reaction (RT-PCR) one-step. donor Methods: Blood samples were from collected and HCV nucleic viral acids were evaluated incombination with detection by reverse transcriptase - hepatitis C virus. attempt to reduce transmission blood-born of human (HIV) andvirus immunodeficiency tion and implemented has been at city inthe banks some of blood SãoPaulo, Brazil, inan Nucleic acidamplification testing (NAT) was recently recommended by Brazilian- legisla ABSTRACT serum in blood bank screening Magnetic bead technology forviralRNA extraction from expensive, and symptoms of an ongoing infection. absencethe of markers serological and/or signs bor large amounts of in infectiousparticles viral phase, this During an donor infected may har recipientin the by shortening window the phase. reducedhas residual the further risk of infection gen/antibody-based assays for donor screening, cation test (NAT), together with existing the anti - At present, single-donation testing is The implementation of amplifi - RODUCTION : HCV; HIV; infection; reverse transcriptase polymerase chain banks. reaction; blood 7 and in larger banks, it blood 1-3 : ManualObjective for methods extraction HIV magnetic particle-based 2-6 - -

in repeat donors. Barreto et al. donations following HIVantibody screening precipitation of particles. viral al. United States. Sabino et al. donorsers inblood is than higher that inthe of HCV infections annually. mately of 5cases HIV-1 infections and 56cases NAT prevented transmission the of approxi- United States and concluded that of use the from firstthe 4 years of NATscreening the in based on publishedbased window periods. an NAT yield of 1million donations 10.8per São Paulo, Brazil. Using a less sensitive EIA has on incidence the among first-time donors in sidual risk for HIVof 1million based 4.1per reduced to 1:2-4million. NAT for HCV and HIV, residual the risk was Statesed yearin the 2000.After introduction of and 1:1.3million for HIVinEurope and Unit- transmission of less than 1:250,000for HCV tion of efficiency, extraction such as centrifuga- 12 Allain In Brazil, residual the risk for HIV-1 mark- Stramer etal. described an estimate described of 100,000 1.56 per 8,9 or addition the of agents that support 10 described a residual risk of described viral 2 reported NAT yield data 11 and Wendel et 13 reported a re- : The man -

Carine Prisco Arnoni Prisco Carine Albertoni Guilherme Ambrozio Authors 1 José Augusto Barreto Schor Nestor Castello Girão Manoel João Batista Carvalho Fabrício Oliveira Sheila Siqueira Andrade Araujo Patricia Barboza Regina All rights reserved. All Editora©2011 Elsevier Ltda. interest. We declare no conflict of [email protected] +55 115055-6588 Fax: +55 115055-6588R-39 Phone: Brazil São Paulo, SP 04080-006 Indianópolis Av. Jandira, 1260 Guilherme Albertoni, Correspondence to: Approved on: 06/14/2011 Submitted on: 03/29/2011 5 4 3 2 Paulo, SP, Brazil deSangue,de Coleta São BeneficenteAssociação Brazil Sangue, SãoPaulo, SP, Coleta de Beneficente de Brazil São Paulo (UNIFESP), SP, Universidade Federal de Paulo, SP, Brazil deSangue,de Coleta São BeneficenteAssociação Paulo, SP, Brazil deSangue,de Coleta São BeneficenteAssociação PhD Students, – Colsan Head, –Associação Colsan Full Professor, Full Professor, – Colsan Director, – Colsan 1

2 1 4 3 5 1 1

547 ORIGINAL ARTICLE Magnetic bead technology for viral RNA extraction from serum in blood bank screening

Brazilian legislation does not yet recommend the imple- to each tube. The samples were vortexed and incubated at mentation of NAT screening for HIV and HCV. The Brazilian room temperature for 10 minutes. government has decided to create a commission to analyze Subsequently, 0.150 mL of the magnetic particles (magnet- the “advantages and disadvantages” before implementation of ic beads) were added and incubated at room temperature for NAT, 14 discussing the high cost of implementation of this test 20 minutes with agitation every 10 minutes (Figure 1A). The and the possible ways it could be used in the blood banks. tubes were placed in a magnet specific to the separation of NAT technology has been implemented in our ser- the magnetic beads containing the viral RNA, and the superna- vice since October 2008, using minipool samples (n = 6, tant was removed (Figure 1B); 0.200 mL of wash solution was 1.5 mL) for HIV/HCV, in a real-time PCR one-step test. In then added, and again the tubes were homogenized to wash the our laboratory (Colsan, São Paulo, Brazil), in the period of magnetic beads. The tubes were again transferred to the mag- 2005 to 2010, the median point estimates of residual risk net to separate the beads containing the captured viral RNA. of transfusion-transmitted infections were 1 per 93,227 for A second wash was performed with 0.150 mL, after which HCV and 1 per 97,940 for HIV. the beads were resuspended with 0.050 mL of elution buffer The introduction of real-time PCR analysis has led to (Figure 1C). The isolated RNA (0.015 mL) was used in the real- considerable progress in automating the amplification and de- time PCR for the detection of viral RNA (HIV or HCV). tection steps, but nucleic acid isolation remains laborious when After , in a first step, magnetic beads with a specific performed manually. Compared with standard separation functionalization are added to the biosuspension in order procedures, magnetic-separation techniques have several ad- to bind the RNA target. Because of the small size of the par- vantages, including simplicity in handling, a high automation ticles, thorough mixing in the next step achieves very fast potential and low cost.15 binding of the product. Subsequently, the loaded magnetic beads are retained selectively by magnetic separation and METHODS separated from the supernatant. Final washing and elution steps deliver the target product free from solids in a purified, Samples and standards often also concentrated form. Samples from Colsan Blood Bank in São Paulo, Brazil, were Reverse tested for HIV and HCV with a NAT assay (DNA Tech- RNA was reverse transcribed into cDNA by the addition of nology of Brazil) by real-time PCR (Applied Biosystems) Superscript III Platinum® One-Step in the mix containing according to the manufacturer’s instructions. In parallel, all necessary reagents for amplification. Real-time amplifica- anti-HIV and anti-HCV screening were performed by tion was performed using a 7500 Real-Time Sequence De- enzyme-linked immunosorbent assay (ELISA) testing, as tection System (SDS; ABI Prism 7500; Applied Biosystems, determined by Brazilian legislation. CA, USA). Real-time PCR product accumulation was moni- st th From January 1 , 2010 to September 30 , 2010, 20,808 tored using a TaqMan probe. donations were enrolled in this study. Two hundred-fifty mi- croliters of each serum sample were mixed in a serum pool Quantitative real-time PCR of 6 samples. If a minipool was reactive for HIV and/or HCV, Relative gene expression was calculated using conditions at the it was re-analyzed in individual reactions to determine the early stages of PCR. At this point, amplification is logarithmic identity of the reactive donor. For individual tests, 300 µL and can thus be correlated to the initial copy number of gene were used as a sample. transcripts. Reactions were cycled 40 times under the condi- Working reagents for nucleic acid amplification techniques tions previously determined by conventional PCR. At the end from DNA Technology of Brazil were used to establish the ana- of PCR reaction, the temperature was increased from 60oC to lytical sensitivity of the multiplex assay.

Control reagents contained 100 IU/mL of the RNA HCV A B C (genotypes 1-6) and HIV-1/2. Amplification of an internal con- trol was included in every assay to exclude false negative results due to interference of inhibitors and to ensure the performance of the nucleic acid extraction procedure. RNA isolation by magnetic beads Total RNA was extracted and purified using magnetic bead technology with the appropriate kit (Chemagic Viral DNA/ Addition of Specific Supernatant Wash and magnetic beads binding removal elution RNA kit). Pools of 0.250 mL of serum or plasma of each sample were centrifuged for 1 hour at 25.000 rcf to form Figure 1: The basic principle of viral isolation by magnetic the viral pellet, and then 0.100 mL of lysis buffer was added beads.

548 Albertoni, Arnoni, Araujo, et al.

95oC at a rate of 2oC/min. During this time, fluorescence was technology of Brazil). Internal controls monitored the valid- measured every 15 s to construct a melting curve. A non-tem- ity of the PCR in each reaction. plated control was run with each assay. PCR was performed with commercial primers that are selec- Statistical method tive for HIV, HCV and an internal control (DNA Technology of Kappa statistic was used to evaluate the agreement between Brazil). PCR was performed with primers HIV sense (5´ - TAA tests.16 Kappa is calculated from the observed and expected AGC TTG CCT TGA GTG CT - 3´), HIV antisense (5´- GTC frequencies on the diagonal of a square table of frequencies, TGA GGG ATC CTA GTT ACC AG - 3´), and probe HIV with a 95% interval, used to study the reliability of nominal P (5´ - FAM - AGT AGT GTG GCC GTC TGT TGT GTG - or categorical variables, where: 3´) are located in the LTR region of HIV-1 genome Accession K = (po - pe) / (1 - pe) Number - HM583638. HCV sense (5´- CGG GAG AGC CAT k = kappa statistic AGT GGT 3´), HCV antisense (5´- CGC GAC CCA ACA CTA po = proportion of observed agreement of diagnoses. CTC – 3´) and probe HCV P (5´- FAM - TGC GGA ACC GGT pe = expected proportion of diagnostic agreement GAG TAC ACC - 3´) Accession Number - HM594184.1 due to chance Enzyme-linked immunosorbent assay (ELISA) In this study, the kappa statistic should be interpret- Testing for HCV antibodies was performed with a third ed as the proportion of agreement between the two tests generation enzyme immunoassay (EIA-3; bioMérieux used beyond chance. In the case of perfect agreement, the Co.) following the manufacturer’s instructions. S/CO ra- value of Kappa = 1 (tests are consistent in all cases); Kappa tio (S, sample ratio and CO, cutoff ratio) was determined is zero if the correlation is entirely due to chance, and nega- for all the blood samples and based on this we created two tive when the correlation is lower than expected by chance. groups: Group A in which S/CO ratio was < 3, and group B in which S/CO ratio was > 3.16,17 That algorithm was devel- RESULTS oped in order to establish the relationship between S/CO ratio and the probability RT-PCR been positive.18 A total of 3,468 minipools, representing 20,808 blood dona- tions were tested simultaneously by ELISA and NAT. All pools Amplification and detection that tested negative for the presence of these viral sequences Fifteen microliters of eluate were used as a template for were in accordance with results obtained from serological rou- HCV/HIV RNA reverse transcription polymerase chain re- tine screening. In the period studied, were obtained 29 samples action (RT-PCR), by a 7500 real-time PCR system (Applied positive for anti-HCV and 24 samples positive for anti-HIV. All Biosystems). The PCR amplification and detection was per- samples were also detected by NAT. No samples were positive formed according to the manufacturer’s instructions (DNA to anti-HCV/HIV and negative by NAT (Table 1). All of the

Table 1. Samples tested for HCV/HIV NAT – Colsan (Associação Beneficente de Coleta de Sangue)

Period Number of samples Number of positive Number of positive tested in minipool samples to HIV RNA samples to HCV RNA and anti-HIV and anti-HCV

January 2010 2.421 0 4 February 2010 2.193 1 3 March 2010 4.472 2 4 April 2010 1.838 0 0 May 2010 1.628 3 2 June 2010 2.033 3 5 July 2010 2.153 6 4 August 2010 2.240 4 2 September 2010 1.830 5 5 Total 20.808 24 29

Kappa = 0.99.

Braz J Infect Dis 2011; 15(6):547-552 549 Magnetic bead technology for viral RNA extraction from serum in blood bank screening

results were validated by the internal positive control (IPC). In this study, we used magnetic beads that bind nucleic ac- There was almost perfect agreement (Kappa > 0.99) accord- ids to their proprietary magnetic polyvinyl alcohol-based parti- ing to the classification of Landis et al.17 cles. The bound nucleic acids are then transferred through the steps of the extraction process. This extraction technology has DISCUSSION successfully been evaluated to screen for viral nucleic acids in routine blood donations. Transfusion safety is one of the major concerns in the blood Pichl et al.24 used a total of 102 minipools, represent- bank setting. In the past 30 years, great progress towards ing approximately 612 blood donations that were replicated safer transfusions was made, especially concerning trans- from routine NAT pools, to analyze for B19 (parvovirus) mission of viruses. Such strides were made due to the DNA, HBV DNA, and HAV (hepatitis A virus) RNA. In our adoption of good manufacturing practices, better donor study, 3,468 minipools representing 20,808 blood donations screening policies, an increase of repeat donors, and new replicated from routine NAT pools were analyzed for HIV technologies used in determining the biological qualifica- and HCV RNA. tion of donors. Nine minipools were positive for HIV RNA, and 18 The NAT was first described in 1985 and began to be used in the screening of plasma products. In the late 1990s, were positive for HCV RNA. These minipools were bro- Europe and the USA started screening blood donations for ken out to determine which sample was reactive. All pools, HIV and HCV. Now, it is well established that implementa- positive and negative, were in agreement with results tion of NAT is a crucial step to further reduce the residual obtained from routine screening. All were valid due to a risk of transfusion-transmitted diseases, notably HIV and negative result for the IPC RNA, indicating that the tests HCV, to produce a safer blood inventory. are highly specific and reliable. Although the introduction of real-time PCR has led to The limit of detection determined for Parvovirus B19 24 considerable progress in automating the amplification and by Pichl et al. was 875 IU/mL, 260 IU/mL for HAV and detection steps of NAT, nucleic acid isolation remains very 1274 IU/mL for HBV. In our study, the detection limit for labor-intensive when performed manually. Traditional HCV and HIV was 100 IU/mL. 25 phenol-chloroform extraction and ethanol precipitation Hourfar et al. compared two methods of extraction: methods are complicated, time-consuming, hazardous, using a kit for extraction by silica column (QIAamp) and and unsuitable for processing high numbers of samples. another by magnetic beads. The detection limit was 23.1 IU Several methods of nucleic acid isolation have been de- for HIV-1 and 10.7 IU of HCV per PCR for the bead extrac- veloped, i.e., extraction based on the method described by tion and 21.6 IU for HIV-1 per PCR and 5.4 IU for HCV per Boom,19 which uses the principle of adsorption of nucleic PCR for the QIAamp extraction (Qiagen). 26 acids to silica matrices in the presence of “chemotropic Jarvis et al. uses the “decant” method on NucliSens au- salts” and alcohol.20 tomated extractor systems (bioMérieux UK Ltd., Basingstoke, According to the literature, the salt guanidine thiocy- UK) to extract nucleic acid from 2 mL plasma minipools as anate is one of the most powerful chemotropic agents and described by the manufacturer. The assays have 95% detection is commonly used in the .20 Guanidine thio- levels of 7.25 IU/mL for HCV and 39.8 IU/mL for HIV. cyanate has both RNase-inactivating and lysing properties In our study, we observed that magnetic bead technology for viral particles and cellular structures21,22 and promotes was very effective. The detection limits were 100 UI/mL, using binding of nucleic acids to silica surfaces.23 The Boom minipools and duplex PCR reactions (Table 2). method is most commonly implemented in conjunction EIA results are reported as “reactive” or “nonreactive,” and with spin columns. the EIA signal to cutoff (S/CO) ratio may also be reported as “high” or “low.” The EIA S/CO ratio is a comparison of the op- tical density of the patient’s positive EIA result to the optical Table 2. Advantages and disadvantages between density of the laboratory’s positive EIA control. If the ratio is extraction methods RNA high (> 3.0 using the most widely employed ELISA diagnostic kits), the positive predictive value (that the patient really has Silica Column Magnetics beads HCV antibodies in the blood) of the patient’s result is high. 27 Sample volume 0.140 mL 0.250 mL Albertoni et al. observed that 90% of the samples with an Assay time 30 min 3 hours S/CO ratio of > 3 were positive in the RT-PCR test, confirming the serological tests; in contrast, 13% samples with S/CO ratio Pool No Yes < 3 were confirmed as positive by RT-PCR. It was suggested Limit detection 310 IU/mL 100 IU/mL that another 87% patients were false-positive by ELISA. In Cost/test US$ 13.00 US$ 6.00 our study, we could see that all samples with S/CO < 3 were negative by RT-PCR, and samples with S/CO > 3 were con-

550 Albertoni, Arnoni, Araujo, et al.

firmed positive by RT-PCR, indicating that there is indeed 5. Coste J, Reesink HW, Engelfriet CP, Laperche S. Implementa- a strong relationship between the levels of HCV antibodies tion of donor screening for infectious agents transmitted by and HCV RNA. From the data obtained, it can be conclud- blood by nucleic acid technology. Vox Sang. 2005;88:289-303. 6. Velati C, Fomiatti L, Baruffi L, Romano L, Zanetti A. Impact ed that using magnetic beads for viral nucleic acid extrac- of nucleic acid amplification technology (NAT) in Italy in the tion from minipool samples is feasible as a manual process, three years folowing implementation (2001-2003). Euro Sur- shows high sensitivity, and offers the possibility to perform veill. 2005;10:12-4. different PCR assays from one sample eluate. 7. Jackson BR, Busch MO, Stramer SL, AuBuchon JP. The cost- It is important to note that, on average, the cost for the effectiveness of NAT for HIV, HCV and HBV in whole-blood donations. Transfusion. 2003; 43:721-9. extraction of viral RNA by the column method is about 8. Roth WK, Buhr S, Drosten C, Seifried E. NAT and viral safety $13.00 per test. However, when the magnetic beads are in blood transfusion. Vox Sang 2000;78(Suppl):257-9. used for extraction, the cost is reduced by half. Therefore, 9. Braham S, Gandhi J, Beard S, Cohen B. Evaluation of the Roche it can be concluded that the magnetic bead technology for lightCycler parvovirus B19 quantification kit for the diagnosis extraction of viral RNA is trustworthy, easy to perform, of parvovirus B19 infections. J Clin Virol. 2004;31:5-10. 10. Allain JP. Transfusion risks of yesterday and of today. Transfus and cost-effective. Clin Biol. 2003;10:1-5. 11. Sabino EC, Salles NA, Sáez-Alquezar A, Santos GR, Chamone CONCLUSION DF, Busch MP. Estimated risk of transfusion-transmitted HIV infection in São Paulo, Brazil. Transfusion. 1999;39:1152-53. The manual magnetic bead-based extraction in combination 12. Wendel S, Fachini RM, Levi JE, et al. A single window-period with RT-PCR detection can be used routinely to screen blood donation detected by human immunodeficiency vírus p24 an- donations for HIV and HCV viruses to further increase the tigen after 5 years of routine screening in a group of Brazilian safety of blood products. blood banks. Vox Sang. 2002;83:309-12. 13. Barreto CC, Sabino EC, Gonçalvez TT, et al. Prevalence, in- In this study we noticed that the centrifugation for 1 hour cidence and residual risk of human immunodeficiency virus of biological material to be analyzed has been crucial for the among community and replacement first-time blood donors formation of the pellet for extraction of viral RNA. Moreover, in São Paulo, Brazil. Transfusion. 2005;45:1709-14. it was extremely difficult to standardize the concentration of 14. Ministério da Saúde, Secretaria de Atenção à Saúde, De- primers and probes in the multiplex assay (HIV + IPC) and partamento de Atenção Especializada, Coordenação Geral de Sangue e Hemoderivados. Estágio Atual da Implantação (HCV + IPC) dos Testes de Ácido Nucléico para HIV e HCV – NAT HIV/ Author’s contribution: All authors participated in the de- HCV na rede de Serviços de Hemoterapia do Brasil. Nota sign, interpretation of the studies and analysis of the data técnica, 2010. and review of the manuscript; GAA, CPA, PRBA, SSA, FOC, 15. Hourfar MK, Schimidt M, Seifried E, Roth WK. Evaluation of MJBCG, NS, and JAB conducted the experiments, and GAA, an automated high-volume extraction method for viral nucleic acids in comparison to a manual procedure with preceding en- CPA, SSA and FOC wrote the manuscript. richment. Vox Sang. 2005;89:71-76. 16. Fleiss JL. Statistical methods for rates and proportion. 2nd ed. ACKNOWLEDGEMENTS New York, John Wiley & Sons 1981; 212-25. 17. Landis JR, Koch GG. The measurement of observer agreement This work was supported by grants from Conselho Nacional for categorical data. Biometrics. 1977;33:159-174. de Desenvolvimento Científico Tecnológico (CNPq), Finan- 18. Gerberding JL, Fleming DW, Snider DE. Guidelines for labo- ciadora de Estudos e Projetos (FINEP), Fundação Oswaldo ratory testing and result reporting of antibody to hepatitis C Ramos (FOR), Fundação de Amparo à Pesquisa do Estado de virus. Epidemiology Program Office, Centers for Disease con- trol and Prevention (CDC), Department of Health and Hu- São Paulo (FAPESP), and Coordenação de Aperfeiçoamento man Services 2003; 52:1-15. de Pessoal de Nível Superior (CAPES). 19. Boom R, Sol JCR, Salimans MM, Jansen CL, Wertheim-van Dillen PM, Van Der Noordaa J. Rapid and simple method for purification of nucleic acids. J Clin Microbiol. 1990;28:495-503. REFERENCES 20. Vogelstein B, Gillespie D. Preparative and purification of DNA from agarose. Proc Natl Acad Sci USA. 1979;76:615-69. 1. Alter HJ, Houghton M. Hepatitis C virus and eliminating post 21. Castellino FJ, Barker R. The denaturing effectiveness of guani- transfusion hepatitis. Nat Med. 2000;6:1082-86. dinium, carbamoylguanidinium, and guanylguanidinium 2. Stramer SL, Glynn SA, Kleinman SH, et al. Detection of salts. . 1968;7:4135-38. HIV-1 and HCV infections among antibody-negative blood 22. Chirgwin JM, Przbyla AE, MacDonald RJ, Rutter WJ. Isolation donors by nucleic acid-amplification testing. N Engl J Med. of biologically active ribonucleic acid from sources enriched in 2004;351:760-68. . Biochemistry. 1979;8:5294-99. 3. Roth WK, Weber M, Seifried E. Feasibility and efficacy of rou- 23. Melzak KA, Sherwood CS, Turner RFB, Haynes CA. Driving tine PCR screening of blood donations for hepatitis C virus, forces for DNA adsorption to silica in perchlorate solutions. hepatitis B virus, and HIV-1 in a blood-bank setting. Lancet. J Colloid Interface Sci. 1996;181:635-44. 1999;53:359-63. 24. Pichl L, Heitmann A, Herzog P, Oster J, Smets H, Schottstedt 4. Laperche S. Blood safety and nucleic acid testing in Europe. V. Magnetic bead technology in viral RNA and DNA extrac- Euro Survey. 2005;10:3-4. tion from plasma minipools. Transfusion. 2005;45:1106-10.

Braz J Infect Dis 2011; 15(6):547-552 551 Magnetic bead technology for viral RNA extraction from serum in blood bank screening

25. Hourfar MK, Michelsen U, Schmidt M, Berger A, Seifried E, 27. Albertoni GA, Arnoni CP, Araújo PRB, Carvalho FO, Barreto Roth WK. High-throughput purification of viral RNA based JA. Signal to-cut-off (S/CO) ratio and detection of HCV geno- on novel aqueous chemistry for nucleic acid isolation. Clin type I by real time PCR one-step method: is there any direct Chemistry. 2005;7:1217-22. relationship? Braz J Infect Dis. 2010; 14:147-52. 26. Jarvis L, Becker J, Tender A, et al. Evaluation of the Roche cobas 201 system and cobas TaqScreen multiplex test for blood screening: European multicenter study. Transfusion. 2008;48:1853-61.

552