xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader nam Ka
controlfield tag 001 001917296
007 cr mnu|||uuuuu
008 071119s2007 flu sbm 000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0002027
Cameron, Melissa A.
Evaluation of TaqMan real-time PCR for the detection of viable Cryptosporidium parvum oocysts in environmental water samples
h [electronic resource] /
by Melissa A. Cameron.
[Tampa, Fla.] :
b University of South Florida,
ABSTRACT: Cryptosporidium parvum is of growing public health concern due to its ability to survive typical water treatment processes. In order to protect the public from infection, the Environmental Protection Agency developed Method 1623 for the detection of Cryptosporidium oocysts in environmental water samples. Execution of this method is time consuming, and the results do not provide an accurate estimation of viability. Therefore, current research is focused on creating a real-time PCR method for the accurate detection of viable Cryptosporidium parvum in environmental water samples. This thesis presents the development of a real-time PCR method, and the results obtained in its use on field samples. The assay was standardized using multiple dilution series in addition to positive and negative controls. Environmental water samples were tested using this method and Method 1623 for comparison. The results were compared statistically to determine the degree of correlation between methods. The data show that the real-time PCR method correlates well to Method 1623. In addition, the assay was determined to be more cost effective and less labor intensive than Method 1623. Although these early findings are promising, additional research and development are needed before the proposed assay can be used in industry.
Thesis (M.S.)--University of South Florida, 2007.
Includes bibliographical references.
Text (Electronic thesis) in PDF format.
System requirements: World Wide Web browser and PDF reader.
Mode of access: World Wide Web.
Title from PDF of title page.
Document formatted into pages; contains 65 pages.
Advisor: Boo H. Kwa, Ph.D.
x Public Health
t USF Electronic Theses and Dissertations.
Evaluation of TaqMan Real-Time PCR for the Detection of Viable Cryptosporidium parvum Oocysts in Environmental Water Samples by Melissa A. Cameron A thesis submitted in partial fulfillment of the requirements for the degree of Master of Science in Public Health Department of Global Health College of Public Health University of South Florida Co-Major Professor: Boo H. Kwa, Ph.D. Co-Major Professor: Lillian M. Stark, Ph.D. Donna J. Haiduven, Ph.D. Date of Approval: February 14, 2007 Keywords: DAPI, EPA, IFA, method 1622, method 1623 Copyright 2007, Melissa A. Cameron
Dedication To my mother who always told me I had the ability to achieve any goal and continuously encouraged and motivated me to pursue my dreams.
Acknowledgments Conducting my research at the Florid a Department of Health Bureau of Laboratories, Tampa, I was afforded the opportunity to work along side of several outstanding individuals in the fi eld of public health. I want to thank Lillian Stark, Ph.D., Boo Kwa, Ph.D. and Donna Haiduven, Ph.D., fo r their guidance and support. I would also like to thank Deno Kazanis, Ph.D., for his valuable advice and assistance. Additionally, I greatly appreciate the time spent with me by Christy Ottendorfer and Lea Larson throughout this study, particularly for training me in EPA Method 1623. A special thanks is extended to Heidi He rnandez, Priscilla Iwakawa, and Frank Reeves, my friends at the lab who always went out of their way to help me. Also, I want to thank my family for always supporting me. Finally, and most importantly, I want to thank my brother, Ed Mitchell, and best frie nd, Ed Garbin, Ph.D., for all of their help and encouragement. I could not have a ccomplished this without their support.
i Table of Contents List of Tables iii List of Figures iv List of Symbols and Abbreviations vi Definitions vii Abstract viii Introduction 1 Cryptosporidium 1 Discovery 1 Epidemiology 2 Transmission 3 Clinical Features 4 Treatment and Prevention 6 Water Treatment and Detection Methods 7 Water Testing (EPA Methods 1622 & 1623) 7 Analysis 8 Real-Time PCR 9 Objectives 13 Materials and Methods 15 Water Sample Submission and Processing 15 Sample Size and Selection 15 Water Sample Analysis 16 EPA Method 1623 16 Elution 16 Immunomagnetic Separation 17 Disassociation 18 Staining 19 Microscopy 20 Real-Time PCR 21 Extraction 21 Positive and Negative Controls 21
ii Standardization of PCR Assay 22 Real-Time PCR Protocol 23 TaqMan Analysis 24 Enumeration and Inhibition 24 Results 38 Real-Time PCR Standardization 38 Sample Classification and Determination of Inhibition 39 Statistical Analysis of Assay Results 40 Discussion 54 Conclusion 60 References 63 Bibliography 65
iii List of Tables Table 1 Comparison of DAPI and Real-t ime PCR results for field sample data 50 Table 2 Correlation analysis of the field sample data set comparing real-time PCR and DAPI 51 Table 3 Comparison of DAPI and Real-t ime PCR results for adjusted field sample data 52 Table 4 Correlation analysis of the adjust ed field sample data set comparing real-time PCR and DAPI 53 Table 5 Cost Comparison of the c ontinuation of Method 1623 vs. the Experimental Real-Time PCR Assay for 8 samples 62
iv List of Figures Figure 1 Iddex elution device used in EPA Method 1623 for the elution of Cryptosporidium oocysts and Giardia cysts 26 Figure 2 Flow chart of sample processi ng and time requirements for processing 8 samples using Method 1623 vs. real-time PCR 27 Figure 3 Immunomagnetic separation procedure: D ynal rotation instrument with flat sided tubes mounted 28 Figure 4 Immunomagnetic se paration procedure: Dynal flat sided tubes in magnetic holder for the removal of the supernatant 29 Figure 5 Immunomagnetic se paration procedure: Washed beads are added to microfuge tubes in a magnetic holder for additional washes 30 Figure 6 Cryptosporidium producing apple-green fluorescence with IFA 31 Figure 7 Cryptosporidium producing bright blue fluorescence with DAPI staining 32 Figure 8 TaqMan template for master mix calculations and well placement of samples, spikes, controls and dilution series 33 Figure 9 Real-time PCR procedure: Load ing of sample into 96-well MicroAmp plate for analysis 34 Figure 10 Real-time PCR procedure: Mi croAmp 96-well plate loaded on the ABI 7500 35 Figure 11 TaqMan analysis: Placement of the threshold in the center of the exponential phase of the curve for determination of Ct values 36 Figure 12 TaqMan analysis: Displayed Ct values after analysis of PCR results 37
v Figure 13 Comparison of correlation coeffi cients for positive control dilution series vs. Ct values for determination of optimum primer and probe concentrations 43 Figure 14 Screen capture i llustrating the linear plot of fluorescence with all samples converging at zero for de termination of th e start cycle setting 44 Figure 15 Screen capture i llustrating the cycle at which amplification begins for determination of the end cycle setting 45 Figure 16 Percentage of samples inhibiti ng the real-time PCR process by matrix type 46 Figure 17 Percentage of ef fluent water samples inhibi ting the real-time PCR process by degrees of inhibition 47 Figure 18 Percentage of raw water sample s inhibiting the real-time PCR process by degrees of inhibition 48 Figure 19 Percentage of reclaimed water samples inhibiting the real-time PCR process by degrees of inhibition 49
vi List of Symbols and Abbreviations Symbol and Abbreviations Description % Percent Prime Degrees C Degrees Centigrade ABI Applied Biosystems Inc. bp Base pairs CDC Centers for Disease Control and Prevention C. parvum Cryptosporidium parvum Ct Cycle threshold DAPI 4Â’,6-diamidino-2-phenylindole DIC Differential Interference Contrast DNA Deoxyribonucleic Acid EPA Environmental Protection Agency g Gravitational acceleration HCl Hydrochloric Acid hsp Heat Shock Protein ID50 Infective Dose, 50% IFA Immunofluorescence assay IMS Immunomagnetic Separation ml milliliter N Normal NaOH Sodium hydroxide nm nanometers NPV Negative Predictive Value PBS Phosphate Buffered Saline PCR Polymerase Chain Reaction PPV Positive Predictive Value l microliter m micrometer M micromolar U.S. United States UV Ultraviolet
vii Definitions Cyst : a phase or form of an organism char acterized by a thick and environmentally resistant cell wall. It is produced either in response to environmental conditions or as a normal part of the life cycle of the organism Effluent : the outflow of water usually from a waste water facility that has been treated in order to be released back into the envi ronment. It may be further treated for use as reclaimed water Fomites : an object (such as an article of clothing) that may be contaminated with infectious organisms and serve in their transmission ID50: used for the dose of an infectious orga nism required to produce infection in 50 percent of the experimental subjects Inhibition : something that forbids, debars, or restricts Oocyst : an encysted zygote of certain sporozoans, e.g. Cryptosporidium characterized by a thick and environmentally resistant cell wall. It is a phase or form of an organism produced either in response to environmental conditions or as a normal part of the life cycle of the organism Raw water : water taken from the environment (gr ound and surface) that is subsequently treated or purified to produce potable water in a water purification works Reclaimed water : wastewater (sewage) that has b een treated and purified for reuse, rather than discharged into a body of water. It is frequently used to irrigate golf courses and parks and to fill decorative fountains
viii Evaluation of TaqMan Real-Time PCR for the Detection of Viable Cryptosporidium parvum Oocysts in Environmental Water Samples Melissa A. Cameron ABSTRACT Cryptosporidium parvum is of growing public health concern due to its ability to survive typical water treatment processes. In order to protect the public from infection, the Environmental Protection Agency deve loped Method 1623 for the detection of Cryptosporidium oocysts in environmental water samples. Execution of this method is time consuming, and the results do not provide an accurate estimation of viability. Therefore, current research is focused on creating a real-time PCR method for the accurate detection of viable Cryptosporidium parvum in environmental water samples. This thesis presents the development of a real-time PCR method, and the results obtained in its use on field samples. The assay was standardized using multiple dilution series in addition to positive and negative co ntrols. Environmental water samples were tested using this method and Method 1623 for comparison. The results were compared statistically to determine the degree of correlation between me thods. The data show that the real-time PCR method co rrelates well to Method 1623. In addition, the assay was determined to be more cost effective and less labor intensive than Method 1623. Although these early findings are promising, additional research and development are needed before the proposed assay can be used in industry.
1 Introduction Cryptosporidium Cryptosporidium parvum commonly known as Â“crypto,Â” originally thought to only cause disease in animals, has become a major public health concern. Over recent decades, Cryptosporidium has been linked to most waterborne outbreaks in the United States (Leav et al., 2003). It resi sts chlorination and is difficult to remove by filtration due to its small size. Hence, Cryptosporidium has become a major threat in United StatesÂ’ water supplies (Guerrant, 1997) Subsequently, it has become the most common cause of human waterborne diseas e in the United States (CDC, 2005). The disease caused by Cryptosporidium known as Cryptosporidiosis, is an enteric illness in humans and animals and has become recognized as a significant cause of diarrhea in humans. Though the disease is self-limiting in those with healthy immune systems, it is potentially life threatening in the growing number of individuals with compromised immune systems (Guerrant, 1997). For this reason, water sources must be closely monitored to assure the health of the public. Discovery Cryptosporidium was first described as an intrace llular organism in the mucosa of mice by E.E. Tyzzer in 1907 (Hannahs, 2007). It is a minute coccidean parasite and the only genus in the family of Cryptosporidiidae. Cryptosporidium was originally known as an intracellular parasite and the cause of en terocolitis in many animal species including
2 mammals, birds, and fish (Markell et al., 1999). This ability to infect a vast variety of hosts sets it apart from other coccideans. Nevertheless, it was not until 1976 that Cryptosporidium was first discovered to cause dis ease in humans (Leav et al., 2003). The first reported case involved a 3 year old girl from Tennessee who developed a severe yet self-limiting enterocolitis (Markell et al., 1999). An intestinal biopsy was performed and examination of the intestinal mucosa showed the causative organism to be Cryptosporidium parvum There are approximately 20 different species of Cryptosporidium with the primary cause of illness in humans and most mammals attributed to Cryptosporidium parvum (Roberts & Janovy, 2000). Epidemiology Cryptosporidiosis may be acquired from domestic animals as a zoonosis with mainly bovine and human reservoirs (Leav et al., 2003). It is a common cause of shortterm diarrhea (Roberts & Janovy, 2000). Humans acquire the parasite by ingesting it in its oocyst form after it is excreted in the stoo l of infected animals or people (Leav et al., 2003). Though distributed world wide a nd endemic in developing countries, Cryptosporidiosis is only seen in developed countries in sporadic outbreaks mainly affecting children and people who are im munocompromised (Leav et al., 2003). Cryptosporidiosis is often significantly under diagnosed due to its self-limiting nature. Among diagnosed cases, a vast major ity have been linked to the ingestion of water contaminated with Cryptosporidium oocysts. There have been several waterborne outbreaks of cryptosporidiosis in the U.S. (M arkell et al., 2000). An environmental study found between 67% and 95% of the surface wa ter throughout the U.S. is contaminated
3 with Cryptosporidium oocysts (Markell et al., 1999). Due to their small size, 4 to 5m in diameter, Cryptosporidium oocysts are difficult to filter from water supplies. They are also resistant to chlorination. This has l ead to numerous outbreaks throughout the U.S despite water treatment efforts. The first reported outbreak occurred in 1984 and was due to the fecal contamination of an artesian well in Texas. Another highly publici zed outbreak occurred in Milwaukee, Wisconsin in 1993 (Leav et al., 2003). The outbreak affected approximately 403,000 people and was the larg est waterborne outbreak in the U.S. (Guerrant, 1997). Of these, many became seve rely ill and several of those who were immunocompromised died. Drinking water has been identified as s ource of infection. Mo st outbreaks have occurred in communities where the local wate r utilities were meeting state and federal guidelines. Other sources of infection have been linked to public wave pools and to consumption of unchlorinated well water (Marke ll et al., 1999). The vast possibilities for infection and outbreaks illustrate the importance of routine monitoring of water supplies. Transmission The transmission of Cryptosporidium parvum is via the fecal-oral route. Infected individuals and animals shed the parasite in th eir feces in the form of oocysts as few as five days after initial infection and for up to five weeks after the diarrheal illness ends (Roberts & Janovy, 2000). A single bowel moveme nt from an infected individual or animal may release millions of oocysts (C DC, 2005). Oocysts can contaminate water, soil and food. Cryptosporidium oocysts have a thick wall which allows them to survive
4 well in the environment and withstand chlorination (Guerrant, 1997). Once the oocysts are shed into the environment there are mu ltiple means of transmission. For example, from animal to person: a person comes in co ntact with an infected animal, picks up the oocysts through contact and accidentally ingests the parasite. This means of transmission has been seen on dairy farms where the pathoge n is present in 50% of the calves on 90% of dairy farms; however, it is rare in other environments (Hannahs, 2007). Another mode of transmission is person to person contact. This occurs predominately in child daycare centers, nursing homes, and hospitals where the occupants need supervision and assistance wi th personal hygiene. Infection may spread quickly if special attention is not given to cleanliness when changing diapers or handling contaminated fomites of infected patients. Like animal to human transmission, person to person transmission is rare (Hannahs, 2007). The most common means of transmission for Cryptosporidium parvum is via contaminated food or water. Most outbreak s world wide have been transmitted through contaminated drinking water and recreational water parks (Leav et al., 2003). Source water is easily contaminated by water runoff fr om farms and grazing areas. Once oocysts are in the water supply, they are very di fficult to remove. Upon ingestion of the contaminated water, the indi vidual becomes infected, begi ns excreting oocysts and the cycle continues. Clinical Features Cryptosporidium parvum is the etiologic agent of cryptosporidiosis. Based on human studies C. parvum has an infective dose (ID50) of 132 oocysts. However,
5 infection may occur upon ingestion of as few as 30 oocysts by hea lthy individuals and a single oocyst in those immunocompromis ed (Guerrant, 1997). Once ingested, sporozoites are released and pa rasitize the brush lining of the epithelial cells located in the gastrointestinal or respiratory tract. Th is may cause a variety of symptoms two to ten days post infection depending upon the age and immune status of the individual (Leav et al., 2003). In individuals with healthy immune systems, the disease may be asymptomatic but generally causes a watery or mucous-like diarrhea that may or may not be accompanied with abdominal pain. Other symptoms may include varying degrees of nausea and vomiting accompanied with dehydr ation, low grade fever and weight loss (Markell et al., 1999). Symptoms in healthy individuals may cycle, causing a period of a few days during which the individual seems to be improving before the symptoms return (CDC, 2005). The symptoms are usually self-l imiting and mild lasting one to two weeks. However, children and pregnant women s hould be closely monitored due to their increased sensitivity to dehydration (Hannahs, 2007). Individuals with a compromised or deficien t immune system, such as people with HIV/AIDS, those who have undergone a tran splant or chemotherapy, and individuals with inherited immune disease, are much more likely to suffer from more severe symptoms (CDC, 2005). Since the early 1980Â’s, Cryptosporidiosis has become an important contributory factor in the death of AIDS pa tients (Guerrant, 1997). The infected individual may experience cholera-like watery diarrhea with as many as 6 to 25 bowel movements per day, with a stool fluid loss of up to 20 liters per day. They may
6 also experience severe abdominal pain, naus ea and vomiting. Complications may occur due to prolonged diarrhea, malabsorbtion a nd dehydration. In immunocompromised and deficient individuals, infection may occur in multiple areas of the body in addition to the intestines by penetration of th e luminal surface (Hannahs, 2007). Cryptosporidium has been found in the sputum, lung biopsy materials and the biliary tract. Symptoms may subside, but often become chronic and life-threatening. There have been many deaths attributed to Cryptosporidiosis in immunocompromised individuals (CDC, 2005). Treatment and Prevention Cryptosporidiosis is diagnosed by iden tification of the organism in biopsy material or detection of oocysts in stool samples (Guerrant, 1997). The disease may not require any treatment in healt hy individuals due to its self -limiting nature however some may be treated with nitazoxanide (CDC, 2005). Treatment for individuals with poor health or weakened immune systems is more di fficult. The effectiveness of drugs such as nitazoxanide is unclear in the immunocomp romised (CDC, 2005). The illness is not usually curable in these indi viduals and, as the immune st atus worsens, the symptoms may recur and worsen producing a chronic infection (CDC, 2005). There are few measures that may be taken to prevent possible infection by Cryptosporidium parvum The best way to prevent illn ess is to abstain from drinking water or consuming food that may be contaminated with the parasite. One must also practice good hygiene and use caution when trav eling; especially abroad, i.e. proper hand
7 washing to avoid fecal-oral contamination, drinking bottled water and consuming only cooked foods (CDC, 2005). Water Treatment & Detection Methods Public health and municipal water author ities have taken action in an attempt to assure the safety of public drinking water. The U.S. Safe Drinking Water Act (1974) requires drinking water utilities to meet st ringent standards for maximum levels of microbiological and chemical contaminants (Viessman & Hammer, 1998). The treatment process used at each facility is determined by the type of raw water source and the quality of finished water desired. To successfully remove a protozoan like C. parvum there must be effective chemical treatment and f iltration (Viessman & Hammer, 1998). It is difficult and very costly to remove all Cryptosporidium from water supplies because of the parasiteÂ’s resistance to chlorination and small oocyst size (Guerrant, 1997). Minor problems in the treatment process may go unnoticed and allow C. parvum to enter the water supply. Therefore, it is very important that munici pal water supplies are treated and monitored regularly to prevent the public from becoming ill. In order to do this, improved and faster methods of detection are necessary. Water Testing (Environmental Protection Agency Methods 1622 & 1623) The U.S. Environmental Protection Ag ency (EPA) created Methods 1622 and 1623 as a means of routine monitoring of wa ter sources to preven t the occurrence of outbreaks like that witnessed in Milwaukee. Method 1622 is specific for the detection and enumeration of Cryptosporidium oocysts and Method 1623 is specific for the detection and enumeration of both Cryptosporidium oocysts and Giardia cysts in
8 environmental water samples (EPA, 2001). Th e methods are performed by filtration of the water to be assayed, elution of the organisms from the filter, immunomagnetic separation (IMS) of the oocysts from the matrix and immunofluorescence assay (IFA) microscopy and differential interference c ontrast (DIC) microscopy (EPA, 2001). Analysis The primary method for identifying and enumerating the number of cysts and oocysts in environmental water samples, per EPA Methods 1622 and 1623, is by immunofluorescent staining and microscopic exam ination. Two types of stain are used. The first stain used is an immunofluorescen t stain, for example EasyStain C&G (BTF, Cat. #ESTAIN80) which is designed to bi nd specifically to the protein coat of Cryptosporidium and Giardia using monoclonal antibodies that cause the oocysts and cysts in the sample to fluoresce a bright a pple green color. This allows for quick identification of potentially viable oocysts and cysts. Giardia and Cryptosporidium may also be assayed with a secondary stain, 4Â’,6diamidino-2-phenylindole, commonly known as DAPI. DAPI is a fluorescent dye that binds to A-T rich double stranded DNA, produc ing a sky blue color when viewed with a UV filter block (excitation 550nm, emissi on 600nm) (Polysciences, INC., 1999). Positive DAPI staining is an indicator of potential viability, whereas DAPI negative oocysts are considered nonviable since they lack intact DNA (EPA, 2001). The slides are also examined with different ial interference contrast (D IC) microscopy for internal structures characteristic of Giardia cysts and Cryptosporidium oocysts.
9 These staining methods are capable of providing a rough estimate of the number of potentially viable oocysts in a water samp le; however, there are problems with these methods. Because they do not di fferentiate between different Cryptosporidium species, oocysts counted in this process may be of a sp ecies that is not known to cause harm to the general public. In addition, DAPI has been s hown to overestimate the number of viable oocysts in a sample by studies in which mice were infected with DAPI positive samples yet never produced infecti on (Jenkins et al., 2000). In addition to the non-species specific natu re of these stains, they are also labor intensive to perform and examine, taking hours to complete. Examination of the slides must be performed by experienced laboratorians trained in the science. However, a great deal of the interpretation of the results is le ft to the discretion of the examiner. Although both of these stains allow us to assess the pos sibility of contaminants in water samples, there needs to be a less subjective, more spec ific and efficient mean s of detecting viable Cryptosporidium in environmental water samples in order to significantly reduce the possibility of misi nterpretation. Real-Time PCR In attempts to find a more specific and timely method for the detection of Cryptosporidium in environmental water samples, scientists are looking toward real-time polymerase chain reaction (PCR). It is a method that allows for the logarithmic amplification of short strands of DNA and detection in Â“real-timeÂ” by the reporting of fluorescent probes. Theoretical ly, a single copy of a particul ar sequence can be amplified and detected, through the use of appropriate pr imers and probes, with a direct relationship
10 between the starting target and the amount of product at a given cycle (Ambion, 2007). This is accomplished by cy cling the sample through various thermal cycles, usually ranging in number from 40 to 50, during wh ich the DNA is replicated (Ambion, 2007). As the DNA replicates, the probe searches fo r a specific target nucleic acid sequence. The probe attaches to the target DNA and cleav es, creating a fluorescence that is detected by the real-time PCR instrument after each thermal cycle is completed. This method is commonly used for detec tion and quantification of various viruses and parasites in numerous sample mediums. Th e ability to detect a specific sequence and the fast result time (2 to 3 hours) makes real-time PCR a good candidate for future Cryptosporidium testing and water monitoring. Many st udies have been done in attempts to create a method that detects viable Cryptosporidium in water samples; however, few have been done using environmental samples as opposed to spiked laboratory samples. In 1995, Wagner-Wiening and Kimmig detected viable Cryptosporidium parvum oocysts using traditional PCR. The study used oocysts that were placed in media and excysted to assure viability. The PCR gene rated a product 873 base pairs (bp) in length encoding an oocyst protein. Th e procedure was successful in detecting viable oocysts; however, the results were not easily replicated or pr edictable. Kaucner and Stinear (1997) detected viable C. parvum oocysts in large volumes, 20 to 1,500 liters, of spiked creek and river wa ter. They devised a method for detecting a smaller segment of DNA measuring only 307 bp in length, from a heat shock protein found in C. parvum The method was not tested on e nvironmental samples. These two
11 studies did not permit enumeration of oocysts but they did set the course for further research in the area. A Most-Probable-Number assay was de veloped by Slifko, Huffman, and Rose (1999) that enumerated infectious C. parvum oocysts in cell culture systems. They also were able to determine from this study that the age of the oocysts affected its infectivity. This discovery illustrated the ability of oocysts enumeration and a need for a more precise method for determining the viability of oocysts. Gobet and Toze (2001) conducted a study to determine the relevance of heat shock protein (hsp) 70 messenger RNA a nd DNA to determine the viability of Cryptosporidium parvum oocysts. The study compared an assay using this protein with methods utilizing mouse infectivity and imm unofluorescent dyes. The poor specificity and sensitivity of immunofluorescent dyes and th e unreliability of infectivity assays lead to the determination that DNA encoding for hsp70 was the best indicator of viability. They also noted that the amount of DNA detect able in the oocysts decreased quickly after they became nonviable allowing for more pr ecise detection of only viable oocysts. In 2003, Fontaine and Guillot developed a method for an immunomagnetic separation real-time PCR fo r the quantification of C. parvum in water samples. The method followed the previously used EPA Method 1622 for the detection of Cryptosporidium species in water samples. They were able to detect as few as 5 oocysts in spiked samples. Though this method was su ccessful with laboratory spiked samples, it did not use hsp70 as the target and was not designed to detect only viable oocysts.
12 During the same period of time, LeCheval lier, et al. (2003) were comparing EPA Method 1623 with a cell culture PCR method. The two me thods produced similar results. Their study showed the usefulness of hsp70 for detection of viable C. parvum in water samples. Though the method was successful and able to detect low quantities of viable oocysts, it utilized traditional gel based PCR and was labor inte nsive to perform. The results of both 2003 studi es just described were us ed in the development of the present research presented herein. Th e method developed for this study used immunomagnetic separation follo wed by real time PCR with hs p70 as the target in order to detect only viable oocysts in spiked water samples and environmental samples received at the Florida Department of H ealth, Bureau of Laboratories in Tampa.
13 Objectives Many methods have been used to assess the viability of Cryptosporidium parvum oocysts. For example, one method consists of growing cell cultures and inoculating the cells with oocysts. This is done to determine if the oocysts are viable (capable of causing infection). This method is time-consuming, ta king weeks to perform, and underestimates of the number of viable oocysts. Met hod 1623, developed by the US Environmental Protection Agency, is the current test used by most laboratori es. Although this method is widely used, it is also a time-consuming met hod requiring at least one day for processing. The results depend on the interpretation of the technologist and may lead to an overestimate of the number of vi able oocysts in the sample. The inaccuracy and inefficiency of curre nt testing methods necessitates finding a more precise and practical method for laborator ies to determine oocyst viability in public water sources. Current research utilizing th e real-time PCR shows promise. Primers and probes have been report ed in the literature for testing the viability of Cryptosporidium parvum oocysts, but the tests are no t standardized and are not cu rrently in routine use, nor have they been evaluated to determine thei r efficiency in a public health laboratory. The hypothesis of this study is that a prot ocol can be developed for real-time PCR based assay for testing of Cryptosporidium parvum in environmental water samples that is as sensitive as the current testing methods. This method is anticipated to give a more accurate estimate of oocyst vi ability due to the specific na ture of the test and the
14 elimination of examiner interp retation. This new method w ill decrease the total time needed to complete the assay and allow for qui cker reporting of results. This will, in turn, aid in the prevention of illness and allow control measures to be implemented in a more timely manner should an outbreak occur. This study evaluates the potential us e of real-time PCR to determine Cryptosporidium parvum oocyst viability by comparing the viability of C. parvum oocysts estimated by EPA Method 1623 with TaqMan real-time PCR. The study has four specific aims: 1) to standardize and validate a protocol for using Real Time PCR detection on viable Cryptosporidium parvum oocysts in various types of water samples; 2) to determine the sensitivity and statistical relationship of real-time PCR method as compared to IFA/DAPI staining methods; 3) to determine if the statistical relationship varies by water sample types, i.e. raw, treated waste water (effluent ), reclaimed water, or dr inking water (potable); and, 4) to determine the best testing method for viable Cryptosporidium oocysts by analyzing cost and time efficiency. The overall goal of this study is to e nhance public health by improving current Cryptosporidium detection methods in environmental water samples, and to aid in the prevention and control of infection.
15 Materials and Methods Water Sample Submission and Processing Water samples are received weekly at the Florida Department of Health Bureau of Laboratories in Tampa, to be tested for Cryptosporidium and Giardia The samples are submitted from various water and waste water utilities throughout the state of Florida. Sample types received vary depending on the facility requesting the testing which include raw (ground or surface), efflue nt (treated waste water), re claimed (re-use), and drinking (potable) water. The raw water used in this study is from ground or surface water sources. Raw water is treated and purified to be used as potable or drinking water. Effluent is the outflow of water usually from a waste water facility that has been treated in order to be released back into the environment. Efflue nt can be further treated and purified, known as reclaimed water, for use in irrigation systems and outdoor fountains. On average the Tampa Laboratory receives approximately 130 samples a year; however this number is steadily increasing. Upon submission, samples are logged, numbered and stored at 4C until testing. All samples are test ed within 96 hours of collection per EPA Method 1623 requirements. Sample Size and Selection A sample size of 32 was needed in this study to assure statis tically significant results. The sample size was determined us ing the sample size cal culator developed by Cameron and Baldock based upon a population of 130 (the average number of samples
16 collected per year), a sensitivity and specific ity of 80%, a prevalence of 45% (based on an average of 59 positive samples per year), a level of significance of =.05, and a power of 95% (AUSVET, 1998). Due to the limite d number of samples collected each year, every sample received between September and December of 2006 was assayed in this study. This resulted in a final sa mple size of 40 for this study. Water Sample Analysis EPA Method 1623 All water samples received are tested for Cryptosporidium and Giardia protozoa using EPA Method 1623. This method will be su mmarized here. A complete protocol is available at http://epa.gov/ waterscience/methods/1623.pdf. In the field, water sources are passed through filters (IDDEX, Cat. # FMC10601) which are designed to trap Cryptosporidium Gardia and extraneous material. The filters are sent on ice to the Department of Health for processing and testing. Upon arrival at the laboratory, the samples are numbered and held between 0 and 8C until tested. Elution The Filta-Max system (IDDEX Cat. #FMC 10102, FMC 10301, FMC 12001) is used to elute the material off of the submitted filter. The apparatus consists of an elution tube and plunger containing a membrane desi gned to concentrate the oocysts and cysts (Figure 1). The filter is placed in the elut ion tube along with any remaining liquid from the filter carriage. The filter is processed us ing a series of washes to assure all protozoa are concentrated on or above the membrane at the base of the tube. The membrane, with protozoa attached, is remove d from the device and washed to remove the oocysts and
17 cysts from the membrane. The membrane is discarded. The oocyst suspension is placed in 50 ml sterile polypropylene conical t ubes and centrifuged at 1,500g for 15 minutes. This results in the formation of a pellet, c onsisting of oocysts, and the supernatant. The supernatant is removed to leave a volume of approximately 5ml above the pellet. At this point, the pellet size is meas ured and recorded. The pellet is divided in two with one portion continuing on to the immunomagnetic separation step, the other preserved with 10% neutral buffered formalin and archived. For the purpose of this study, the second portion of the pellet conti nued through the IMS step and was processed for real-time PCR testing (Figure 2). Immunomagnetic Separation Dynal flat-sided tubes (Dynal Inc., Ca t. #740.03) are labeled and 1ml each of buffer A and buffer B from the Dynal IMS kit (Dynal Inc., Cat. #730.02) is added to each tube. Both pellets are resuspended by vorte xing for 2 minutes. The sample pellets are removed with a pre-wetted pipette and added to the appropriate Dynal tube. The conical tubes are rinsed twice with the appropriate am ount of sterile water in order to produce a total volume of 12ml in the Dynal tube. Magnetic beads with antibodies specific for Giardia and Cryptosporidium are added. The Dynabeads Giardia-Combo and CryptoCombo vials (Dynal Inc., Cat. #730.02) are vort exed and 100l from each vial are added to each tube. The Dynal flat-sided tubes ar e then placed on a rota tion device (Dynal Inc., Cat. #947.01) and rotated at a speed of 18 ro tations per minute for one hour (Figure 3). After the samples comp lete the rotation, the Cryptosporidium oocysts and Giardia cysts present in the samples should be att ached to the magnetic beads by an antigen-
18 antibody reaction. The Dynal flat -sided tubes are removed from the rotator and placed in a magnetic holder with the flat portion of the tube facing the magnet (Figure 4). The tubes are rotated in the hold er 90 end to end for two mi nutes at one tilt per second causing the magnetic beads to adhere to th e flat portion of the Dynal tube. The supernatant is immediately decanted; the tube s are removed from the magnetic holder and placed in a test tube rack. The tube contai ning the magnetic beads is rinsed three rinses with a 1X buffer A solution to assure rem oval of the magnetic beads from the Dynal tubes: 0.5ml of the 1X buffer A solution is ad ded to the tube. Th e tube is caped and gently rocked to rinse the beads from the flat side of the tube. The solution is removed using a pre-wetted pipette and added to a micr ofuge tube. The process is repeated with an additional wash using 0.5ml and a final wa sh of 0.4ml. The microfuge tube is capped, placed in a holder and a magnetic strip is ad ded to the holder (Figure 5). The tubes are gently rocked back and forth through 180 for one minute at one rock per second allowing the magnetic beads to adhere to th e side of the tube. The supernatant is removed leaving the magnetic beads with any a ttached protozoa. At this point, the two samples separated at the beginning of the IMS proceed to different steps. One microfuge tube continues to the IMS disassociation st ep, staining with DAPI and EasyStain, and microscopic examination while the other pr oceeds to the experimental nucleic acid extraction, real-time PCR and analysis. Disassociation The magnetic bar is removed from the microfuge tube holder and 50l N HCl is added to the tube. The tube is vortexed for 50 seconds, placed back into the tube holder
19 without the magnetic strip a nd let stand for 10 minutes. This causes the oocysts to disassociate from the beads. The tube is vor texed for 30 seconds and the tube is tapped to assure any drops in the cap return to the base of the tube. The magnetic strip is placed back into the holder. The holder is placed at a slant and let stand for 10 seconds. A well slide is labeled and 10l of 1.0 N NaOH is added to the well. Without removing the magnetic strip the supernatant is removed from the microfuge tube and added to the well slide. The disassociation step is repeated a nd the supernatant is adde d to the slide. The slide is now ready for staining. Staining During the staining process a positive and negative slide are also prepared for controls. The slides containing the sample and the controls are placed on a slide warmer set to 37C and allowed to air dry. Each slid e is treated with 50l of absolute methanol and allowed to air dry. The slide is removed from the wa rmer and 50l of DAPI staining solution (Sigma, Cat. #D9542-1MG) is added to each slide and allowed to stand for 2 minutes. The solution is removed from the slide by tilting it on a paper towel. A volume of 50l of sterile water is a dded to the well and the slide stands for one minute. The water is removed as before by tilting the slide on a paper towel. 50l of EasyStain, consisting of an immunofluores cence reagent designed for use on Cryptosporidium oocysts and Giardia cysts in water samples, (BTF, Cat. #ESTAIN80) is added to the slide and the slide is placed in a humid, dark ch amber at room temperature for 30 minutes. The stain is removed from the slide by tilting it on a paper towel. At this time, 300l of Fixing Buffer from the EasyStain Kit (BTF, Cat. #ESTAIN80) is added to the slide. The
20 slide stands for 2 minutes and the buffer is removed by placing the slide tilted on a paper towel. 10l of mounting buffer is added to the well and a cover slip is put in place. The cover slip is sealed in pla ce using two coats of clear nail polish around its edges. Once dried, the slide is ready for microscopic analysis. Microscopy A Zeiss epifluorescent microscope (Zeiss, model #AXIOSKOP2) is used to scan the entire well at 200X or 400X for an a pple-green fluorescence which indicates the possible presence of cysts and oocysts. Wh en bright apple-green fluorescing ovoid or spherical objects, rangin g in size from 4 to 6 m in diameter, are observed with highlighted edges they are counted and recorded (Figure 6). A UV filter block (excitation 550nm, emission 600nm) is put in place for DAPI examination. The oocysts may exhibit one or more of th e following characteristics: a light blue internal staining and no distinct ive nuclei, an intense blue in ternal staining, or no more than four distinct sky blue nuclei (Figure 7) The results are recorded for the first ten organisms examined. If the organism exhibits a green rim without in ternal structures, it is characterized as DAPI negative. Organisms are recorded as DAPI positive if there is a strong blue internal stain or distinctly st ained nuclei are present. The UV filter is removed; magnification is increased to 1000X (oil emersion) for DIC examination of possible internal structures.
21 Real-Time PCR Extraction Samples that do not undergo the IMS di sassociation step mentioned above continue to the experimental real-time PCR method. 1ml of 1X phosphate buffered saline (PBS) added to each microfuge tube containing the magnetic beads to wash away any residual Buffer A solution. The samp les are vortexed, returned to the magnetic holder, and rocked 180 for 2 minutes allowing th e beads to attach to the side of the tube containing the magnet. A solution is prepar ed at a 25% weight-t o-volume concentration of Chelex 100 resin (BioRad, Cat. #143-2832) and sterile reagent grade water to be used in the extraction. The wash supernatant is removed from the microfuge tube containing the magnetic beads and 200l of Chelex 100 re sin solution is added to the tube. The DNA is extracted using a simple freeze-thaw method. The tubes are submerged into liquid nitrogen for 2 minutes Upon removal, they are immediately placed into a water bath at 95C for 2 minutes This freeze-thaw process is repeated for four more cycles. After the final cycle, th e sample is placed in a microcentrifuge and spun at 10,000g for 10 minutes to separate the beads and Chelex 100 resin from the supernatant containing the DNA. The sample ly sate is now ready for PCR testing and is stored in a -20C freezer until real-time PCR is performed. Positive and Negative Controls Positive and negative controls are used to assure accuracy in the PCR testing. Viable Cryptosporidium parvum oocysts in 1X PBS were obtained from Waterborne Inc. (New Orleans, LA) with a guaranteed number of viable oocysts at 1 x 106 per 4ml. A
22 volume of 500l is removed from the vial and placed in a screw top microfuge tube marked positive. An additional 500l is removed from the vial, boiled for one minute in order to render the Cryptosporidium oocyst nonviable and is placed in a screw capped microfuge tube marked negative. Both t ubes undergo the same freeze thaw extraction method used on the samples. Standardization of PCR Assay The PCR assay that was developed for th is study was based on methods used by LeChevallier et al. for the detection of viable Cryptosporidium and Fontaine et al. for IMS detection of Cryptosporidium parvum The methods were combined to allow for testing of viable oocysts in environmental water samples. Primers and probes specific for Cryptosporidium parvum hsp70 DNA were obtained from Operon (Huntsville, AL). Primer and probe sequences to be used are as follows: forward primer: 5' TCCTCT GCCGTACAGGATCTCTTA 3'; reverse primer: 5' TGCTGC TCTTACCAGTACTCTTATCA 3'; TaqMan probe: 5' 6-carboxyfluorescein TGTTGCTCCATTATCACTCGGTTTAGA 6carboxytetramethylrhodamine 3'. Based on experimental results the optimal conc entration for the primers and probe in this study are 100M and 25M respectively and are subsequently used to test all collected samples.
23 Real-Time PCR Protocol Real-time PCR is performe d with the Applied Biosys tems (ABI) 7500 Fast RealTime PCR System (Applied Biosystems, CA ) and TaqMan One Step RT-PCR Master Mix Kit (Applied Biosystems, Cat. #4309169). A template is created for calculating the quantities of each master mix ingredient per sample being tested; an example is seen in Figure 8. For each sample to be assayed, a master mixture is made containing: Dnase/Rnase free water, 6.35l; TaqMan Univ ersal 2X PE Master Mix, 12.5l; forward primer, 0.25l; reverse primer, 0.25l; probe 0.15 l; enzyme, 0.50l. These ingredients are combined in a SafeLock 1.5 microfuge tube (Eppendorf, Cat. #0540334B), vortexed, and briefly pulsed to remove any drops from the lid. Samples are tested on a MicroAmp Fast Optical 96 well reac tion plate (Applied Biosystems, Cat. #4346906) along with positive and negative controls, a positive control dilution series for quantification and spiked sa mples to determine if the is any inhibition of the PCR run. The template sheet also il lustrates the well lo cation of each sample, spiked samples, controls, and dilution series (F igure 8). Each plate is loaded with 15l of master mix per well. A volume of 10l of DNA template from the extracted sample is added to the appropriate well creating a tota l reaction volume of 25l (Figure 9). The mixture is mixed by pipetting up and down caref ully so as to not create bubbles or aerosols. Spiked samples are loaded with 15 l of master mix, 8l of sample and 2l of a known positive control. A well is also loaded with the same volumes of master mix, sterile water and positive control for comparison against spiked samples. An optical
24 adhesive cover (Applied Bios ystems, Cat. #4311971) is placed on the plate to seal the wells. The sample plate is loaded in the AB I 7500 (Figure 10). The detectors, FAM and TAMRA, are selected and applied to the enti re plate. The optimum cycling times are programmed. These include an initial dena turing step of 95C for 10 minutes and an amplification step of 50 cycl es of 95C for 30 seconds fo llowed by 60C for 1 minute. The sample volume is set to 25l and a r un mode of 9600 emulation with detection occurring at the 60C stage. The run takes approximately 3 hours to complete. TaqMan Analysis Once the run is completed, the manual Ct and manual baseline settings are selected. Start and end cycles are set to 13 and 25 respectively, based on experimental results. The threshold is moved to the mid point of the exponential portion of the curves (Figure 11). The data is analyzed and cycl e threshold (Ct) values are displayed following the instrument instructions. The Ct values displayed for each sample allow for enumeration and determination of run validity (Figure 12). Enumeration and Inhibition Based on the Ct values for the positive a nd negative controls one can assess if the run is valid. Positive controls should have a Ct value of less than 45 and the negative control should be listed as undetected. The Ct value gi ves a reference value for the enumeration of the amount of viable Cryptosporidium in the sample. For example, if a known number of 250 oocysts are placed in the positive control well, the Ct values of
25 each sample can be compared to the Ct of the positive control, estimating the number of oocysts in each sample. Some field samples may inhibit the PCR r eaction. In order to account for this, the spiked samples are examined. The Ct value for the well containing the water spike (8l water and 2l positive control) is compared to the Ct value of each spiked sample. If the Ct value is less than two standard deviations from the spiked control Ct, the sample is not considered to be inhibitory. Ct values on spiked samples that are 2 to 4 standard deviations greater than the c ontrol spike are reported with minor inhibition. Samples that are 4 to 6 standard deviations above the control are reported with major inhibition. Samples 6 or more standard deviations above the control and CtÂ’s li sted as undetected are reported with major or complete inhibition, respectively.
26 Figure 1 Iddex elution device used in EPA Method 1623 for the elution of Cryptosporidium oocysts and Giardia cysts
27 Figure 2 Flow chart of sample processi ng and time requirements for processing 8 samples using Method 1623 vs. real-time PCR
28 Figure 3 Immunomagnetic separation proce dure: Dynal rotation instrument with flat sided tubes mounted
29 Figure 4 Immunomagnetic separation pro cedure: Dynal flat sided tubes in magnetic holder for the removal of the supernatant
30 Figure 5 Immunomagnetic separation pro cedure: Washed beads are added to microfuge tubes in a magnetic holder for additional washes
31 Figure 6 Cryptosporidium producing apple-green fluores cence with IFA (obtained from www.griffin.uga.edu)
32 Figure 7 Cryptosporidium producing bright blue fluorescence with DAPI staining (obtained from www .griffin.uga.edu)
33 Figure 8 TaqMan template for master mix calculations and well placement of samples, spikes, controls and dilution series
34 Figure 9 Real-time PCR procedure: Load ing of sample into 96-well MicroAmp plate for analysis
35 Figure 10 Real-time PCR procedure: MicroA mp 96-well plate loaded on the ABI 7500
36 Figure 11 TaqMan analysis: Placement of the threshold in the center of the exponential phase of the curve for determination of Ct values
37 Figure 12 TaqMan analysis: Displayed Ct values after analysis of PCR results
38 Results Real-Time PCR Standardization A required minimum sample size of 32 was determined by use of the Cameron and Baldock sample size calculator. Due to the unpredictable schedule of sample submission by water and waste water utilities, every sample received between September and December 2006 was used in this study inst ead of a random sampling. This resulted in a final sample size of 40. All field sample s were processed and tested for viable Cryptosporidium parvum oocysts using the method developed for this study. Positive and negative controls were tested against various primer and probe concentrations ranging from 30 to 300M for primers and 10 to 200M for probes. Using the 3 primer and probe concentra tions (60/25, 30/10, 100/25) that accurately detected positive and negative controls, multiple dilution series of positive controls were assessed to determine the best concentrations to be used in the re al-time PCR protocol. This was determined by finding the concentratio ns of primers and probes that resulted in a correlation coefficient closest to one to illustrate perfect unity. Based on this data, a concentration of 100M for the primers and 25 M for the probe was selected for use in this study (Figure 13). Three ratios of master mix to DNA template were tested to determine the best ratio to use for the assay. The first test util ized a master mix volume of 20l with 5l of template. The second used a 10l master mix volume and a 15l template volume.
39 Amplification curves produced from these two tests were jagged and did not have a distinctive exponential phase. The third mixture tested incl uded 15l of master mix and 10l of DNA template. This run generated am plification curves that were smooth with a distinctive exponential phase, indicating the 15:10 ratio of master mix and DNA template to be most suitable for use in this PCR protocol. Initially, 45 cycles of DNA amplification we re used in the PCR protocol. Results from this first run produced nicely shaped curves but the curves did not complete the exponential phase prior to the run termination. Therefore, the real-time PCR runs were assessed at 50 and 55 cycles. The 50 cycle run resulted in smooth and clearly defined exponential curves at 50 cycles, while the 55 cycle run produced results similar to those of the 45 cycle trial. Once cycling times were determined, the linear graph of the dilution series was examined to find the cycle number at which all fluorescent curves were at zero and converge. This cycle was identified as 13, and the start cycle wa s set to this value (Figure 14). Initial amplifi cation occurred at 30 cycles. The end cycle which determines the end of the background noise, was set to 5 cycl es before this point, or 25 (Figure 15). Sample Classification and Determination of Inhibition Water samples were categorized into 4 groups based on the type of water noted on the submission paperwork. Of the 40 samples collected, 21 were reclaimed water, 10 were raw (ground or surface) water, 5 were e ffluent, and 4 were potable water samples. These samples were analyzed using real-time PCR and further classified. Real-time PCR assays utilizing positive control dilution series were analyzed and a standard deviation from the mean of 0.699 was calculated. The inhibition of PCR by
40 the sample matrix was determined in increments of 2 standard deviations. Ct values less than two standard deviations above the sp iked control Ct were reported as having no inhibition. Ct values 2 to 4 standard deviations above the spiked control were reported with minor inhibition. Samples 6 or more sta ndard deviations from the control and CtÂ’s listed as undetected were reported with ma jor and complete inhibition, respectively. Inhibition was detected in all water types in varying degrees except in the potable water samples (Figure 16). Effluent water samples illustrated the highest degree of inhibition at 80%, with 20% of the samples showing minor inhibition and 60% displaying complete inhibition (Figure 17). Similar results were observed in the raw water samples (70%); complete inhibition was observed in 60% of the samples and 10% produced minor inhibition (Figure 18). Reclaimed water samp les produced an overall inhibition of 52%, where 28% of the samples produced complete inhibition, 5% produced major inhibition, and 19% showed minor inhibition (Figure 19) Samples tested from potable water sources were the only samples that produced no inhibition. Statistical Analysis of Assay Results A statistical analysis was performed on the results obtained from each real-time PCR on field samples, with triplicate runs used to account for vari ability. Using two-bytwo tables comparing real-time PCR and DAPI results from the split IMS pellet, the sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) of this PCR assay were determined. The real-time PCR assay had a specificity of 100% and a sensitivity of 56%, a PPV of 100% and a NPV of 73% (Table 1).
41 In addition to the aforementioned tests, statistical tests which study equivalency and measure the association between vari ables with dichotomous outcomes were performed on the data obtained from this pr oject. Additional tests included the Kappa measure of agreement, PearsonÂ’s Correlati on, and YoudenÂ’s J. These were performed using diagnostic effectiveness software (Simpl e Interactive Statistical Analysis, 2007). The Kappa measure of agreement determines the degree of agreement between compared tests. A value of zero is produced if there is no agreement between tests and a value of one results if the te sts are in perfect agreement (i .e., they correctly predict the outcome). Values that fall between zero and one are classified by degrees of agreement (Szklo & Nieto, 2000). The Kappa value fo r this test was determined to be 0.59, indicating a substantial agreement be tween DAPI and real-time PCR. PearsonÂ’s correlation indicates the amount of correlation between the expected Ct, based on percent DAPI positive oocysts, and mean Ct values. A value close to one indicates a good linear correlati on between the values whereas values of zero indicate there is no correlation between the two va lues. This assay produced a PearsonÂ’s Correlation value of 0.64, indicating a posit ive correlation between the two tests. The final statistical test performed on the data set was YoudenÂ’s J, which determines if the results are in agreement or produced solely be chance. A value of one indicates the tests are in perf ect agreement. A value of zer o indicates the results of the test occurred due to chance alone (Szklo & Nieto, 2000). The real-time PCR assay had a YoudenÂ’s J of 0.56 (Table 2).
42 Due to the high level of inhibition, a st atistical analysis was repeated on all samples that did not produce inhibition to de termine the overall perf ormance of the realtime PCR assay in the absence of inhibition. The samples that produced inhibition were removed from the data set and counted as failed runs. The additional statistical analysis performed on the revised data set increased the sensitivity of the test to 89%. It also resulted in an increase in the NPV and PearsonÂ’s correlation to 90% and 0.89, respectively. The Kappa value also incr eased to 0.89, showing a better correlation between the two testing methods (Tables 3 & 4).
43 Figure 13 Comparison of correlation coefficients for positive control dilution series vs. Ct values for determination of optimum primer and probe concentrations Primer Concentration Standard Curve Primer Set A R2 = 0.9769 Primer Set B R2 = 0.9825 Primer Set C R2 = 0.999715 20 25 30 35 40 45 50 55 600123456789 Dilution FactorCT Value Primer Set A Primer Set B Primer Set C Linear (Primer Set A) Linear (Primer Set B) Linear (Primer Set C)
44 Figure 14 Screen capture illustrating the linear plot of fluorescence with all samples converging at zero for determ ination of the st art cycle setting
45 Figure 15 Screen capture illustrating the cy cle at which amplification begins for determination of the end cycle setting
46 Figure 16 Percentage of samples inhibiti ng the real-time PCR process by matrix type 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% EffluentRawReclaimPotable N=4 N=7 N=12 N=0
47 Figure 17 Percentage of effluent wate r samples inhibiting the real-time PCR process by degrees of inhibition 0% 10% 20% 30% 40% 50% 60% 70% CompleteMajorMinor N=0 N=3 N=1
48 Figure 18 Percentage of raw water sample s inhibiting the real-time PCR process by degrees of inhibition 0% 10% 20% 30% 40% 50% 60% 70% CompleteMajorMinor N=5 N=0 N=2
49 Figure 19 Percentage of reclaimed wate r samples inhibiting the real-time PCR process by degrees of inhibition 0% 5% 10% 15% 20% 25% 30% CompleteMajorMinor N=6 N=1 N=4
50 Table 1 Comparison of DAPI and Real-time PCR results for field sample data DAPI Stain Results Real-time PCR DAPI +DAPI Totals PCR + 9 0 9 PCR 9 22 31 Totals 18 22 40 Sensitivity 56% Specificity 100% PPV 100% NPV 73%
51 Table 2 Correlation analysis of field sample data set comparing real-time PCR vs. DAPI Correlation Statistical Analyses Assay PearsonÂ’s YoudenÂ’s J Kappa PCR 0.64 p=0.0000 0.56 0.59
52 Table 3 Comparison of DAPI and Real-tim e PCR results for the adjusted field sample data DAPI Stain Results Real-time PCR DAPI +DAPI Totals PCR + 8 0 8 PCR 1 9 10 Totals 9 9 18 Sensitivity 89% Specificity 100% PPV 100% NPV 90%
53 Table 4 Correlation analysis of adjusted field sample data set comparing realtime PCR vs. DAPI Correlation Statistical Analyses Assay PearsonÂ’s YoudenÂ’s J Kappa PCR 0.89 p=0.0001 0.89 0.89
54 Discussion This study describes the development of a real-time PCR protocol for the detection of Cryptosporidium parvum in environmental water samples. Due to C. parvumÂ’s low infective dose and its ability to evade conventi onal water treatment, it is important to devise an assay that has the abil ity to reliably detect viable oocysts in a variety of water types. Te sting of municipal water syst ems was established by the EPA to ensure the safety of the public, and Met hod 1623 has been successful in the detection of Cryptosporidium and Giardia in environmental water supplies. However, the method is very labor intensive and tedious to perform. Als o, Method 1623 does not have the ability to distinguish between different species of Cryptosporidium or Giardia and it does not have the ability to give an accurate estimate of viability. A real-time PCR assay for the detection of C. parvum would be beneficial to facilities that regularly pe rform water testing. Realtime PCR has the ability to dramatically decrease the time required to obtai n and report results (Fi gure 2). This type of assay has the ability to na rrow the range of species of Cryptosporidium detected to C. parvum specifically. Therefore, it will only detect those species that are of human concern. It also has the ability to accurately estimate oocysts viability due to the use of the hsp70 DNA as a target, since hsp70 quick ly degrades once the oocyst become nonviable. Real-time PCR was shown to accurately detect C. parvum oocysts in cell cultures in a study by Fontaine et al. ( 2003). It has also been demons trated by LeChevallier et al.
55 that viable C. parvum oocysts are detectable with high specificity in samples when using hsp70 gene as a target (2003) These two methods brought the option of real-time PCR for the detection of viable C. parvum oocysts to the forefront. This study built upon the results of these two studies to produce a protocol for a real-time PCR assay that detected viable C. parvum in environmental water samples. The testing of various primer and probe concentrations yielded a final concentration of 100M for the primers and 25M for the probe used in this assay. These concentrations produced a correlation coefficient of 0.9997 illustrating an almost perfect relationship between the number of vi able oocysts and the Ct values. As the number of oocysts increased in the sample the Ct values accurately and predictably decreased linearly. The assay was further evaluated to de termine the optimum master mix to DNA template ratio. The amplification plots for each ratio were compared and a 15:10 (master mix: template) ratio was determined to be best suited for this assay. Whereas other concentrations produced jagge d curves and minor exponentia l phases that were difficult to decipher, the amplification plot for this mixture was smooth and free of excess background noise. It also produced a clear e xponential phase with a distinct plateau, allowing for data analysis. Amplification plots were also assessed to determine the number of cycles that should be used in each PCR run. Initially 45 cy cles were used, resulting in plots that had a very small exponential phase and never reached a plateau. The small exponential phase did not allow for accurate analysis of the run, due to the difficulty of aligning the
56 threshold in the center of the phase. The number of cycles were reset to 50 resulting in a much larger exponential phase and plateau in the final cycles. This adjusted the placement of the threshold and allowed for the accurate reporting of Ct values. The assay showed great potential for dete cting viable oocysts with the successful run of multiple dilution series The results of these assa ys accurately detected the presence of viable oocysts and did not re port oocysts that were rendered nonviable. Assays of dilution series of viable oocysts indicated the method was able to detect low numbers of oocysts ranging from 1 to 4, maki ng real-time PCR a useful tool for testing potable water samples, as these ar e required to be oocyst free. The data generated by the 2x2 tables was used to determine the specificity, sensitivity, PPV and NPV of the PCR assay. The assay had a specificity of 100% and a NPV of 73% with field samples. This was an indication that the assay was successful at correctly identifying negative field samples and had a low tendency to produce false negative results. In other words, samples re ported as negative were true negatives and did not contain viable C. parvum oocysts. A sensitivity of 56% and PPV of 100% were determined for the PCR assay. Though the assa y was not capable of correctly identifying all samples positive for viable oocysts, the samples that were reported were true positive samples containing viable oocysts. Correlation studies were al so performed comparing the real-time PCR assay to the DAPI stained portion of the pellet. Thes e tests included the Kappa measure of agreement, YoudenÂ’s J, and PearsonÂ’s correl ation. The Kappa and YoudenÂ’s J values of 0.59 and 0.56, respectively, indicate there is ag reement between the two testing methods
57 and the results are not due to chance alone. A value of 0.64 was obtained for the PearsonÂ’s correlation. This i ndicates the assay does show moderate correlation with the DAPI results. Though there is agreement be tween the testing methods and a moderate correlation in the results, th e assay didnÂ’t perform well enough to be used on all water matrices. The statistical analysis may have been influenced by the high number of samples that inhibited the real-time PCR runs. A hi gh percentage (80%) of waste water effluent samples displayed inhibition to PCR. This caused many true positive samples to report as negative because the runs failed. This was also observed with the raw water and reclaimed water samples, having 70% and 52% inhibition, respectivel y. The levels of inhibition produced an interes ting correlation to current water treatment. Raw water showed a level of inhibition of 70%, sugge sting the inhibition was not caused by the treatment but by a impurity already in the wa ter supply. Effluent may show a high level of inhibition due to the high level of chemi cal and biological agents it contains. Since effluent may be further treated and purified to produce reclaimed water, it is logical for reclaimed water to have a lower inhibition that effluent. However, reclaimed water treatment is not as rigorous as potable water treatment, since the water is not used for human consumption. This may give insight as to why reclaimed water produced an inhibition level of 52%. Sin ce potable water has to endur e a more rigorous treatment procedure, the contaminants that were in the original raw water sample have been removed producing an inhibition of 0%.
58 Due to the high level of inhibition, it was necessary to determin e the true value of the PCR assay in the absence of inhibition. Samples that produced inhibition were removed from the original data set, and the runs were listed as failed. The total number of successful PCR runs was 18. The data was th en analyzed again using only values from the runs that produced a PCR result. Th is produced a noticeable difference in the performance ratings of the assay. The specifi city and PPV were unaffected and remained at 100%, however, the sensitivity increased to 89% and the NPV to 90% indicating a lower probability of reporting false negativ e results. The PearsonÂ’s correlation and Kappa values both increased to 0.89 indicati ng the results were not due to chance and there is a more substantial correlation between the two tests. This analysis shows the inhibition of the assay by matrix factors not removed in sample processing lead to the low proficiency of identifying positive samples and was not caused by a flaw in the PCR assay design. Focusing in particular on the necessa ry materials and the Â“hands-onÂ” time required for each method, that is, the amount of time a sample must be handled in some manner by the person executing the protocol, a cost comparison of Method 1623 to realtime PCR was completed. There was a marked decrease in the cost of performing realtime PCR on 8 samples as opposed to completing Method 1623. The cost associated with completion of Method 1623 on 8 samples is approximately $750.00. The EasySeed, used for positive control, and EasyStain, the fluor escent antibody stain, used in Method 1623 add substantially to the cost of Method 1623. In addition, on average a hands-on time of
59 9 hours is needed to complete the disassociati on step, staining, to ex amine the slides and report the results. In contrast, the real-time PCR assay w ould cost approximately $145.00 to test the same 8 samples. The major cost associated with this assay is attribut ed to the cost of the 96 well plates and covers; however it is offset by minimal hands-on labor of only 2 hours. The result is that the real-time PCR assay is much more cost eff ective than Method 1623 (Table 5). These results further show the potential for a real-time PCR assay in the absence of inhibition. The varying levels of inhibi tion may give some insi ght into the underlying cause, however further research should be pe rformed to determine the inhibiting factors and methods for their elimination before pe rforming the real-time PCR assay. The assay works well with potable water samples, produc ing no inhibition and accurate detection of viable C. parvum oocysts. In addition, the ability to test multiple samples in a single run paired with lower total cost make the assay more efficient to perform. Another important consideration in imple menting a new assay is the amount of time needed to execute the test and repor t the results. The cu rrent Method 1623 takes approximately 9 hours to test 8 samples, from disassociation of the beads to reading the slides and reporting the results. The hands -on time required to prepare 8 samples for real-time PCR is 2 hours. Since the PCR run is completed in 3 hours, a total of 5 hours is needed to complete the assay. This cuts the time needed for reporting results almost in half. Another benefit of the PCR assay is its ability to run multiple samples without
60 drastically increasing the turn around time. Nevertheless, for both methods, processing of the water filter through IMS ma y require 4 to 6 hours to perform. Conclusion The monitoring of environmental water sources is important in the implementation of control measures to prev ent outbreaks and protect the population from possible infection. EPA Method 1623 is a practical and effective method for the detection of Cryptosporidium in water samples; however it is costly to perform and requires substantial time to report results. It also does not have the ability to distinguish between various species of Cryptosporidium or give an accurate determination of viability. In summary, the objectives of this stu dy were achieved. The real-time PCR assay developed was validated with the use of pos itive and negative controls. The results of this assay are comparable to the DAPI re sults obtained using EPA Method 1623. The study determined that this real-time PCR assa y may be capable of providing detection of viable C. parvum oocysts in potable water samples in a more cost effective manner than Method 1623. The real-time PCR assay develope d in this study has the potential to be used on other types of water samples once the problem of inhibition is solved. The rapid detection of viable C. parvum in environmental water samples by realtime PCR would allow for a more accurate determin ation of the risk to public health. It would allow proper authorities to issue boil wa ter warnings and potentially decrease the risk of infection. This method may potentia lly be used to assess treatment methods for
61 use in reclaimed water. The increased us e of reclaimed water in residential and recreational areas raises the possible risk of individuals to become infected with C. parvum. The use of real-time PCR may more accurately assess this risk reclaimed water imposes on the public and prevent infection. This real-time PCR a ssay has the potential to allow for faster detection of viable C. parvum in environmental water samples which may aid in the prevention and c ontrol of future infections.
62 Table 5 Cost comparison of the continua tion of Method 1623 vs. the experimental real-time PCR assay for 8 sample Continuation of Method 1623 Experimental Real-time PCR Item Cost Item Cost 1.5 Microfuge Tubes $0.50 1.5 SafeLock Tubes $2.80 Cover slips $2.00 1.5 Microfuge Tubes $0.50 DAPI $0.20 0.6 Microfuge Tubes $0.60 Easy Seed $440.0096 Well MicroAmp Plate $30.00 EasyStain $74.40 Chelex 100 Resin $0.40 Well Slides $8.00 Liquid Nitrogen $2.00 Optical covers $14.00 One Step PCR Master Mix $41.60 Primers $0.60 Probe C. parvum oocysts $2.00 $0.80 Labor Hours @ $25.00/Hr $225.00 Labor Hours @ $25.00/Hr $50.00 Total:$750.10Total: $145.30
63 References Cited Ambion The RNA Company (2007). RT -PCR Basics. Available from URL http://www.ambion.com/techlib /basics/rtpcr/index.html AusVet (2007). Sample Size Ca lculator. Available from URL http://www.ausvet.com.au/conten t.php?page=res_software#freecalc Centers for Disease Control and Prevention [CDC] (2005). Cryptosporidiosis Fact Sheet. Available from URL http://www.cdc.gov/ncidod/dpd/parasites/cr yptosporidiosis/factsht_cryptosporidio sis.htm Environmental Protectio n Agency [EPA] (2001). Method 1623: Cryptosporidium and Giardia in Water by Filtration/IMS/ FA. Available from URL http://epa.gov/wate rscience/methods/1623.pdf Fontaine M and Guillot E (2003). An immunomagnetic separation-real-time PCR method for quantification of Cryptosporidium parvum in water samples Journal of Micro Methods. 54: 29-36. Gobet P and Toze S. (2001). Relevance of Cryptosporidium parvum hsp70 mRNA Amplification as a Tool to Discrimina te Between Viable and Dead Oocysts Journal of Para. 87.1: 226-229. Guerrant RL (1997). Synopses: Cryptospor idiosis: An Emerging, Highly Infectious Threat. Emerging Infectious Diseases 3.1: 51-57. Hannahs G, Kenyon College (unknown). Cryptosporidium parvum : an emerging pathogen. Hannahs Inc. Productions. Available from URL http://biology.kenyon.edu/slonc/b io38/hannahs/crypto.htm. Kaucner C and Stinear T (1998). Sens itive and Rapid Detection of Viable Giardia Cysts and Cryptosporidium parvum Oocysts in Large Volume Water Samples with Fiberglass Cartridge Filters and Reverse Transcription-PCR Applied and Environmental Micro. 64.5: 1743-1749.
64 Jenkins MC, Trout J, Abrahamsen MS, et al. (2000). Estimating viability of Cryptosporidium parvum oocysts using reverse tran scriptase-polymerase chain reaction (RT-PCR) directed at mRNA encoding amyloglucosidase Journal of Micro. Methods 48: 97-105. Leav BA, Mackay M, and Ward HD (2003). Cryptosporidium Species: New Insights and Old Challenges. Clinical Infectious Diseases. 36: 903-908. LeChevallier MW, Di Giovanni GD, Clan cy JL, et al. (2003) Comparison of Method 1623 and Cell Culture-PCR for Detection of Cryptosporidium spp in Source Waters Applied and Environmental Micro. 69.2: 971-979. Markell, John, and Krotoski (1999). Ma rkell and VogeÂ’s Me dical Parasitology, 8th ed. W.B. Saunders Company, Pennsylvania, USA. Polysciences Inc. (1999). Tec hnical Sheet 315. Available from URL http://www.polysciences.com/shop/pr oduct.asp?dept_id=99&pf_id=09224 Roberts L and Janovy J (2000). Gerald D. Schmidt & Larry S. RobertsÂ’ Foundations of Parasitology, 6th ed. McGraw Hill, Massachusetts, USA. Slifko TR, Huffman DE, and Rose JB (1999) A Most-Probable-Number Assay for the Enumeration of Infectious Cryptosporidium parvum Oocysts. Applied and Environmental Micro. 65.9: 3936-3941 Szklo M and Nieto F (2000). Epidemiology Beyond the Basics. Aspen Publishers Inc., Maryland, USA. Viessman W and Hammer M (1998). Wa ter Supply & Pollution Control, 6th ed. Addison Wesley Longman Inc., California, USA. Wagner-Wiening C and Kimmig P ( 1995). Detection of Viable Cryptosporidium parvum Oocysts by PCR Applied and Environmental Micro. 61.12: 4524-4516.
65 Bibliography Fontaine M and Guillot E (2002). Developm ent of a TaqMan quantitative PCR assay specific for Cryptosporidium parvum FEMS Microbiology Letters. 214: 13-17 Giovanni G and LeChevallier M (2004) Quantitative-PCR Assessment of Cryptosporidium parvum Cell Culture Infection. Applied and Environmental Micro. 71.3: 1495-1500 Hallier-Soulier S and Guillot E (2003). An Immunomagnetic separation-reverse transcription polymerase chain reaction (IMS-RT-PCR) test for sensitive and rapid detection of viable waterborne Cryptosporidium parvum Environmental Micro. 5.7: 592-598 Higgins J, Fayer R, Trout j, Xiao l, Lal A, Kerby S, and Jenkins M (2001). Real-Time PCR for the detection of Cryptosporidium parvum Journal of Microbiological Methods. 47: 323-337 Jenkins M, Trout J.M, Higgins J, Dorsch M, Veal D, and Fayer R (2003). Comparison of tests for viable and infectious Cryptosporidium parvum oocysts Parasitology Res. 89: 1-5 McCuin R and Clancy J (2003). Modificat ions to Untied States Environmental Protection Agency Methods 1622 and 1623 for Detection of Cryptosporidium Oocysts and Giardia Cysts in Water. Applied and Environmental Micro. 69.1: 267-274 Stinear T, Matusan A, Hines K, and Sandery M (1996). Detection of a Single Viable Cryptosporidium parvum Oocyst in Environmental Water Concentrates by Reverse Transcription-PCR. Applied and Environmental Micro. 5.7: 592-598 Widmer G, Orbacz E, and Tzipori S (1999). -Tubulin mRNA as a Marker of Cryptosporidium parvum Oocysts Viability. Applied and Environmental Micro. 65.4: 1584-1588 Wikipedia (2007). Cryptosporidium Available from URL http://en.wikipedia.org/wiki/Cryptosporidium