Microfluidic based Sample Preparation for Bloodstream Infections

Sahar Ardabili

Kungliga Tekniska högskolan, KTH

Royal Institute of Technology

School of Biotechnology

Stockholm, 2014

2

© Sahar Ardabili

Stockholm, 2014

Royal Institute of Technology

Science for Life Laboratories

SE-171 65 Solna

Printed by Universitetservice US-AB

Drottning Kristinas väg 53B

SE-100 44 Stockholm

Sweden

ISBN 978-91-7595-385-4

TRITA-BIO Report 2014:19

ISSN 1654-2312

3

To my parents

4

Abstract

Microfluidics promises to re-shape the current health-care

system by transferring diagnostic tools from central laboratories to close vicinity of the patient (point-of-care). One of the most important operational steps in any diagnostic platform is sample preparation, which is the main subject in this thesis. The goal of sample preparation is to isolate targets of interest from their surroundings. The work in this thesis is based on three ways to isolate bacteria: immune-based isolation, selective cell lysis, size-based separation.

The first sample-preparation approach uses antibodies against lipopolysaccharides (LPS), which are surface molecules found on all gram-negative bacteria. There are two characteristics that make this surface molecule interesting. First, it is highly abundant: one bacterium has approximately a million LPS molecules on its cell-wall. Second, the molecule has a conserved region within all gram-negative bacteria, so using one affinity molecule to isolate disease-causing gram-negative bacteria is an attractive option, particularly from the point of view of sample preparation. The main challenge, however, is antigen accessibility. To address this, we have developed a treatment protocol that improves the capturing efficiency.

The strategy behind selective cell lysis takes advantage of the differences between the blood-cell membrane and the bacterial cell-wall. These fundamental differences make it possible to lyse (destroy) blood-cells selectively while keeping the target of interest, here the bacteria, intact and, what is more important alive. Viability plays an important role in determining antibiotic susceptibility.

Difference in size is another well-used characteristic for sample- separation. Inertial microfluidics can focus size-dependent particle at high flow-rates. Thus, particles of 10 µm diameter were positioned in precise streamlines within a curved channel. The focused particles can then be collected at defined outlets. This approach was then used to isolate white blood cells, which account for approximately 1% of the whole blood. In such a device particles of 2µm diameter (size of bacteria) would not be focused and thereby present at every outlet. To separate bacteria from blood elasto-inertial microfluidics was used. Here, e blood components are diverted to center of the channels while smaller bacteria remain in the side streams and can subsequently be separated

5

Populärvetenskaplig sammanfattning

Blodförgiftning (sepsis) är en livshotande sjukdom som årligen

drabbar omkring 15-19 millioner människor globalt[1–3]. Den bakomliggande anledningen är oftast en bakterieinfektion, men blodförgiftning kan även orsakas av virus, parasiter eller svampinfektioner[4]. I Sverige är blodförgiftning den 13:e mest resurskrävande sjukdomen inom slutenvården[5]. Sepsis är en följd av en så kallad systemisk inflammationsrespons (SIRS) orsakat av vårt eget immunförsvar som svar på invaderande patogener (sjukdomsorsakande organismer). Detta är ett exempel på hur vårt immunsystem, som i vanliga fall ska skydda oss från faroämnen, kan orsaka mer skada än nytta. Om sjukdomen lämnas obehandlad kan det på sikt leda till cirkulationskollaps, multiorgansvikt och död[6]. Ett snabb omhändertagande med rätt antibiotikabehandling räddar liv. Därför ges en kombination av olika antibiotika omgående om en patient misstänks ha drabbats av sepsis[6]. Idag används tekniken blododling för isolering och identifiering av den invaderande bakterien, samt fastställande av eventuell antibiotikaresistens. Nackdelen med denna teknik är att svarsresultat kan dröja upp till 72 timmar och då är det ofta för sent[7–11]. Läkare behandlar i blindo då situationen lätt kan bli livshotande. Behovet för en diagnostisk plattform med snabb patogenidentifiering är därmed stort.

Målet med studierna som föregått denna avhandling är prov-preparing med hjälp av mikrofluidik. Mikrofluidik är ett interdisciplinärt forskningsfält där mikrosystemteknik, fysik och bioteknik möts för att skapa ett system i miniatyr där diverse laboratorieanalyser kan utföras. Ett stort fokus har lagts på bakterieisolering med förhoppningen om att detta ska ta oss ett steg närmare förbättrade diagnostiska verktyg för sjukdomar som sepsis.

Avhandlingen är uppdelad i 4 kapitel. Kapitel 1 ger en sammanställning över motiveringen bakom denna forskning samtidigt som den beskriver de diagnostiska verktyg som finns tillgängliga på marknaden idag. Kapitel 2 redogör för framstegen som gjorts inom mikrofluidik för att isolera celler från komplexa lösningar så som blod. Kapitel 3 beskriver det eftersträvade slutmålet med forskningen, och beskriver koncepten ”point-of-care” och ”lab-on-a-chip”. Slutligen

6

redogör Kapitel 4 för den bakomliggande forskningen som jag har utfört tillsammans med mina kollegor under min forskningsutbildning. Avslutningsvis skulle jag vilja poängtera att bakterier inte är ondskan inkarnerad. Vår kropp innehåller ungefär 10 gånger så många bakterieceller än av våra egna celler[12]. Dessa bakterier hjälper oss på olika sätt, t.ex. med matsmältning, vitaminproduktion, och immunförsvar[12]. Vår kropp kommer bara till skada när de hamnar vid fel plats.

7

Numinous - English, (adj) - "describing an experience that makes you fearful yet fascinated, awed yet attracted; the powerful, personal feeling of being overwhelmed and inspired.

8

List of publications

Ardabili, S., Zelenin, S., Ramachandraiah, H., Russom, A. Epitope unmasking for improved immuno-magnetic isolation of Gram-negative bacteria. Manuscript

Zelenin, S., Hansson, J., Ardabili, S., Ramachandraiah, H., Brismar, H., and Russom, A. Microfluidic-based isolation of bacteria from whole blood for sepsis diagnostics. Biotechnology Letters, 2014, DOI: 10.1007/s10529-014-1734-8

Ramachandraiah, H.*, Ardabili, S.*, Faridi, A. M., Gantelius, J., Kowalewski, J. M., Mårtensson, G., & Russom, A. Dean flow-coupled inertial focusing in curved channels. Biomicrofluidics, 2014, 8(3), 034117.

Faridi, A.M., Ramachandraiah, H., Ardabili, S., Zelenin, S., and Aman Russom, Elasto-Inertial microfluidics for bacteria separation from whole blood for sepsis diagnostics. Manuscript

Pavankumar, A.S.*, Ardabili, S*, Zelenin, S., Shulte, T., Lundin, A. and Russom, A. Recombinant Shigella flexneri apyrase enzyme for bioluminescence based diagnostic applications Manuscript

* Authors contributed equally.

All papers reproduced with permission of the copyright holders.

9

Contribution to the papers

Paper I:

Major parts of the experiments and writing.

Paper II:

Minor parts of the experiment and writing

Paper III:

Major parts of the experiments. Minor parts of the writing

Paper IV:

Minor parts of the experiments and writing

Paper V:

Major parts of the experiments. Minor parts of the writing

10

TABLE OF CONTENT

Abstract ............................................................................................. 4 Populärvetenskaplig sammanfattning ......................... 5 List of publications ...................................................... 8

Contribution to the papers ................................................................ 9

Thesis Road Map ............................................................................. 12

Bloodstream infection ..................................................................... 13 Infectious disease ...................................................... 14 Sepsis ........................................................................ 16

Epidemiology .......................................................................... 17 Misdiagnosis: a fatal error. ..................................................... 18

Current diagnostic assays .......................................... 19 Nucleic acid-based techniques ............................................... 20 Positive blood culture ............................................................. 21 Diagnosis directly from blood ................................................ 22

Isolation techniques for complex fluids ......................................... 25 The Challenge: Taking Blood Apart ........................... 26 Microfluidics – A Laboratory Time Saver? ................ 27 Microfluidic-Based Separation .................................. 30 Cell-Wall Composition ............................................... 31

Chemical approach ................................................................. 34 Affinity-Based Approaches .................................................... 34 Size-Based Approaches .......................................................... 36 Inertial Microfluidics ............................................................. 36

Point-of-Care: The Final Goal ......................................................... 37 Point-of-Care: An Overview ....................................... 38 Operational Steps within Point-of-Care ..................... 39 Point-of-Care for Bacterial Identification .................. 40

Verigene .................................................................................. 40 Film Array ............................................................................... 41 Concluding Remarks .............................................................. 42

11

Present investigation ...................................................................... 43 Aim of the Thesis ....................................................... 44 Paper I ....................................................................... 46

Paper II……………………………………………………………….50 Paper III .................................................................... 54 Paper IV .................................................................... 57 Paper V ...................................................................... 60

Conclusion and Future Work ......................................................... 63

Acknowledgement .......................................................................... 66

Abbreviation ................................................................................... 69  

12

Thesis Road Map

This thesis focuses on sample preparation with an emphasis on microfluidics. The objective is to apply such strategies to the development of diagnostics for infectious disease.

There are four chapters of which Chapters 1-3 are introductory. Chapter 1 (Blood Stream Infection) gives an overview of the motivation behind our research as well as a brief review of the diagnostic tools available on the market today. Chapter 2 reviews the work in microfluidics to isolate cells from complex fluids. Chapter 3 presents the concept of point-of-care and lab-on-a-chip: the intention is, in principle, to miniaturize a full-scale laboratory onto a tiny chip for integrated bioassays. Finally Chapter 4 presents my work during my years as a PhD student.

13

CHAPTER 1

Bloodstream infection

“Our arsenals for fighting off bacteria are so powerful, and involve so many different defense mechanisms, that we are in more danger from them than from the invaders. We live in the midst of explosive devices; we are mined. It is the information carried by the bacteria that we cannot abide. The Gram-negative bacteria are the best examples of this. They display lipopolysaccharide endotoxin in their walls, and these macromolecules are read by our tissues as the very worst of bad news. When we sense lipopolysaccharide, we are likely to turn on every defense at our disposal; we will bomb, defoliate, blockade, seal off, and destroy all the tissues in the area.

- Thomas Lewis (The Lives of a Cell: Notes from a Biology Watcher) [1]

14

Infectious disease

The unwanted presence of multiplying pathogens in our bodies can lead to a range of infectious diseases. The infection-causing pathogens may be viruses, bacteria, fungi, protozoa, parasites or prions [2]. Infectious disease affects a vast number of people world-wide. According to a report from the World Health Organization in 2011, infectious diseases such as lower respiratory infections, HIV/AIDS, diarrheal diseases, malaria and tuberculosis are the leading causes of death in low-income countries. In high-income countries, on the other hand, only one in ten deaths are caused by infectious disease [3] (Figure 1.1).

Figure 1.1: Leading causes of deaths world-wide in 2011. Chart adapted from WHO- fact sheet. [3].

Nevertheless, health-care associated infection in the intensive care units (ICU) remains a global problem and is associated with high mortality and costs [4–8]. The risk of acquiring an infection is large even in high-income countries. An 2009 ICU study covering 75 countries world-wide with data collected from 13,796 patients on one single day, reported that 51% of all of the patients were considered infected [4]. Within the infected cohort, only 70% had a culture-positive result, but antibiotics were administered to 71% of all the patients [4]. An earlier,

15

one-day study from 1992 with data from 17 European countries showed a similar percentage of infected patients (45%) [8].

It should not be surprising that the risk of acquiring an infection in hospital settings is high, especially in view of increasing number of entry points created by invasive procedures. Incorrect antibiotics for bacterial infections can have dire life-threatening consequences [9]. The number of patients receiving inappropriate antibiotics has been estimated to be 20-30% [10]. The level of antibiotic resistance has steadily increased over the years while the production of new effective antibiotics has decreased. Together this constitutes an alarming scenario in a health-care era in which diseases that could have been easily treated in the past can now have mortal outcomes [11–13].

The yearly incidence of sepsis, the systemic inflammatory response to an infection, is approximately 18 million people worldwide [14–17]. This would correspond to the total population of Denmark, Finland, Norway, and Slovenia World bank figures from 2012 [18]. According to the Society of Critical Care Medicine, sepsis is the second leading cause of death in non-coronary ICUs in the USA [19–21].The mortality rate of sepsis is estimated to be between 20% and 80% [14,21–23]. These studies demonstrate the importance of infection control, which can be achieved by improving hospital guidelines, providing new medicines; developing better strategies for diagnostics and simply by providing diagnostic tools applicable in resource limited areas.

In addition to clinical settings pathogen detection is of great interest and importance in many fields, such as the food industry and in water and environment health and safety (Figure 1.2). The present thesis is primarily concerned with pathogen (bacteria) detection for clinical use. Techniques for bacteria isolation (sample preparation) have been given particular importance here.

16

Figure 1.2: The relative amount of literature in specific areas. Chart adapted from Lazcka et al 2007 [24].

Sepsis

Sepsis, severe sepsis and septic shock constitute thee systemic inflammatory response syndrome (SIRS) of infectious origin with escalating severity, symptoms and signs such as organ dysfunction and hypotension [15,22,25–28]. The systemic inflammatory response is part of the body’s defense mechanism against harmful invasions. However, in the case of sepsis, severe sepsis and septic shock, this response has gone awry and causes more harm than good. Infection is not the sole cause of SIRS [25,26] (Figure 1.3). It can be triggered by a range of events such as trauma, burns or pancreatitis. Additionally, Figure 1.3 sows that the causative agent in sepsis is not always a bacteria but could also be fungi, parasites or viruses [25]. Nonetheless, as the Figure 1.3 indicates, bacteria (bacteremia) are the leading cause [21].

SIRS patients may display a range of different clinical manifestations such as; fever (> 38°C), rapid heart rate (> 90 beats/min), hyperventilation and changes in white blood-cell counts [22,25,28,29]. Two or more of these signs are needed to fulfill the criteria for SIRS. In order to confirm that the underlying cause is actually an infection, further investigation is required [28,29]. Because of the seriousness of the condition, doctors administer broad-spectrum antibiotics immediately when sepsis is suspected. There is simply no time to wait for laboratory results, which can take up to 72 hours before the complete picture, including a potential antibiotic-resistance profile are available [30–34].

17

Epidemiology In the United States, sepsis is rated as the 10th leading cause of

death [15,21,26]. A 22-year American study showed that the number of sepsis patients increased from 82 to 240 per 100,000 in 1979 and 2000. Even though the overall mortality rate had decreased from 28% to 18%, the total number of deaths was three times higher due to an overall increase of incidence [21]. Globally, sepsis is estimated to affect between 15 and 19 million people every year [15–17]. In spite of having a quite high disease burden (e.g. incidence, mortality and cost), sepsis still not attract as much public attention as do diseases such as breast cancer and AIDS [16,35–38].

Figure1.3: The relationship between SIRS, sepsis and infection. Sepsis is a systemic inflammatory response syndrome (SIRS) caused by an infection. An infection may have several different origins: bacteremia, fungemia, parasitemia and viremia. There are several medical scenarios other than sepsis in which SIRS can appear. This image is reproduced with the permission of the copyright holder [25].

According to the American Center of Disease Control and Prevention (CDC), the mortality rate per 100,000 populations in 2010 was 25.9 for breast cancer, 2.5 for AIDS and 41.4 for stroke [39–41]. If we compare these numbers with the equivalent values for sepsis mortality, one can easily see why sepsis is one of the ten leading cause of death in the US and the 2nd leading cause of death in non-coronary ICU [21,26,30,35,42–44]. Martin et al. estimated the overall death rate due to

18

sepsis at 43.9 per 100,000 for the year 2000 [21], while Wang et al. estimated a mortality rate of 65.5 in a study ranging between 1999 and 2005 [45]. In contrast, Melamed et al. estimated 50.5 deaths per 100, 000 for the same period (1999-2005). Daniels et al. made a similar comparison between sepsis and diseases with high public awareness in UK (Figure 1.4) [37]. The incidence of sepsis in the European Union was estimated to be 90.4 deaths per 100, 000. In comparison the incidence of breast cancer that was determined to be 58 per 100,000 [37]. The occurrence of severe sepsis in Europe, on the other hand, seems to lie between 50 and 100 cases per 100,000 individuals [35,36,46–51]. These differences can be attributed in part to seasonal variations, variations in the length of study, and variations in the diagnostic criteria [42].

Although it might be difficult in some cases to determine the true sepsis incidence/mortality rates, these studies still confirm its place among the common causes of deaths worldwide. Sepsis deserves increased attention equal to AIDS, prostate cancer, breast cancer and other better-known conditions.

Figure1.4: Mortality rate for various diseases in the UK This image is reproduced with the permission of the copyright holder [37]

Misdiagnosis: a fatal error. As mentioned earlier, the definition of sepsis is quite broad and to

some extent overlaps with other diseases. This definition was established as recently as in 1992 by the ACCP/SCCM conference committee [25].

19

But does it help or is there still a lot of confusion? One thing that is absolutely certain and is agreed by all is that speed saves lives: the sooner a potential sepsis patient is identified the better the survival chances. So how is this translated into the clinics and hospitals around the world? An interesting survey performed by Poeze et al. looked into the perception, attitude and the ability of healthcare professionals around the world to diagnose sepsis [52]. The great majority, 86% of the physicians were of the opinion that sepsis could easily be misdiagnosed and 65% believed physical examination to be inadequate. Only a small fraction of the participants (22%) used the ACCP/SCCM criteria to define sepsis and this eight years after the definition was set [52]. The problem with the criteria is that they have high sensitivity but low specificity. A vast majority of the patients in ICUs and general wards fulfilled these criteria at some point [53]. A new attempt was made in 2001 by the ACCP/SCCM conference to further improve the sepsis criteria [54]. But it does not seem to have had the desired effect. A comparative study showed little difference between the 1991 and the 2001 criteria [53]. Indeed, much depends on the clinician’s ability to predict sepsis. Consequently, the confusion around sepsis criteria is worrisome, since any delay in treatment can severely reduce the chances of survival [30,55–60]. Kumar et al. have shown that chances of surviving septic shock decreased by 8% for every hour the correct treatment is delayed [58].

Current diagnostic assays

Although today’s microbiological gold standard, blood culturing,

is highly affected by external technical factors (proper skin preparation, sample volume, transport time, incubation atmosphere, blood-to-broth ratio, culture media and so on), no method has yet been able to replace it fully [30,61–66]. The main challenge of microbiological diagnostics is the low bacteria load in the sample. As the clinical signs become manifest, the blood stream still contains as few as 1-30 colony-forming units (CFU) per ml [33]. The bacteria load might increase to 1000 CFU/mL, but this is encountered only in severely ill patients [67]. According to Towns et al., almost 50% of all patients have less than 1 CFU/mL [10]. Hence, sufficient sample volume is a highly important parameter.

The ability of blood culturing to detect only viable bacteria can be seen as both strength and a weakness. Blood contains antimicrobial

20

agents that inhibit bacterial growth. This inhibition is further increased in patients with antibiotics already in their systems, which give rise to false negative results [68]. Although the sensitivity of blood culture is considered to be 1 CFU/mL, only a third to a half of all sampled septic patients yield positive cultures [10,31,55,69–71]. The difference in yield is directly associated with the factors cited above, but the major drawback is the time it takes to grow enough cells needed for analysis (up to 72 hours). However, classic microbiological methods are far superior when it comes to determine antibiotic-resistance profiles. Molecular methods will give a yes/no answer based on already known resistance mechanisms/genes. However, the absence of a resistance gene will not necessarily mean the organism is susceptible to a particular antibiotic. There is seldom a situation where a single gene can give rise to resistance, since phenotypic resistance to a certain antibiotic may be caused by a whole array of different genes. For example, the resistance to beta-lactams among the Enterobacteriaceae family has been attributed to several hundred mechanisms [72]. False positive signals will occur for silent genes or pseudo-genes since resistance is dependent not only on the presence of the gene but also in its expression level. Furthermore, genetic methods will not give any information regarding the minimal inhibitory concentration (MIC). The occurrence of false negatives is also a possibility as in the case of primer binding-site mutations [73–76]. There is also the barrier of sample preparation for complex samples, which might contain assay inhibitors. However, molecular methods open up the possibility of circumventing time-consuming culturing thereby decreasing the turnaround time. They may also be very advantageous when it comes to slow-growing and fastidious organisms since they can hardly be detected with today’s gold-standard method, thus giving rise to false negative results. Even though the idea of a universal molecular-based method for detecting all bacteria with all possible combinations of antibiotic resistance may well be overly ambitious, developing a rapid test for a few clinically relevant strains is not. Even rapid gram determination would be useful in clinical settings, as it could narrow down the antibiotic spectrum.

Nucleic acid-based techniques A number of the nucleic-acid-based techniques (NAT) for sepsis

diagnosis are available on the market today. These methods can be divided into two groups: techniques that involve cell enumeration

21

(positive cell culture) and techniques that directly use blood as a sample. Nucleic-acid-based techniques can be sub-divided in function of their analysis: pathogen-specific assays, universal broad-range assays and multiplex assays [30,33]. There are a number of parameters to take into consideration when evaluating these tools: the actual hands-on time/number of assay steps, assay performance when it comes to poly-microbial samples, the total turnaround time, sensitivity, and specificity. Particularly in the context of sepsis, however the most important parameter is the extent of its diagnostic spectrum. A common barrier, regardless of the nucleic-acid-based approach, is sample preparation (which will be discussed in more detail in the following chapter), and a common limitation is the antibiotic-resistance profile. This is also the reason why molecular methods are seen as a compliment to the gold standard.

Positive blood culture While all of the techniques discussed in this section require a pre-

culturing step, they differ in their diagnostic spectrum, turnaround time and actual hands-on time. As can be seen in Table 1.1, these assay methods have their strengths but, as yet, none provide a complete all-in-one-tool. One must compromise between the number of target species, the assay time, and the number of resistance markers. Two of these methods (Verigene and Filmarray) provide a closed-box system with minimal hands on time (5 minutes), which is basically what one looks for in a new diagnostic tool, but there is still room for improvement when it comes to the diagnostic spectrum.

22

Table 1: 1 Techniques that require positive blood culture.

Method Analysis Multiplex Nr. of pathogens

Resistance genes

Hands- on time

Assay time

Prove-it sepsis

[7,9,61,77]

Multiplex PCR,

microarray colorimetric

read-out

Yes 74 3 90 min 3-3.5h

Film Array [9,78]

PCR based No 24 3 5 min 1h

Verigene [9,71,79]

Microarray Optical

detection Yes 13 gram+

5 gram - 3 5 min 2-2.5h

Diagnosis directly from blood It would be highly advantageous to eliminate the blood-culturing

step and identify the disease causing pathogen directly from blood. Consequently, the turnaround time could be drastically reduced and specific antibiotic therapy could be administered. The first attempt to isolate bacteria from blood was made in 1993 [77]. There are a number of assays available on the market that offers pathogen detection/isolation directly from blood: Septifast (Roche), Septites (Molezym), MagicPlex Sepsis Real-Time Test (Seegene), Vyoo (Sirs lab) and Polaris (Biocartis) are a few examples. SeptiTest from Roche has been available on the market since 2004 and detects 19 of the most common bacteria and six of the most common fungi [78]. The newest addition is Polaris from Biocartis, a platform that is currently under development. One of the interesting aspects of Polaris is that it selectively removes human DNA (deoxyribonucleic acid) before the pathogen lysis takes place. The selective removal of contaminating human DNA or enrichment of bacterial DNA prior analysis is a strategy used not only by Polaris (Biocartis). Both Molzyme and VYOO (SIRS lab) remove human DNA or, as in the case of VYOO, enrich bacterial DNA by using an affinity chromatography (PureProve) [30,79,80]. PureProve (SIRS lab) uses characteristic motifs of prokaryotic DNA, non-methylated CpG (cytidylatephosphate-deoxyguanylate) motifs, to bind bacterial DNA in a resin while human DNA is washed away [80]. In a study by Loonen et al., the Polaris chemical-based method is compared to Molzyms’ (MolYsis) enzymatic removal of human DNA [81]. To my knowledge no other reports have been published about the Polaris platform. An interesting

23

study that was recently published used the DNA-binding section of a bacterial topoisomerase (Gyrase) to selectively isolate prokaryotic DNA [82].

Table 1.2 summarizes the differences between these assays. Two things are striking in this table: the first is that all of the methods use PCR (polymerase chain reaction) to some extent; the second is that the size of the sample volume (1-5mL) is very low compared to the gold standard (20-30 mL) [33,83]. According to Jordana-Lluch et al., low volumes are a necessity when working directly with blood because of the presence of large amounts of human DNA, which may hinder detection or inhibit the PCR reaction. Moreover, there are several natural components in blood besides leucocyte DNA that might reduce the PCR capacity, one such being hemoglobin [83–85]. It has also been shown that even the presence of immunoglobulin G (igG) could have possible inhibitory effects [84–86]. The list of inhibitors makes the sample preparation especially important for it concerns not only isolating the pathogen but also to getting rid of any inhibitory substances. Another component affecting the assay outcome, apart from the various inhibitory substances, is the choice of DNA polymerase (a crucial component in PCR). DNA-polymerases differ in their capacity to withstand the presence of inhibitors in blood. Al-Soud et al. found that AmpliTaq Gold was highly sensitive to the presence of blood. An amount of 0.004% (vol/vol) blood resulted in complete inhibition [84]. Other polymerases (HotTub, Pwo, rTfl and Tli ) could tolerate up to 20% (vol/vol) blood [84]. Human DNA is not the only negative aspect when working directly with blood. Nucleic acid-based tests (NAT) risk producing medically irrelevant findings due to the presence of circulating bacterial DNA, transient bacteremia and dead bacteria

Recently, Laakso and Mäki have shown that the Prove-It sepsis technology from Mobidiag could be used to analyze 1 mL of spiked whole-blood samples. Together with two other technologies, the SelectNA blood-pathogen isolation kit (Molzyme) and the Nordiag Arrow automated extraction device, they were able to detect a number of different bacterial species with a detection range of 11-600 CFU/mL (depending on the species) [87].

24

Table 1.2: Diagnostics directly from blood

Method Volume Analysis LOD CFU/mL

Nr. of Pathogens TAT

SeptiFast (Roche) 1.5mL PCR, Melting curve analysis 3-30 25 6h

SeptiTest (Molzym)

[9,83]

1-5mL PCR, Sequencing N/A >300 6-

12

VYOO/LOOXTER (SIRS lab)

[7,9,77,81,92 5mL PCR, gel-

electrophoresis 3-10 ~40 6-8h

MagicPlex (Segeene)

[77,83,85,93] 1mL 3x PCR reactions

required N/A 27 6h

Polaris (Biocartis) [83,94] 1-5mL PCR, real time

fluorescence N/A N/A N/A

25

CHAPTER 2

Isolation techniques for complex fluids

“Blood is a treasure of information about the functioning of the whole body. Every minute, the entire blood volume is recirculated throughout the body, delivering oxygen and nutrients to every cell and transporting products from and toward all different tissues. At the same time, cells of the immune system are transported quickly and efficiently through blood, to and from every place in the body where they perform specific immuno-surveillance functions. As a result, blood harbors a massive amount of information about the functioning of all tissues and organs in the body”.

- Mehmet Toner and Daniel Irima (from Blood on a Chip) [88].

26

The Challenge: Taking Blood Apart

Toner and Irima’s observation above describes the importance of

blood and the potential amount of information hidden within it. However, accessing that information is challenging, the main one being the complexity of blood itself relative to the low number of target molecules (Table 2.1). One milliliter (mL) of blood contains 5 × 109 red blood cells (RBC), 2-5 × 108 platelets, and 5 to 10 × 106 white blood cells (WBC) [88], while clinically interesting samples often have very low target concentrations. One such example, besides bacteria and blood-stream infection, are circulating tumor cells (CTC), which are of great interest within cancer research, but the number of target cells could be as few as 10 cells/mL blood [89]. Another example are basophils in allergology, which are mast cells that comprise less than 1% of the total number of white blood cells [90] These two examples have one thing in common: the low number of target cells in a sea of different blood cells so sample preparation plays a significant role in each application area. Sample preparation has often been described as the “bottleneck” or the “road-block” for new technologies as well as the “forgotten beginning”.

Table 2.1: Number of cells found in mL of blood.

Cell-type/Target molecule Cells/mL Ref

RBC 5 × 109 [88] WBC 5 to10 × 106 [88]

Platelets 2-5 × 108 [88]

Basophil 103-105 [88]

Circulating tumor cells (CTC) 0,1-10 [89]

Bacteria (Blood stream infection) 1-1000 CFU /mL

The goal of sample preparation is to separate target cells from their surroundings in order to remove inhibitory substances that may hamper downstream analysis while reducing the heterogeneity, the complexity, of the sample itself [91]. Microfluidics brings promises to re-

27

shape the current health-care system by transferring diagnostic tools from central laboratories to the vicinity of patient. To achieve this, sample preparation must be included in the workflow, which would be an improvement over the current situation in microfluidics in which “of-chip” macro-scale solutions are often the only ones available [92]. The ideal scenario would be a “plug-and-play” system, an objective that many researchers are striving to achieve [31,93].

Microfluidics – A Laboratory Time Saver?

As the word implies, microfluidics is about handling and manipulating fluids in micro-scale dimensions. Here, micro technology, engineering, physics, chemistry and biotechnology overlap. Microfluidic aims to replace tedious laboratory work that often requires repeated pipetting by one single automated closed box or hand-held device. Other important objectives are high-throughput and multiplexing. Imagine the advantage of screening multiple targets and running numerous tests simultaneously. Throughput is defined as the number of assays a system can perform during a certain period of time [94]. In the macro-world, the ability to run 96 or 384 samples at once is considered high-throughput. Now, with the aid of microfluidics, throughput can be moved beyond the micro-plate. High-throughput microfluidics can be achieved serially or in parallel. With parallel processing, a high number of samples can be run simultaneously in order either to reduce the overall processing time or to increase multiplexing by running different assays simultaneously. Serially, throughput is achieved by taking a sample from processing to analysis, that is, different steps and functionalities are integrated serially in one setup. This is usually the case in Point-of-Care systems (Chapter 3).

The economic benefits of miniaturization are obvious: smaller size would inevitably mean less reagent consumption, less waste and less manufacturing cost [95–98]. When moving into the sub-millimeter scale, there are certain characteristics that become prominent, one such being laminar flow. Although Figure 2.1 is not a micro-scale example, it gives quite a striking image of how two stream lines in a laminar-flow condition would look. In such a system, mixing is caused primarily by diffusion.

28

Figure 2.1: An example of how a laminar flow streamline would look like in a microfluidic device. Two parallel streams flow without mixing. Photograph taken by Jozef Kowalewski.

When designing physical systems that involve fluids it is important to be able to predict the behavior of the flow. The

dimensionless Reynolds number (Re = ρvd/µ) describes whether or not a flow can be considered turbulent or laminar. The Reynolds number is in essence the ratio between inertial to viscous forces [99]. A high Reynolds number (>4000) indicates turbulent flow, while a low Reynolds number (<2300) gives laminar flow. The intermediate values (2300<Re<4000) indicate flow with both the laminar and turbulent flow regimes present [100]. In the laminar regime, the flow will have a parabolic profile. Our blood vessels are one example where this occurs. Here, layers of the blood cells travel parallel to the vessel wall in an orderly fashion (Figure 2.2) with no disturbance between the streamlines [101]. In contrast, in situations where turbulent regimes take place, there will be swirling motions and more unpredictable mixing. High flow-rates will generally result in turbulent flows [102].

29

Figure 2.2 Parabolic flow profile in a blood vessel. Blood travels parallel with no disturbances between the streamlines (arrows). The maximum velocity (Vmax) occurs at the center line, and the lowest velocity is found by the vessel wall (V=0).

Smaller reaction chambers give shorter diffusion paths, which in turn enable more rapid reactions than their macro-scale counterparts can provide [103]. One example of this is conventional enzyme-linked immunosorbent assay (ELISA), which is typically performed in micro-titer plates with mm-scale diffusion distances. As a consequence, an ELISA can take from several hours up to 2 days to complete. Studies have shown that a microfluidic setup can reduce the assay time from hours to minutes [103–106]. A high surface-to-volume ratio is another scaling-down effect that could be advantageous, especially for surface-bound affinity assays [107,108].

Low cost, fast reactions, high-surface-to-volume ratio, multiplexing, high-throughput, and automation are keywords that are usually positively associated with microfluidics. However, there are also challenges when working in micro-scale dimensions (Table 2.1). Depending on the objective, a laminar flow regime can be strength or a weakness. On the positive side, it provides more precise placement of particles and reagents, thus making multiplexed chemical dilutions possible [97]. On the downside however, mixing becomes challenging as the driving force is mainly diffusion-based and occurs at stream-line interfaces [97,109]. The disadvantage of a high surface-to-volume ratio is that there are more surfaces available for adsorption, which leads the discussion to the choice of material. The material most often used in the field of microfluidics is polydimethylsiloxane (PDMS). One of its advantages is biocompatibility so it has been used in catheters, drainage tubing, pacemakers, prostheses and various implants (blood vessels, heart valves, breast implants) [110]. Biocompatibility is to a large degree

30

determined by cellular responses, which in turn are based on their reaction to adsorbed proteins on biomaterial surfaces. When an implant is made, the very first event is protein adsorption. This takes place within seconds of implantation, thereby making the biomaterial a biologically active surface. Therefore, cells in our bodies do not encounter the biomaterial itself but the proteins adsorbed on its surface [110]. PDMS is hydrophobic in its nature. To avoid non-specific adsorption, various surface modifications are needed [104], which is particularly important when working with blood [111]. Blood plasma encompasses a myriad of proteins readily adsorbed on surfaces. Among these proteins there are some (fibrinogen, fibrnoectin, vitronectin, von Willebrand factor, etc.) that induce platelet adhesion. Surface modification may help evade blood-cell activation, platelet adhesion, and coagulation upon blood-surface contact [110,112]. Such modification may involve heparin or poly (ethylene glycol) (PEG).

Table 2.2: Pros and cons of using microfluidics.

Pros Cons Multiplexing Mixing High-surface to volume ratio Non-specific adsorption Automation Interfacing Faster reaction Clogging Fluid control Bubbles Low cost

Microfluidic-Based Separation

Traditionally in microfluidics, separation has been divided into

active and passive forms. Simply put, an active separation requires an external force while a passive separation relies more on channel geometry and inherent forces [113–115]. Table 2.3 gives an overview of the different active and passive separation methods as well as the separation criteria used in them. These methods are beyond the scope of this thesis. For further reading see the suggested references [113,115–118].

Different cell characteristics are often used to differentiate the target from its surroundings: size, density, deformability, surface antigen,

31

surface charge and cell-wall composition. In the following section, the focus will be on cell-wall composition and size.

Table 2.3: Active and passive microfluidic separation

Category Method Separation criteria

Ref

Active separation

Acoustophoresis size, density, compressibility

[113,116–119]

Dielectrophoresis Surface charge, density, size

[113,117,118]

Electrophoresis Surface charge [113] Optical Size, refractive

index, polarizability

[113,118]

Hydrodynamic Size [113,115] Magnetic Magnetic

susceptibility [116]

Passive Separation

Deterministic lateral displacement

Size, deformability

[113,114,116]

Inertial Size [113,116,120,121] Filtration Size [113,116]

Cell-Wall Composition

Exploiting differences in cell-wall composition is, perhaps, one of

the first approaches that come to mind when separating targets from their surroundings. Antibody-based systems would be one of the most obvious implementations. Antibodies are proteins that are part of the body’s defense system as well as being invaluable tools in today’s modern diagnostic toolbox [122–125]. These proteins bind targets called antigens with high specificity (“lock and key”) and sensitivity even in complex solutions [105]. Assays that involve antibodies are called immunoassays. There is a large selection of different types of immunoassays (e.g.,enzyme-linked,fluorescent-based, chemiluminescent-based and radio immunoassays). Perhaps the most important aspect of antibodies is the ability to custom-make them against almost anything with high affinity. Ever since they were introduced in the late 1950s, they have

32

become standardized tools, and there are a great variety of them [126]. You can find them coupled to fluorophores, enzymes, quantum dots, and beads of different sizes and materials. It is, therefore, not surprising that the annual sales of immunoassay material have been estimated to be as high as $7.2 billion worldwide [123].

A less obvious tactic when using differences in cell-wall

composition is a chemical approach (selective lysis) that takes the whole cell-wall into account as opposed to only certain surface molecules. This method takes advantage of the large difference between the protective cell barrier of bacteria and blood cells. First, blood cells lack a cell-wall. The enclosing barrier of a blood cell is a cell membrane. A fundamental function of the cell membrane is to segregate the liquid interior of the cell from the watery environment outside the cell [127]. However, since bacteria often live in harsh conditions, they need the extra protection against the environment that comes in the form of a cell-wall [128]. For instance, Escherichia coli can be found in the mammalian gut and Salmonella in the gall bladder [128]. Here, the bacteria must be able to tolerate detergents such as bile salts and gastric juices.

This difference in cell-wall/cell-membrane composition has been used in both macro- and micro-setups. In addition to centrifugation as a means of blood fractionation, there are also chemical methods that selectively lyse red blood cells, a process that typically involves ammonium chloride [88,129]. This process is called hemolysis and has been used extensively in the study of white blood cells (WBC). Here, cell separation/isolation is necessary since the WBC make up only less than 1% of the whole blood [115,130]. With a selective lysis method, the difference in cell-wall composition is used to lyse the majority of erythrocytes with minimal damage to the leukocytes. In addition to ammonium chloride as a lysis buffer, there are also a few commercially available solutions: FACSlyse solution (Becton Dickenson), Molysis (Molzyme), Zap-oglobin and Coulter Q-Prep lysis solution (Beckman Coulter) [129]. Despite being commonly used, long incubation times with ammonium chloride have been shown to activate leukocytes, thereby changing their membrane expression pattern. Cell activation and cell-membrane alteration due to the isolation technique is always an undesirable effect when attempting to study any cell type [129,131]. On the macro-scale, lysing small volumes of blood (1mL) takes approximately

33

5 minutes. This is longer than necessary for an actual lysis event but it is nonetheless needed due to macro-scale diffusion limitations [129]. This is a perfect scenario where the advantages of microfluidics would come in handy.

As mentioned earlier, bacteria cells distinguish themselves from mammalian cells by having a cell-wall (as opposed to only a cell-membrane). Traditionally, bacteria have been categorized as gram-positive or Gram-negative on the basis of the Gram-stain. Those bacteria strains that maintain the dye are called Gram-positive and those that do not are defined as Gram-negative [128]. The rough and smooth serotype is a gram-negative trait that is based on their lipopolysaccharide (LPS) structure. LPS is found in abundance in the outer membrane (OM) of gram-negative bacteria and can be divided into three regions: Lipid A, core-oligosaccharide and polysaccharide (O antigen) regions. Bacteria that lack the polysaccharide part (O antigen) of the LPS molecule are called rough strains [69,128,132,133]. What follows is a schematic overview (Figure 2.3) of the differences between the cellular barrier of erythrocytes, leukocytes and Gram-negative bacteria, which is a fundamental difference that is taken advantage of in these macro- and micro-based approaches.

Figure 2. 3: Schematic overview of differences between the cellular barrier of red-blood cells, white-blood cells and Gram-negative bacteria. The cell membrane of RBCs and WBCs consists mostly of phospholipids while a Gram-negative cell-wall has an outer membrane, a peptidoglycan layer and an inner membrane.

34

Chemical approach Sethu et al. showed complete lysis of erythrocytes with

ammonium chloride after only 30 seconds in a microfluidic device. The throughput of this particular device is quite low considering the small flow-rate (3.5 µL/min) [129]. In a later publication, Sethu et al. used deionized water to lyse erythrocytes selectively: they managed to lyse all of the RBC while keeping the WBC in a near non-activated-state, something that is considered difficult to achieve in a macro-scale setting where longer incubation times will inevitably activate or lyse cells. All in all, it takes 30 minutes to process 0.6 mL blood (20µL/min), which is approximately a 6-fold improvement from the previous setting [131].

The strategy of selective cell lysis was further improved in one of our studies. By taking advantage of the differences between the blood cell membrane and bacteria cell-wall, we managed to lyse both erythrocytes and leukocytes selectively while keeping the target of interest, the bacteria, intact and viable. This will be further explored in the present investigation section (Chapter 4).

Hwang et al. used another interesting approach in which they treated bacteria samples with sodium acetate to induce adherence. To increase the surface-to-volume ratio even further, a microfluidic device containing pillar arrays was fabricated. A capturing efficiency of 70% was reached for all the sodium acetate samples within the concentration range of 103-107. Bacteria (107 CFU/mL) spiked in 50% blood (diluted with sodium acetate) reached a capturing efficiency of 40% [134]. In a follow-up article, Hwang et al. reported a complete assay by executing everything from sample preparation to PCR on the chip. Here they reached a capturing efficiency of 40% for bacteria samples (104- 107 CFU/mL) spiked in blood with a flow rate of 100-200µL/min [135].

Affinity-Based Approaches There are a few interesting research articles that use affinity-

based assays in a microfluidic system to isolate bacterial cells from blood [136–138]. One of the most interesting approaches has been used by Daniel Kohane’s laboratory at the Boston’s Children Hospital [136], a synthetic ligand, zinc-coordinated bis(dicopolylamine) (bis-Zn-DPA) with high affinity toward both gram-negative and Gram-positive bacterial cell-walls [136,139,140]. As a proof of concept, they were able to isolate E.coli

35

with a concentration of 106 CFU/mL from whole blood using a flow rate of 60mL/h [136]. For this to be applicable in a diagnostic setting the limit of detection must be improved and the binding properties of bis-Zn-DPA with a range of different bacteria species, both Gram-positive and gram-negative, must be tested. However, they did use whole blood as a sample and were able to process the blood quite rapidly (60mL/h) [136]. This ligand was first described in 1964 [139] and has been extensively used by Bradley D. Smith’s laboratory at University of Notre Dame (USA) [139–143]. Here, they demonstrated both the binding of this ligand against both Gram-negative (smooth and rough serotype) and Gram-positive bacteria. Interestingly, they reported a change in the binding properties of bis-ZN-DPA when conjugated to quantum dots. In this configuration, the ligands were only able to bind rough-mutants of E.coli. They also failed to bind any of the Gram-positive prototypes they used, which they attributed to the overall size of the conjugated ligand (15-20 nm), a size too large to enter the pores in the cell-wall (maximum 10nm in diameter) [143]. In the work of Lee et al. (Kohane’s lab), E.coli Stbl3 was used, which is a derivate of E. coli HB101 [136,144,145]. E.coli HB101 is K-12 derivate with a truncated form of LPS, a rough serotype [128,146]. As a comparison, it would be interesting to know if their setup can be applied to both rough and smooth strains [136,144].

A recently published technical report from nature medicine describes a recombinant opsonin-based method to cleanse blood from bacteria and toxins. This recombinant Mannose-binding-lectin holds promise of binding a large panel of different bacteria, fungi, viruses and toxins [147]. The authors provide a unique interpretation of the sepsis dilemma. They do not provide means of identification but a new treatment direction for sepsis patients. If the assay sensitivity is able to match clinical relevant sepsis cases (1-1000 CFU/mL, for symptom showing patients) this could revolutionize treatment strategies. In a constant battle against ever increasing drug-resistant pathogens, this could indeed become a future strategy. Today this method has efficiently cleansed blood containing a bacteria concentration of 104 CFU/mL and a toxin (LPS) level of 10µg/mL. Sepsis patients have an endotoxin level of 300-400pg/mL [148,149].

36

Size-Based Approaches Difference in size is another well-used characteristic for sample

separation. Deterministic lateral displacement (DLD) and filter-based methods are two examples of microfluidic systems that use differences in cell diameters as a separation criterion. However, these methods are associated with clogging, fouling and low flow rates [116]. Another method that involves size as a separation criterion is inertial microfluidics. The most desirable feature of this method is particle focusing at high flow-rates. As a consequence, it is possible to process large volumes sample, which is not as time consuming as with other microfluidic-based systems.

Inertial Microfluidics With inertial microfluidics, particles travel across streamlines

instead of keeping their original inlet position as is seen with laminar flow profiles. Inertial focusing brings the possibility of directing particles of a particular size to a precise equilibrium position (particle focusing) within the flow. As a consequence, particles of different sizes can be sorted out at different outlets. These particle focusing positions arise at high flow rates due to two counteracting forces that act on the particles: shear gradient lift forces and wall-induced lift forces. Smaller particles such as bacteria (1-3 µm) are more difficult to focus since they undergo smaller forces than do the larger blood cells (8-20µm) (Table 2.4) and so maintain a more uniform distribution [150]. However, there are a few articles within the field of inertial microfluidics that focus on separating bacteria from larger surrounding blood cells [99,151,152].

Table 2.4 Different cell sizes.

Cell-type Size

RBC 8 µm in diameter x 2 µm thick WBC 5-20 µm in diameter

Platelets 1-3 µm in diameter

CTC 16-20 µm

E.coli 1-3 µm in length

37

CHAPTER 3

Point-of-Care: The Final Goal

”Point of care testing describes testing using handheld or benchtop technology, where the result will be used in the screening for, or the diagnosis and/or the management of, disease. It is an alternative to using the services of a centralized facility such as a laboratory”.

Christopher P. Price (from Disease Management & Health Outcomes)

38

Point-of-Care: An Overview

Point-of-care (POC) and lab-on-a-chip (LOC) are two major

concepts in medical microfluidics where “cheap”, “fast” and “reliable” are the ultimate goals. The terms of themselves are quite descriptive but still deserve a short conceptual overview. Lab-on-a-chip integrates several laboratory functions into one or a series of miniaturized compartments, providing a potential black-box system, where the sample goes in and an answer comes out. Point-of-care, on the other hand, refers to devices used in close-proximity of the patient. It could be a hand-held device used by medical staff, a home monitoring system used by the patients themselves, or a small bench-top technology that would reduce the dependence on large central laboratory testing sites [153]. The driving forces behind point-of-care development are a number of benefits such as rapid decision making, early therapy initiations, improved treatment optimization and reduced hospital stays [154] .

Point-of-care instrumentation can be categorized in two different subclasses based on their target groups: those developed for resource-limited settings and those made for developed countries [155]. The limiting factor will be significantly different depending on the target group with substantially different requirements. Consequently, all of the devices that are considered “point-of-care” are not suitable for all settings. Some e of the requirements of resource-limited settings are robustness, environmental considerations (temperature, humidity), portability, minimal hands-on-time, no need of highly trained personnel, and easily interpreted results [155,156]. Taking all this into account, it is easy to see that the spectrum of point-of-care devices is very large, ranging from more advanced benchtop devices to dip-stick assays [153]. Regardless of the end user, the point-of-care field would greatly benefit from simplified sample preparation[31]. Two examples of automatic systems that have been mentioned in point-of-care reviews are GeneXpert (Cepheid) and BD Max (BD Diagnostics), both of which use off-line sample pre-treatment [93,157,158]. However, these systems would not be well suited in resource-limited settings since they are high in energy consumption and cost [159].

Numerous studies have shown that the turnaround time (TAT) can be drastically reduced by POC instruments with respect to

39

standardized laboratory tests by removing transport time, sample preparation (centrifugation, separation), validation, and the need to forward results [153]. Although there are several investigations showing improved TATs for point-of-care devices over traditional laboratory testing, the reports on the actual therapeutic TAT and/or the impact on hospital stay vary [153].

Operational Steps within Point-of-Care

One of the first objectives of a point-of-care system is the

reduction of the sample volume from mL scale to a µL scale, the so-called macro-to-micro interface, while retaining the target of interest throughout the entire process. This is not a small task since most analytes often occur in low concentrations. Next, the sample needs to be cleared of potential inhibitors and other abundant cells that might interfere with the downstream analysis. Lysis is usually followed by a target amplification step and, finally, the signal read-out (Figure 3.1) [160].

A possible alternative to conventional PCR is isothermal amplification, which uses one amplification temperature (30°-65°C depending on the method) thereby reducing instrumentation complexity [158,160–162]. The final operational step is detection, which can be achieved either at the endpoint (after the reaction) or in real time (during the reaction). For systems aiming for low cost, endpoint analysis is more appropriate than is to real-time analysis.

Point-of-care instruments would not exist without the joint effort of microfluidic systems (the ability to miniaturize) and progress in software-development. A technical concern that needs to be taken into consideration within all of the operational steps is the need for fluid control. This involves valves, mixing, fluid-movement, external/internal heaters, coolers, choice of material, surface treatments and so forth [155,163,164].

40

Figure 3.1 The Operational steps needed for a point-of-care system: sample preparation, signal amplification and signal-readout. Adapted from Hartman et al. (2013) [160].

Point-of-Care for Bacterial Identification

The classes of analytes within point-of-care system vary from

proteins, cells and nucleic acids (RNA, DNA) to small molecules (glucose, blood gases, electrolytes) [155]. There are, however, two interesting platforms available for bacterial identification (Verigene and Film Array) that meet the requirements of the automated operational steps for point-of-care devices.

Verigene Verigene is a bench-top platform developed by Nanosphere, Inc.,

a company founded in 1999. The platform cartridges offer detection of Gram-positive (BC-GP) and Gram-negative bacteria (BC-GN), yeasts and viruses. The platform processes positive blood cultures (blood-stream infection), stool samples (gastrointestinal infection) as well as whole blood (for genotyping of cardiac samples) [165]. The most impressive feature is the minimal hands-on-time needed (< 5 minutes + 2.5 hours run-time).

Nanosphere’s patented technology consists of a sample processor (SP) and a microarray reader. An assay requires three disposable units: an extraction tray (for nucleic-acid extraction), a utility tray (containing enzymes needed for enzymatic digestion) and the test cartridge (for hybridization). The user simply loads the sample with a pipette onto the first of the disposable cartridges and all sample preparation (lysis, DNA fragmentation and isolation) and hybridization is then performed

41

automatically. An automatic pipette transfers samples on the extraction and utilization tray before finally being moved to the test cartridge for hybridization. The fluid movement within the test cartridge is done by microfluidic channels and pumps [166–168].

Figure 3.2. A picture of the Verigene test cartridge (reagent pack and slide) together with its Verigene processor and reader instrument. This image is reproduced with the permission of the copyright holder [169]

Film Array Film array (BioFire Diagnostics) is a multiplex PCR with

integrated sample preparation, amplification and detection. It requires minimal hands-on-time. A plastic pouch, which contains several units as well as freeze-dried material, is provided (Figure 3.3). The sample to be analyzed is transferred from a syringe to a pouch. The movement of samples within the pouch is controlled by pneumatic pumps [32]. In the first unit, lysis is performed through bead beating. All of the released nucleic acids are bound and transferred by magnetic beads. The Target RNA is first reversely transcribed into DNA in a single large-volume PCR reaction. Next, the diluted samples are transferred into small wells. Each well is designed to detect one specific target. The analysis is done by end-point melting curve data [32,170,171]. A positive blood culture with a concentration range of 107 to 108 is needed [171].

42

Figure 3.3. Diagram of the Film Array system. The Film Array pouch (containing all of the required materials) is loaded into a loading block (1). A solution is added through the blue inlet port to re-hydrate freeze-dried reagents stored in the pouch (2). Next, the sample is added at the red inlet port (4). Upon finishing these two steps, the Film Array pouch is transferred from the loading block to the Film Array instrument where the entire assay is initiated (5). This image is reproduced with the permission of the copyright holder [170].

Concluding Remarks Although Verigene and Film Array platforms have very attractive

plug-and-play solutions, there is still room for further improvement. Foremost, a blood culture step, which may take up to 72 hours, is still needed. Another drawback is that both platforms are only able to run one sample at a time. The number of detectable species, 14 for Verigene and 24 for Filmarray, is yet another important aspect. Only the Verigene platform includes resistance markers (three antibiotic resistant genes for Gram-positive bacteria and six antibiotic resistant genes for Gram-negative bacteria. In conclusion, these platforms are not able to cover all possible organisms or resistance mechanisms [172–174]. They do, however, show good performance with respect to traditional blood culturing methods and have a more rapid turn-around time for their pathogen panels [174].

43

CHAPTER 4

Present investigation

44

Aim of the Thesis

This thesis focuses primarily on sample preparation, with an

emphasis on bacteria isolation. The work behind this thesis consists of three different approaches: (i) immuno-based isolation, (ii) selective cell lysis, (iii-iv) size-based separation. An additional study investigated the activity of a recombinant, Shigella spp Apyrase, which is an important sample preparation tool in bioluminescence assays.

Paper I

This study investigates the possibility of specific isolation of Gram-negative bacteria. To achieve this, antibodies targeting the conserved region of the lipopolysaccharide (LPS) have been used. The challenge lies in epitope unmasking. To improve epitope accessibility, sample heat treatment was developed. The results show significantly improved capture efficiency over non-treated cells.

Paper II

The bacterial cell-wall has a more rigid structure than does the mammalian cell membrane and should, therefore, withstand harsher chemical treatment. This physiological difference has been used to selectively lyse blood cells while keeping bacteria intact and viable for downstream analysis.

Paper III

By using inertial microfluidics, size-dependent particle focusing at high flow-rates has been achieved. Particles with a diameter of 10 µm are positioned at precise streamlines within the curved channel. The focused particles can then be collected at a specific outlet with a separation efficiency of 90%. As a proof of principle, white blood cells were separated from diluted whole blood with an efficiency of 78%.

45

Paper IV

Elasto-inertial focusing is used to separate bacteria from blood. With the use of non-Newtonian fluids, the blood components are diverted to center of the channels while smaller bacteria remain in the side streams and can subsequently be separated.

Paper V

The activity of recombinant Shigella flexineri apyrase (rSFA) is compared to commercially available Solanum tuberosum apyrase (STA). In terms of sample preparation, apyrase is an invaluable “cleanup-tool” for bioluminescence assays where contaminating ATP needs to be removed prior to an assay run. Initial studies show that rSFA has a higher activity than does STA in buffer and serum.

46

Paper I Epitope Unmasking for Improved Immuno-magnetic Isolation of Gram-Negative Bacteria

On a single Gram-negative bacterium, there are approximately 2 x 106 lipopolysaccharide (LPS) molecules, thereby making it one of the major components of the outer cell membrane [69,132]. The LPS molecule consists of three distinct regions: Lipid A, the core oligosaccharide and the O- polysaccharides (Figure 4.1). There are at least 160 different O-polysaccharides serotypes for E.coli alone [175]. Bacteria that somehow have lost the o-polysaccharide chain are called rough strains, while those with a full length LPS are called smooth strains [133,176].

The Lipid A region of this molecule is highly conserved within all gram-negative bacteria, which makes it an interesting target for further investigation [69,132,133]. Although the benefits of targeting the highly conserved and abundant Lipid A portion is clear, there is an accessibility challenge. Access to the Lipid A moiety is limited because it is partly embedded in the membrane and thus is hydrophobic and also because of the steric hindrance caused by the outer region of the LPS molecule and the capsular polysaccharide [69]. This has been demonstrated by several studies, which show that the binding of anti-lipid A antibody is interfered with by the smooth full length O- polysaccharides [175,177–179]. The steric hindrance presented by the LPS molecules also affects the binding of antibodies to other cell surface antigens such as outer membrane proteins (OMP) [180–182]. Membrane alternating antibiotics such as ceftazidim have been shown to have a positive outcome on the binding of anti-lipid A antibodies with the smooth chemotype [177].

47

Figure 4.1: An illustration of a Gram-negative cell-wall. The outermost layer of the outer membrane consists of lipopolysaccharides (LPS). The LPS molecule can be divided into the O-antigen, the core (inner and outer) and the Lipid A region [132].

Summary To improve antigen accessibility, a heat treatment with different

temperatures was tested. An indirect immunofluorescence method was used to verify the treatment effect. The results clearly show improved binding between the antigen and the antibody after heat treatment. For all strains, significant antibody binding could be seen around 60°C (Figure 4.2).

Next, an indirect immune magnetic approach was used to isolate bacteria from PBS (Figure 4.3). Bacteria cells were incubated with anti-lipid-A antibodies (Step 2) after being treated with heat treatment (Step 1, 60°C, 10 minutes). This was followed by an incubation step with protein G-coated magnetic beads (Step 3). The isolated bacteria were then analyzed with PCR.

48

Figure 4.2. Epitope unmasking by heat. An indirect immunofluorescence assay was used to

study the effect of heat on epitope retrieval. Clear improvement in antibody-binding can be seen when bacteria cells were heat-treated, especially around 60°C and 70°C.

As shown in Figure 4.4 the capturing efficiency was significantly improved for samples subjected to mild heat-treatment (+AB 60°C) with respect to the untreated samples (+AB RT). The level of non-specific binding (-AB) was negligible for both the heat-treated and the untreated samples.

49

Figure 4.3: Schematic overview of the immune magnetic bacteria isolation. A bacteriasuspension (1) is subjected to heat to improve epitope accessibility. Second (2), anti-lipid A antibodies are added with excess antibodies being removed by means of centrifugation. The bacteria-antibody complex (3) is incubated with magnetic particles conjugated with secondary antibodies. The bacteria bound on the beads (4) are collected and analyzed with PCR

Figure 4.4: QPCR results for immuno-magnetic isolation of E.coli BL21 (~107 cells). A significant difference in bacterial isolation can be seen between the room-temperature (RT) and the heat-treated samples (60°C). An approximate efficiency of 81% was achieved while the untreated cells have a capture efficiency of around 10%.

50

Paper II Microfluidic Bacteria Isolation from Whole Blood for Sepsis Diagnosis

In this study, the difference between bacterial cell-walls and mammalian cell membranes has been used to lyse blood cells selectively while keeping the bacteria intact. Thanks to the more rigid bacterial cell -wall, bacteria remain not only intact but also viable. To lyse blood-cells, a combination of saponin and osmotic shock was used. Detergents such as saponin have traditionally been used to permeabilize cell membranes and can be found in plants and marine organism [183,184]. In this study the susceptibility to osmotic lysis is further increased by saponin. As a result, the more resistant white blood cells can be lysed as well.

Summary As can be seen in Figure 4.5, bacteria are still viable after

treatment with 0.05-1% saponin and osmotic shock. At higher saponin concentrations, the percentage of viable bacteria drops to 80%.

For the on-chip experiment, a saponin concentration of 1% was used because of the incomplete lysis seen with 0.5% saponin and osmotic shock (Figure 4.6). The difference between macro and micro is likely due to the speed of mixing: mixing by vortexing has a more rapid effect than does the herringbone structure used in the microfluidic chip.

51

Figure 4.5: Bacteria (~ 107 CFU/mL) and blood cells were submitted to different saponin concentrations in order to determine the appropriate cut-off point where the bloods cells are destroyed while the bacteria remain viable. A concentration of 0.5-1% seems to achieve this goal.

Blood spiked with bacteria is added to Inlet 1 (Figure 4.6) while the lysis solution containing saponin is added through Inlet 2. After sufficient time has been given for thorough mixing between the blood sample and lysis buffer, water was added through Inlet 3, to facilitate an osmotic shock event. The flow rates (µl/min) of blood sample, lysis buffer and water had a ratio of 1:1:2.

When the macro- to micro-adaptation was accounted for, the same result was achieved, selective lysing of the blood cells while keeping the bacteria intact and viable (Figure 4.7).

52

Figure 4.6 Microfluidic chip design. (A) This system has two functional sections: saponin treatment (red) and osmotic shock (green). In both sections, a herringbone structure is used to enhance the mixing. (B) The series of images illustrates the mixing efficiency of water and fluorescein. As can be seen, complete mixing is achieved after three turns. The first part of the chip is kept long in order to allow saponin and blood sufficient time to mix. The lysis is terminated off-chip by adding PBS.

53

Figure 4.7: Selective lysis of blood components on the chip. A higher concentration of saponin (1%) was needed in order to lyse all of the blood components selectively. The bacteria were still intact and viable after this treatment. The bacteria concentration in these experiments was 107 CFU/mL.

54

Paper III Dean Flow-Coupled Inertial Focusing in Curved Channels

Inertial microfluidics uses inherent flow characteristics in microchannels to focus and separate particles by size. Dominant inertial forces (wall- and shear-induced lift forces) cause particles to move across streamlines and occupy equilibrium positions along the faces of microchannel walls. In curved channel geometries, an additional secondary flow (Dean Flow) acts on particles and affects the particle equilibrium positions (Figure 4.9).

Figure 4.9: An overview of the different forces at play in a microchannel with a curved geometry. A homogeneous mixture of particles focuses into one single streamline as it exits the 180° turn.

Summary

In this study, we introduced the design of microfluidic U-shaped channels with varied widths to analyze systematically the 10-µm particle behavior through curved channels. The height (50 µm) and the length (20 mm before the curvature) were kept constant. Microfluidic U-shapes channels with the following height:width ratios were fabricated: 1:1; 1:2; 1:5; 1:10 and 1:20. Particle focusing was achieved for the aspect ratios of

55

1:1 to 1:10. The focusing position was found to be independent of the radius of curvatures.

Here, favorable scaling down properties were combined with high flow rates to achieve continuous focusing of 10-µm particles in a device with the aspect ratio of 1:10. A filtration efficiency of 92% was achieved with the processing speed of 4.25 mL/min (Figure 4.10). The device was further tested with biological samples: white blood cells were separated from diluted whole blood with an efficiency of 78% while retaining high viability (Figure 4.11). The flow rate used in this experiment was 2.2 mL/min. At this flow-rate, 98% of the 10-µm beads were collected at the second outlet (Figure 4.11).

Figure 4.10: A. high-throughput (4.25 mL/min) filtration of 10-µm particles with an efficiency of 92%.

56

Figure 4.11: (A). High throughput filtration of 10-µm particles with an efficiency of 96%. (B) White-blood-cell filtration with a filtration efficiency of 78%. The flow rate in both cases was 2.2 mL/min.

57

Paper IV Elasto-Inertial Microfluidics Towards Bacteria Seperation From Whole Blood for Sepsis Diagnostics

In paper III, inertial lift forces were used to focus 10µm particles at a single streamline, hence facilitating their separation and subsequent collection. Because of the size dependent nature of inertial lift forces, smaller particles such as bacteria, will not experience an inertial-induced effect. It is therefore not surprising that particles of 2 µm size have a uniform distribution within the whole channel. To overcome this, the use of “elasto-inertial” microfluidics was explored (paper IV). Here, non-Newtonian fluids (such as a high-viscous polymer) and lift-induced forces are combined. As a consequence particles are focused away from the equilibrium position towards the centre-line of the channel cross-section. In such systems, larger particles can “migrate” away from the sample matrix and into the polymer solution. The smaller particles, on the other hand, will remain in the streamline and be effectively separated. Figure 4.12 shows a schematic overview of an elasto-inertial device for bacteria separation.

Figure 4-12. Schematic overview of elasto-inertial based bacteria separation. The viscoelastic flow enables larger particles to migrate towards the centreline of the channel. Blood spiked with bacteria is flown and mixed with a polymer solution (PEO). Initially, all blood cells remain in the streamline of which they entered but will eventually (after a certain channel length) start to migrate into the polymer solution. The migration is strictly based on size. The blood-cells are thereby focused at the centreline and can be fractionated out through the middle outlet while all bacteria remain in the stream closer to the walls and are separated through the side outlets.

58

Summary The focusing behaviour of different particle sizes (10 and 2 µm)

was evaluated together with various polymer concentrations, channel geometries and flow-rates. Based on experimental results micro-channel dimensions of 50 µm in width, 65 µm height and a minimum of 25 mm in length were found to be optimal. Figure 4.13 shows a successful focusing of 2 and 10 µm particles by the time they exit the microchip.

Figure 4.13: Elasto-inertial focusing and sub-sequent separation of 2 and 10 µm particles. As the particles reach the outlet, larger particles (green) have migrated away from the side-stream where they entered while smaller particles remain. Flow-rate (bead solution): 0.5µL/min. The flow-rate (non-Newtonian solution): 8µL/min.

Initial experiments shows promising separation results of bacteria spiked in whole blood (Figure 4.14). At an optimum flow-rate of 0.25 µL/min, blood-cells are focused away from the side-stream toward the centre-line. Bacteria are then collected both through the side and middle outlet and plated on agar plates. As expected, the majority of the bacteria cells are present at the side-outlets.

59

Figure 4.14: Plating results of bacteria spiked in whole blood and flown through the channel. At the optimum flow rate (0.25 µl(min) , the bacteria are mainly recovered through the side outlets. The data are from three independent experiments.

60

Paper V Recombinant Shigella Flexneri Apyrase Enzyme for Bioluminescence-Based Diagnostic Applications

Bioluminescence is an attractive analytical method in many biotechnological applications where light is emitted due to the actions of a luciferase enzyme and its substrate luciferin. Here, adenosine triphosphate (ATP) drives the luciferase-mediated conversion of luciferin to oxyluciferin, which generates visible light (Figure 4.12). Since cellular ATP of a bacteria cell is relatively constant (one attomole), the luciferase-luciferin reaction can be used to determine the number of bacteria cells in a sample. A vital step in such an assay is to eliminate any pre-existing ATP in the sample matrix. To this end, apyrase is added prior to a luciferin-luciferase reaction. Apyrase has the ability to sequentially hydrolyse ATP to diphosphates (ADP) and monophosphates (AMP). The most commonly used Apyrase is obtained from Solanum tuberosum (STA), which is also called potato apyrase.

Figure 4.12: Luciferin-Luciferase reaction yielding Oxyluciferin, monophosphate and light.

The results show that RSFA depletes ATP at much higher rates than does the potato apyrase (STA). Another interesting observation is the level of ATP, which almost reaches zero for RFSA after only 10 minutes (Figure 4.13). This level of ATP depletion is never reached by STA, which means that residual ATP from the sample matrix would always be present in the subsequent reaction. The activity of these

61

enzymes was further tested in the presence of 10% (vol/vol) serum. A simillar ATP depletion and activity rate was noted in serum (Figure 4.14).

Figure 4.13: An activity comparison between Solanum tuberosum (STA) and Shigella flexneri (RSFA) apyrase in buffer. RFSA depletes adenosine triphosphate (ATP) at a higher rate and to lower levels than does STA.

62

Figure 4.14: An activity comparison between Solanum tuberosum (STA) and Shigella flexneri (RSFA) apyrase in 10% (vol/vol) serum.

63

Conclusion and Future Work

Paper I: The virulence of gram-negative bacteria has been attributed to the Lipid A Region of the LPS molecule, which is also referred to as an endotoxin. In the 1990s, attempts were made to produce an anti-Lipid A antibody to neutralize endotoxins in patients suffering from sepsis [133,175,185–187]. The basis of this strategy was to produce cross-reactive antibodies, by taking advantage of the structure similarity of the Lipid A region between different strains. However, the results were contradictory [175,176,185,187,188].

In Paper I, the conserved region of the LPS molecule (Lipid A) was explored. To increase epitope access, an epitope retrieval method (without lysing the cells) was developed. Heat-treatment was shown to have a positive outcome on the antigen access for the proof-of-principle strains used in the study.

There are several possible ways to extend the work of Paper I. First, it would be interesting to study the effect on pathogenic strains and compare immune-magnetic isolation between treated and untreated cells. Second, the treatment can be improved. Lipopolysaccharides are a tightly packed forest on the outer layer of the cell-wall, covering 75% of its surface [128,176,189]. This compressed state is made possible by stabilizing cations (Ca2+, Mg2+) which neutralizes the repulsive forces emitted by each LPS molecule [128,176,189,190]. With the use of chelators such as ethylenediaminetetraacetic acid (EDTA), membrane instability can be induced. An EDTA-induced release of the LPS molecules without leading to a lysis event has been reported [190,191]. Whether or not such disturbance in the tightly-packed LPS forest leads to improved antibody binding or not can only be determined experimentally. Another interesting approach would be to use acetic acid, which has been extensively used in protocols for counting white-blood cells. Here, whole blood is diluted in an acetic acid solution to lyse red blood cells prior counting the white blood cells. In hematology, this solution is referred to as “Turks solution” [192]. An acetic-acid treatment could offer a dual action: lysis of unwanted blood cells (making it easier to work with complex solutions) as well as increasing epitope accessibility. Mild acid treatment is known to cleave the linkage between Lipid A and

64

the core LPS [193]. This linkage exists in all LPS molecules and is called KDO, 3-deoxy-D-manno-oct-2-ulosonic acid [193]. Bellstedt et al. have used this process to create what they call “naked bacteria”, thereby creating an immune-carrier for immunization [194].

Paper II: In Paper II, complete lysis of whole blood was achieved while keeping the bacteria intact and viable. The presence of a rigid bacterial cell-wall protects the bacteria from a lysis event [128]. Bacteria had full survival (100%) in the microfluidic chip for concentrations from 107 to 106 CFU/mL. For concentration from 105 to 104 CFU/mL, bacteria showed a slight decrease in survival rate.

The work of Paper II can be further extended by including parallelization in future versions of the device. Thus, samples can be processed more rapidly. The retained bacteria viability is an interesting trait worth further research. One suggested application would be to combine this device with rapid antibiotic susceptibility testing [195,196].

Paper III: One of the challenges of microfluidic-based sample preparation is the need to process large sample volumes. Often, a compromise is made between speed and performance. In Paper III, inertial microfluidics was used to focus and separate particles continuously at high flow rates. Here, we used the synergetic effect of inertial lift forces and Dean forces to focus particles based on size. An efficiency of 96% was achieved for 10µm beads while the efficiency of white blood cells was 78%. The flow rates used for both cases were 2.2 mL/min. Although inertial microfluidics show great potential for high throughput blood processing, an inherent limitation for sepsis diagnostics is the small bacteria size (1-3 µm), which makes them difficult to focus in the presence of larger blood-cells.

To overcome this, two directions are envisioned: (i) the use of “elasto-inertial” microfluidics (Paper IV) or (ii) integration with selective blood cell lysis (Paper II).

65

Paper IV: By combining inertial-lift forces together with non-Newtonian fluids, larger blood-cells can be focused away from entering side-streams toward the center of the microchip. As a result, blood-cells can efficiently be removed from the sample. Since there is little or no impact on the small-sized bacteria they will remain at the sideline. Although a promising result, setting bacteria and blood cells apart, further validation of the performance and separation efficiency is needed. One struggling point is system clogging, making quantification through plating difficult. In order to meet the need of high-throughput blood processing, multi-parallel-channels can be fabricated.

Paper V: The efficiency of removing ATP traces is much higher in Shigella flexineri apyrase (rSFA) than the commercially available Solanum tuberosum apyrase (STA). Consequently, rSFA has the potential of improving sample-preparation steps for bioluminescence assays and thereby refining the detection limit even further.

66

Acknowledgement

I’ve learned a lot, been to a lot of places and more importantly met a lot of wonderful, inspiring people. This has been the essence of my Ph.D. journey and there are a lot of people I would like to show my appreciation to. Thank you all for making these years memorable!!

Aman! First of all I would like to thank you for accepting me as your Ph.D. student. Without you none of this would have been possible. One has to search very long and very hard to find a more enthusiastic and driven person. Thank you for all the interesting inspiring discussions. I hope you reach your goal of providing valuable tools for resource-limited areas.

Cell physics was my starting point and I have a lot of happy memories from this division. A big thank you to Padideh and Athanasia for making this environment extra quirky and full of energy. Padideh, you’ve become like a sister to me. You’re the sweetest, most generous person I know and I am truly happy that our roads crossed. A lot of people have contributed in making cell physics feel like home; Ida, our lovely downstairs (BioX) neighbor, thank you for patiently answering all my physics related questions and dragging me to Zumba sessions (I needed it!), Bruno, for always believing in me (!), Karolin for being cheerful at all times (I especially liked our adventures in China), Elin for all the fun we had together, Per for being your talkative self, Thomas F for all the gossip and all the fika1 and last but definitely (!) not least Kate for all the laughs we shared! You’ve become one of my closest friends and partner in crime. I will never forget our attempts in making home-made Christmas presents! It would have been a lonely lab without all of you. I would also like to thank the PI’s: Hjalmar, Aman, and Björn for being the foundation of this environment.

*Whoever visits Sweden for an extended time period will find it impossible to avoid this activity/word: 1Fika, roughly means "to drink coffee/tea/," usually accompanied by something sweet. Please note that it can be used as both a verb and a noun.

67

Special thanks to all the members in the Intopsens consortium. Even though my first years were filled with deliverables, tight time-schedules and a high pace, you all made it worth it. I enjoyed all our meetings and discussions, especially Valencia 2010.

My Ph.D journey went from applied physics to the school of Biotechnology. Although, I didn’t have time to properly get to know all of you, I feel that biotechnology have become my home. Helen. Thank you for creating such a fantastic environment. It’s always inspiring to see that research and business can be combined!

I would like to thank all the former and present members of the Nano-biotechnology group. You couldn’t wish for better colleagues and friends. My best moments have been shared with many of you. I hope for countless re-union events in the future. Special thanks to Emilie, Mary, Lovisa, Jånas, Philippa, Eva (honorary member), Gustav, Prem, Staffan, Cesc, Petter, Jorge and Dave for brightening my days. I would also like to express my deepest appreciation to all of you (Sergey, Pavan, Jånas, Asim, Mary, Zenib and Nilay) for helping me in my projects. This work would not have been possible without your valuable input and contributions. Sincere thanks goes to Andrés Veide, for graciously handling my giggling attacks and for teaching me about bacteria work.

During my time at the school of Biotechnology, I especially enjoyed my Albanova office, or more precisely the combination of people in it (Emilie, Camilla, Daniel, Tarek, Thiru and occasionally Jesper). I think we discussed everything between heaven and earth. Nothing was too stupid. This office had quite a high threshold for everything, even grumpiness and sarcastic remarks (always given with love) were quite tolerated. I miss it still. Daniel, thank you for enduring my tendencies to (slowly but ever so steady) take over your desk (I know it was hard on you). Tarek. I very much like your fearless attitude in life, especially when it leads to buying a sailing boat (with limited experience) and quickly learning to master the art. A big thank you, to both you and Eva for letting me join in on those adventures, it was great fun! Emilie, you have been an invaluable friend and I admire your positive outlook on life. I appreciate all your everyday gestures of kindness. Priceless!

68

Beyond the realm of Science (which sometimes seems like a distant dream) I would like to tell all my Uppsala friends how much you all mean to me. I will never forget Gotland and our wonderful ski trips. I hope there will be many to come now that I will have somewhat more time on my hands.

A big thank you to Edward, who knew that authors do strange things but read my thesis anyway. Your emails have been a great source of inspiration and your input and suggestion priceless.

Last but not least I would like to thank my family, my mom, dad, and my sister for their unconditional love and support. I would not be the person that I am, if not for you. You all mean the world to me!

To my extended family, thank you for taking care of me, for all the relaxing moments, delicious home-made dinners and our many board game evenings.

Above all I would like to thank Saeid, for making me laugh whenever I needed it and for all your love and never-ending kindness. There is no-one nicer and more understanding than you. I know, at times, that my temper was particularly trying. You kept me sane during all the ups-and downs of this journey. Love you so very much.

Some final words to all the lovely girls at Biotech (plan 3): You rock!! Hope to see all of you on the dance floor!!!

69

Abbreviation

WHO World health organisation

ICU Intensive care unit

SIRS Systemic inflammatory response

CDC Center of Disease Control

CFU Colony forming units

MIC Minimal inhibitory concentration

NAT Nucleic acid based techniques

DNA Deoxyribonucleic acid

CpG Cytidylatephosphate-deoxyguanylate

PCR Polymerase Chain Reaction

IgG immunoglobulin G

RBC Red blood cells

WBC White blood cells

CTC Circulating tumor cells

ELISA Enzyme-linked immunosorbent assay

PDMS Polydimethylsiloxane

E.coli Escherichia coli

LPS Lipopolysaccharide

OM Outer membrane

bis-Zn-DPA Zinc-coordinated bis(dicopolylamine)

70

DLD Deterministic lateral displacement

POC Point-of-care

LOC Lab-on-a chip

TAT Turnaround time

LAMP Loop-mediated-isothermal amplification

RNA Ribonucleic acid

OMP Outer membrane proteins

PBS Phosphate-Buffered Salin

AB Antibody

EDTA Ethylenediaminetetraacetic acid

KDO 3-deoxy-D-manno-oct-2-ulosonic acid

71

Bibliography

1. Thomas L (1978) The Lives of a Cell: Notes of a Biology Watcher. Available: http://jama.ama-assn.org/content/231/2/199.3.short. Accessed 23 January 2014.

2. Kanki P, Grimes DJ (2012) Infectious Diseases - Selected Entries from the Encyclopedia of Sustainability Science and Technology. Kanki P, Grimes DJ, editors New York: Springer Science+Business Media. doi:10.1007/978-1-4614-5719-0.

3. World health Organisation (2013). Fact Sheet No310. Available: http://who.int/mediacentre/factsheets/fs310/en/index1.html. Accessed 23 January 2014.

4. Vincent J, Marshall J, Anzueto A, Martin CD, Gomersall C (2009) International Study of the Prevalence and Outcomes of Infection in Intensive Care Units. 302: 2323–2329.

5. Klevens M, Edwards JR, Richards CL, Horan TC, Gaynes RP, et al. (2007) Estimating Health Care-Associated Infections and Deaths in U.S. Hospitals, 2002. 122: 160–166.

6. Calfee DP (2012) Crisis in hospital-acquired, healthcare-associated infections. Annu Rev Med 63: 359–371. Available: http://www.ncbi.nlm.nih.gov/pubmed/22017445. Accessed 5 May 2014.

7. Raghavan M, Marik PE (2006) Management of sepsis during the early “golden hours.” J Emerg Med 31: 185–199. Available: http://linkinghub.elsevier.com/retrieve/pii/S0736467906003519. Accessed 12 May 2014.

8. Vincent J, Bihari DJ, Suter PM, Bruining HA, White J, et al. (1995) The Prevalence of Nosocomial Infection in Intensive Care Units in Europe in Intensive Care ( EPIC ) Study. J Am Med Assoc 274: 639–644. doi:10.1001/jama.1995.03530080055041.

9. Peleg AY, Hooper DC (2010) Hospital-acquired infections due to gram-negative bacteria. N Engl J Med 362: 1804–1813. Available: http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=3107499&tool=pmcentrez&rendertype=abstract.

72

10. Towns ML, Jarvis WR, Hsueh P-R (2010) Guidelines on blood cultures. J Microbiol Immunol Infect 43: 347–349. Available: http://www.ncbi.nlm.nih.gov/pubmed/20688297. Accessed 2 September 2013.

11. Shlaes DM (2010) Antibiotics The Perfect Storm. Springer Science+Business Media. doi:10.1007/978-90-481-9057-7.

12. Spellberg B, Guidos R, Gilbert D, Bradley J, Boucher HW, et al. (2008) The epidemic of antibiotic-resistant infections: a call to action for the medical community from the Infectious Diseases Society of America. Clin Infect Dis 46: 155–164. Available: http://www.ncbi.nlm.nih.gov/pubmed/18171244. Accessed 4 May 2014.

13. Jabes D (2011) The antibiotic R&D pipeline: an update. Curr Opin Microbiol 14: 564–569. Available: http://www.ncbi.nlm.nih.gov/pubmed/21873107. Accessed 16 May 2014.

14. Slade E, Tamber PS, Vincent J (2003) The Surviving Sepsis Campaign  : raising awareness to reduce mortality: 1–2. doi:10.1186/cc1876.

15. Chalupka AN, Talmor D (2012) The economics of sepsis. Crit Care Clin 28: 57–76, vi. Available: http://www.ncbi.nlm.nih.gov/pubmed/22123099. Accessed 23 January 2014.

16. Vincent J-L (2012) Increasing awareness of sepsis: World Sepsis Day. Crit Care 16: 152. Available: http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=3682256&tool=pmcentrez&rendertype=abstract. Accessed 8 May 2014.

17. Jacob ST, Ortiz JR, West E (2012) Treatment of sepsis. Lancet Infect Dis 12: 746. Available: http://www.ncbi.nlm.nih.gov/pubmed/23017362. Accessed 16 February 2014.

18. World Bank (n.d.). Available: http://wdi.worldbank.org/table/2.1. Accessed 28 January 2014.

73

19. Townsend S, Dellinger RP, Levy MM, Ramsay G (2005) Implementing the surviving sepsis campaign.

20. The Society of Critical Care Medicine (SCCM (n.d.). Available: http://www.sccm.org/About-SCCM/Pages/default.aspx. Accessed 27 January 2014.

21. Martin GS, Mannino DM, Eaton S, Moss M (2003) The epidemiology of sepsis in the United States from 1979 through 2000. N Engl J Med 348: 1546–1554. Available: http://www.ncbi.nlm.nih.gov/pubmed/12700374.

22. Russell J a (2006) Management of sepsis. N Engl J Med 355: 1699–1713. Available: http://www.ncbi.nlm.nih.gov/pubmed/17050894.

23. Angus DC, Wax RS (2001) Epidemiology of sepsis: an update. Crit Care Med 29: S109–S116. doi:10.1097/00003246-200107001-00035.

24. Lazcka O, Del Campo FJ, Muñoz FX (2007) Pathogen detection: a perspective of traditional methods and biosensors. Biosens Bioelectron 22: 1205–1217. Available: http://www.ncbi.nlm.nih.gov/pubmed/16934970. Accessed 19 February 2014.

25. Bone RC, Sibbald WJ, Sprung CL (1992) The ACCP-SCCM consensus conference on sepsis and organ failure. Chest 101: 1481–1483. Available: http://www.ncbi.nlm.nih.gov/pubmed/20162763.

26. Coelho FR, Martins JO (2012) Diagnostic methods in sepsis: the need of speed. Rev Assoc Med Bras 58: 498–504. Available: http://www.ncbi.nlm.nih.gov/pubmed/22930032.

27. Llewelyn M, Cohen J (2007) Infectious disease in critical care. Jordi Rello MD, PhDMD MK, MD ED, Alejandro Rodríguez MD P, editors Springer-Verlag Berlin Heidelberg New York.

28. Mayr FB, Yende S, Angus DC (2013) Epidemiology of severe sepsis. Virulence 5: 4–11. Available: http://www.ncbi.nlm.nih.gov/pubmed/24335434.

74

29. De Jong HK, van der Poll T, Wiersinga WJ (2010) The systemic pro-inflammatory response in sepsis. J Innate Immun 2: 422–430. Available: http://www.ncbi.nlm.nih.gov/pubmed/20530955. Accessed 28 January 2014.

30. Mancini N, Carletti S, Ghidoli N, Cichero P, Burioni R, et al. (2010) The era of molecular and other non-culture-based methods in diagnosis of sepsis. Clin Microbiol Rev 23: 235–251. Available: http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=2806664&tool=pmcentrez&rendertype=abstract. Accessed 16 August 2013.

31. Ritzi-Lehnert M (2012) Development of chip-compatible sample preparation for diagnosis of infectious diseases. Expert Rev Mol Diagn 12: 189–206. Available: http://www.ncbi.nlm.nih.gov/pubmed/22369378.

32. Lebovitz EE, Burbelo PD (2013) Commercial multiplex technologies for the microbiological diagnosis of sepsis. Mol Diagn Ther 17: 221–231. Available: http://www.ncbi.nlm.nih.gov/pubmed/23636778. Accessed 11 September 2013.

33. Afshari A, Schrenzel J, Ieven M, Harbarth S (2012) Bench-to-bedside review: Rapid molecular diagnostics for bloodstream infection - a new frontier? Crit Care 16: 222. Available: http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=3580598&tool=pmcentrez&rendertype=abstract.

34. Spencer RC (1994) Intensive Care Medicine Epidemiology of infection in ICUs: 2–6.

35. Moss M, Martin GS (2004) A global perspective on the epidemiology of sepsis. Intensive Care Med 30: 527–529. Available: http://www.ncbi.nlm.nih.gov/pubmed/14985955. Accessed 12 February 2014.

36. Martin GS (2013) Sepsis, severe sepsis and septic shock: changes in incidence, pathogens and outcome. Expert Rev Anti Infect Ther 10: 701–706. doi:10.1586/eri.12.50.Sepsis.

75

37. Daniels R (2011) Surviving the first hours in sepsis: getting the basics right (an intensivist’s perspective). J Antimicrob Chemother 66 Suppl 2: ii11–ii23. Available: http://www.ncbi.nlm.nih.gov/pubmed/21398303. Accessed 22 January 2014.

38. Laupland KB (2013) Incidence of bloodstream infection: a review of population-based studies. Clin Microbiol Infect 19: 492–500. Available: http://www.ncbi.nlm.nih.gov/pubmed/23398633.

39. CDC/National Center for Health Statistics (2010) AIDS and HIV. Available: http://www.cdc.gov/nchs/fastats/aids-hiv.htm.

40. CDC/National Center for Health Statistics (2010) Mammography/Breast Cancer. Available: http://www.cdc.gov/nchs/fastats/mammography.htm.

41. CDC/National Center for Health Statistics (2010) Cerebrovascular Disease or Stroke. Available: http://www.cdc.gov/nchs/fastats/stroke.htm.

42. Moss M (2005) Epidemiology of sepsis: race, sex, and chronic alcohol abuse. Clin Infect Dis 41 Suppl 7: S490–S497. Available: http://www.ncbi.nlm.nih.gov/pubmed/16237652.

43. Mossie A (2013) PATHOPHYSIOLOGY OF SEPSIS Flow chart SEPSIS  : TERMS AND DEFINITIONS. 1: 159–168.

44. Townsend S, Dellinger RP, Levy MM, Ramsay G (2005) Implementing the Surviving Sepsis Campaign. the Society of Critical Care Medicine, the European Society of Intensive Care Medicine, and the International Sepsis Forum.

45. Wang HE, Devereaux RS, Yealy DM, Safford MM, Howard G (2010) National variation in United States sepsis mortality: a descriptive study. Int J Health Geogr 9: 9. Available: http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=2831852&tool=pmcentrez&rendertype=abstract.

46. Padkin A, Goldfrad C, Brady AR, Young D, Black N, et al. (2003) Epidemiology of severe sepsis occurring in the first 24 hrs in intensive care units in England, Wales, and Northern Ireland. Crit

76

Care Med 31: 2332–2338. Available: http://www.ncbi.nlm.nih.gov/pubmed/14501964. Accessed 12 February 2014.

47. Harrison D a, Welch C a, Eddleston JM (2006) The epidemiology of severe sepsis in England, Wales and Northern Ireland, 1996 to 2004: secondary analysis of a high quality clinical database, the ICNARC Case Mix Programme Database. Crit Care 10: R42. Available: http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=1550902&tool=pmcentrez&rendertype=abstract. Accessed 1 February 2014.

48. Esteban A, Frutos-Vivar F, Ferguson ND, Peñuelas O, Lorente JA, et al. (2007) Sepsis incidence and outcome: contrasting the intensive care unit with the hospital ward. Crit Care Med 35: 1284–1289. Available: http://www.ncbi.nlm.nih.gov/pubmed/17414725. Accessed 12 February 2014.

49. Engel C, Brunkhorst FM, Bone H-G, Brunkhorst R, Gerlach H, et al. (2007) Epidemiology of sepsis in Germany: results from a national prospective multicenter study. Intensive Care Med 33: 606–618. Available: http://www.ncbi.nlm.nih.gov/pubmed/17323051. Accessed 22 January 2014.

50. Flaatten H (2004) Epidemiology of sepsis in Norway in 1999. Crit Care 8: R180–R184. doi:10.1186/cc2867.

51. Angus DC, Linde-Zwirble WT, Lidicker J, Clermont G, Carcillo J, et al. (2001) Epidemiology of severe sepsis in the United States: analysis of incidence, outcome, and associated costs of care. Crit Care Med 29: 1303–1310. Available: http://www.ncbi.nlm.nih.gov/pubmed/11445675.

52. Poeze M, Ramsay G, Gerlach H, Rubulotta F, Levy M (2004) An international sepsis survey: a study of doctors’ knowledge and perception about sepsis. Crit Care 8: R409–R413. Available: http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=1065059&tool=pmcentrez&rendertype=abstract. Accessed 5 February 2014.

77

53. Zhao H, Heard SO, Mullen MT, Crawford S, Goldberg RJ, et al. (2012) An evaluation of the diagnostic accuracy of the 1991 American College of Chest Physicians/Society of Critical Care Medicine and the 2001 Society of Critical Care Medicine/European Society of Intensive Care Medicine/American College of Chest Physicians/Ame. Crit Care Med 40: 1700–1706. Available: http://www.ncbi.nlm.nih.gov/pubmed/22610176. Accessed 21 February 2014.

54. Levy MM, Fink MP, Marshall JC, Abraham E, Angus D, et al. (2003) 2001 SCCM/ESICM/ACCP/ATS/SIS International Sepsis Definitions Conference. Crit Care Med 31: 1250–1256. Available: http://www.ncbi.nlm.nih.gov/pubmed/12682500. Accessed 21 January 2014.

55. Huttunen R, Aittoniemi J (2011) New concepts in the pathogenesis, diagnosis and treatment of bacteremia and sepsis. J Infect 63: 407–419. Available: http://www.ncbi.nlm.nih.gov/pubmed/21840338. Accessed 20 February 2014.

56. Huttunen R, Syrjänen J, Vuento R, Aittoniemi J (2013) Current concepts in the diagnosis of blood stream infections. Are novel molecular methods useful in clinical practice? Int J Infect Dis 358: 1–5. Available: http://www.ncbi.nlm.nih.gov/pubmed/23871281. Accessed 20 August 2013.

57. Kumar A, Ellis P, Arabi Y, Roberts D, Light B, et al. (2009) Initiation of inappropriate antimicrobial therapy results in a fivefold reduction of survival in human septic shock. Chest 136: 1237–1248. Available: http://www.ncbi.nlm.nih.gov/pubmed/19696123. Accessed 6 August 2013.

58. Kumar A, Roberts D, Wood KE, Light B, Parrillo JE, et al. (2006) Duration of hypotension before initiation of effective antimicrobial therapy is the critical determinant of survival in human septic shock. Crit Care Med 34: 1589–1596. Available: http://www.ncbi.nlm.nih.gov/pubmed/16625125. Accessed 7 August 2013.

59. Ibrahim EH, Sherman G, Ward S, Fraser VJ, Kollef MH (2000) The influence of inadequate antimicrobial treatment of

78

bloodstream infections on patient outcomes in the ICU setting. Chest 118: 146–155. Available: http://www.ncbi.nlm.nih.gov/pubmed/10893372.

60. Leibovici L, Shraga I, Drucker M, Konigsberger H, Samra Z, et al. (1998) The benefit of appropriate empirical antibiotic treatment in patients with bloodstream infection. J Intern Med 244: 379–386. Available: http://www.ncbi.nlm.nih.gov/pubmed/9845853.

61. Dreyer AW (2012) Blood Culture Systems  : From Patient to Result. Available: http://dx.doi.org/10.5772/50139.

62. Weinstein MP (1996) Current blood culture methods and systems: clinical concepts, technology, and interpretation of results. Clin Infect Dis 23: 40–46. Available: http://www.ncbi.nlm.nih.gov/pubmed/8816127.

63. Chinniah TR (2012) Blood Culture Techniques  : Increasing Yield and Reducing Contamination.

64. Shafazand S, Weinacker AB (2002) Blood cultures in the critical care unit: improving utilization and yield. Chest 122: 1727–1736. Available: http://www.ncbi.nlm.nih.gov/pubmed/12426278.

65. Mylotte JM, Tayara A (2000) Blood Cultures: Clinical Aspects and Controversies. Eur J Clin Microbiol Infect Dis 19: 157–163. Available: http://dx.doi.org/10.1007/s100960050453.

66. Reimer LG, Wilson ML, Weinstein MP (1997) Update on detection of bacteremia and fungemia. Clin Microbiol Rev 10: 444–465. Available: http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=172929&tool=pmcentrez&rendertype=abstract.

67. Sengupta S, Gordon JE, Chang H, Dame N (n.d.) Microfluidics for Biological Applications. Finehout E, Tian W-C, editors New York: Springer Science+Business Media. Available: http://link.springer.com/book/10.1007/978-0-387-09480-9.

68. Chandrasekar PH, Brown WJ (1994) Clinical issues of blood cultures. Arch intern med 154: 841–849. Available: http://archinte.jamanetwork.com/article.aspx?articleid=618696.

79

69. Mayeux PR (1997) Pathobiology of lipopolysaccharide. J Toxicol Environ Health 51: 415–435. Available: http://www.ncbi.nlm.nih.gov/pubmed/9233377.

70. Wenzel RP, Pinsky MR, Ulevitch RJ, Young L (1996) Current understanding of sepsis. Clin Infect Dis 22: 407–412. Available: http://www.ncbi.nlm.nih.gov/pubmed/8852954.

71. Scott LJ (2013) Verigene® gram-positive blood culture nucleic acid test. Mol Diagn Ther 17: 117–122. Available: http://www.ncbi.nlm.nih.gov/pubmed/23345071. Accessed 14 February 2014.

72. Ledeboer N a., Hodinka RL (2011) Molecular Detection of Resistance Determinants. J Clin Microbiol 49: S20–S24. Available: http://jcm.asm.org/cgi/doi/10.1128/JCM.00771-11. Accessed 19 February 2014.

73. Sundsfjord A, Simonsen GS, Haldorsen BC, Haaheim H, Hjelmevoll S-O, et al. (2004) Genetic methods for detection of antimicrobial resistance. APMIS 112: 815–837. Available: http://www.ncbi.nlm.nih.gov/pubmed/15638839.

74. Jorgensen JH, Ferraro MJ (2009) Antimicrobial susceptibility testing: a review of general principles and contemporary practices. Clin Infect Dis 49: 1749–1755. Available: http://www.ncbi.nlm.nih.gov/pubmed/19857164. Accessed 26 January 2014.

75. Fluit AC, Visser MR, Schmitz F (2001) Molecular Detection of Antimicrobial Resistance Molecular Detection of Antimicrobial Resistance. 14. doi:10.1128/CMR.14.4.836.

76. Woodford N, Sundsfjord A (2005) Molecular detection of antibiotic resistance: when and where? J Antimicrob Chemother 56: 259–261. Available: http://www.ncbi.nlm.nih.gov/pubmed/15967769. Accessed 14 August 2013.

77. Loonen a JM, Wolffs PFG, Bruggeman C a, van den Brule a JC (2014) Developments for improved diagnosis of bacterial bloodstream infections. Eur J Clin Microbiol Infect Dis. Available:

80

http://www.ncbi.nlm.nih.gov/pubmed/24848132. Accessed 28 May 2014.

78. Tang Y-W, Stratton CW (2013) Advanced Tecnhiques in Diagnostic Microbiology. second. Springer Science+Business Media.

79. Klouche M, Schröder U (2008) Rapid methods for diagnosis of bloodstream infections. Clin Chem Lab Med 46: 888–908. Available: http://www.ncbi.nlm.nih.gov/pubmed/18624614. Accessed 28 May 2014.

80. Horz H-P, Scheer S, Vianna ME, Conrads G (2010) New methods for selective isolation of bacterial DNA from human clinical specimens. Anaerobe 16: 47–53. Available: http://www.ncbi.nlm.nih.gov/pubmed/19463963. Accessed 11 March 2014.

81. Loonen AJM, Bos MP, van Meerbergen B, Neerken S, Catsburg A, et al. (2013) Comparison of pathogen DNA isolation methods from large volumes of whole blood to improve molecular diagnosis of bloodstream infections. PLoS One 8: e72349. Available: http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=3744477&tool=pmcentrez&rendertype=abstract. Accessed 19 March 2014.

82. Sandetskaya N, Naumann A, Hennig K, Kuhlmeier D (2014) Specific enrichment of prokaryotic DNA using a recombinant DNA-binding protein. Anal Bioanal Chem 406: 3755–3762. Available: http://www.ncbi.nlm.nih.gov/pubmed/24718438. Accessed 5 June 2014.

83. Jordana-Lluch E, Giménez M, Quesada MD, Ausina V, Martró E (2014) Improving the Diagnosis of Bloodstream Infections: PCR Coupled with Mass Spectrometry. Biomed Res Int 2014: 501214. Available: http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=4000954&tool=pmcentrez&rendertype=abstract.

84. Al-Soud WA, Rådström P (2001) Purification and Characterization of PCR-Inhibitory Components in Blood Cells. 39: 485–493. doi:10.1128/JCM.39.2.485.

81

85. Rådström P, Knutsson R, Wolffs P, Lövenklev M, Löfström C (2004) Pre-PCR Processing. 26.

86. Al-Soud W a, Jönsson LJ, Râdström P (2000) Identification and characterization of immunoglobulin G in blood as a major inhibitor of diagnostic PCR. J Clin Microbiol 38: 345–350. Available: http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=88721&tool=pmcentrez&rendertype=abstract.

87. Laakso S, Mäki M (2013) Assessment of a semi-automated protocol for multiplex analysis of sepsis-causing bacteria with spiked whole blood samples. Microbiologyopen 2: 284–292. Available: http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=3633352&tool=pmcentrez&rendertype=abstract. Accessed 12 March 2014.

88. Toner M, Irimia D (2005) Blood-on-a-chip. Annu Rev Biomed Eng 7: 77–103. Available: http://www.ncbi.nlm.nih.gov/pubmed/16004567. Accessed 15 August 2013.

89. Hayes DF, Smerage JB (2010) Circulating tumor cells. Elsevier Inc. Available: http://www.ncbi.nlm.nih.gov/pubmed/21075330. Accessed 18 March 2014.

90. Obata K, Mukai K, Tsujimura Y, Ishiwata K, Kawano Y, et al. (2007) Basophils are essential initiators of a novel type of chronic allergic inflammation. Blood 110: 913–920. Available: http://www.ncbi.nlm.nih.gov/pubmed/17409268. Accessed 29 April 2014.

91. Zourob M, Elwary S, Turner A (2008) Principles of Bacterial Detection  : Biosensors, Recognition REceptors and Microsystems. Springer Science+Business Media.

92. Crevillén AG, Hervás M, López MA, González MC, Escarpa A (2007) Real sample analysis on microfluidic devices. Talanta 74: 342–357. Available: http://www.ncbi.nlm.nih.gov/pubmed/18371648. Accessed 1 November 2014.

82

93. Bissonnette L, Bergeron MG (2012) Infectious Disease Management through Point-of-Care Personalized Medicine Molecular Diagnostic Technologies. J Pers Med 2: 50–70. Available: http://www.mdpi.com/2075-4426/2/2/50/. Accessed 17 January 2014.

94. Mark D, Haeberle S, Roth G, von Stetten F, Zengerle R (2010) Microfluidic lab-on-a-chip platforms: requirements, characteristics and applications. Chem Soc Rev 39: 1153–1182. Available: http://www.ncbi.nlm.nih.gov/pubmed/20179830. Accessed 23 March 2014.

95. Whitesides GM (2006) The origins and the future of microfluidics. Nature 442: 368–373. Available: http://www.ncbi.nlm.nih.gov/pubmed/16871203. Accessed 21 January 2014.

96. Beebe DJ, Mensing G a, Walker GM (2002) Physics and applications of microfluidics in biology. Annu Rev Biomed Eng 4: 261–286. Available: http://www.ncbi.nlm.nih.gov/pubmed/12117759. Accessed 23 March 2014.

97. Breslauer DN, Lee PJ, Lee LP (2006) Microfluidics-based systems biology. Mol Biosyst 2: 97–112. Available: http://www.ncbi.nlm.nih.gov/pubmed/16880927. Accessed 23 March 2014.

98. Thorslund S (2006) Microfluidics in Surface Modified PDMS Towards Miniaturized Diagnostic Tools.

99. Wu Z, Willing B, Bjerketorp J, Jansson JK, Hjort K (2009) Soft inertial microfluidics for high throughput separation of bacteria from human blood cells. Lab Chip 9: 1193–1199. Available: http://www.ncbi.nlm.nih.gov/pubmed/19370236. Accessed 12 August 2013.

100. Hansson J (2012) Microfluidic blood sample preparation for rapid sepsis diagnostics.

83

101. Klabunde RE (n.d.) Cardiovascular Physiology Concepts. Available: http://www.cvphysiology.com/Hemodynamics/H006.htm.

102. Hardt S, Schönfeld F (n.d.) Chapter 1 Microfluidics  : Fundamentals and Engineering. Microfluidics: Fundamentals and Engineering Concepts.

103. Lim CT, Zhang Y (2007) Bead-based microfluidic immunoassays: the next generation. Biosens Bioelectron 22: 1197–1204. Available: http://www.ncbi.nlm.nih.gov/pubmed/16857357. Accessed 25 February 2014.

104. Bange A, Halsall HB, Heineman WR (2005) Microfluidic immunosensor systems. Biosens Bioelectron 20: 2488–2503. Available: http://www.ncbi.nlm.nih.gov/pubmed/15854821. Accessed 16 July 2014.

105. Lin C-C, Wang J-H, Wu H-W, Lee G-B (2010) Microfluidic Immunoassays. J Assoc Lab Autom 15: 253–274. Available: http://jla.sagepub.com/lookup/doi/10.1016/j.jala.2010.01.013. Accessed 23 September 2013.

106. Liu W-T, Zhu L, Qin Q-W, Zhang Q, Feng H, et al. (2005) Microfluidic device as a new platform for immunofluorescent detection of viruses. Lab Chip 5: 1327–1330. Available: http://www.ncbi.nlm.nih.gov/pubmed/16234960. Accessed 8 March 2014.

107. Davidsson R (2004) MICROFLUIDIC BIOSENSING SYSTEMS BASED ON ENZYMES , ANTIBODIES AND CELLS.

108. Mairhofer J, Roppert K, Ertl P (2009) Microfluidic systems for pathogen sensing: a review. Sensors (Basel) 9: 4804–4823. Available: http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=3291940&tool=pmcentrez&rendertype=abstract. Accessed 19 February 2014.

109. Pennathur S, Meinhart CD, Soh HT (2008) How to exploit the features of microfluidics technology. Lab Chip 8: 20–22.

84

Available: http://www.ncbi.nlm.nih.gov/pubmed/18094758. Accessed 12 June 2014.

110. Ratner B (1996) Bioaterials science - an introduction to materials in medicine. Academic Press, San diego.

111. Thorslund S (2006) Microfluidics in Surface Modified PDMS Towards Miniaturized Diagnostic Tools. Acta Universitatis Uppsaliensis Uppsala.

112. Hemostasis Research Group at Department of Clinical Chemistry. (n.d.) Biomaterial-blood interaction. Available: http://platelets.se/biomaterial/.

113. Bhagat AAS, Bow H, Hou HW, Tan SJ, Han J, et al. (2010) Microfluidics for cell separation. Med Biol Eng Comput 48: 999–1014. Available: http://www.ncbi.nlm.nih.gov/pubmed/20414811. Accessed 24 March 2014.

114. Beech JP (n.d.) SEPARATION AND ANALYSIS OF BIOLOGICAL PARTICLES.

115. Sorting FC (2012) M ICROFLUIDIC  : AN INNOVATIVE TOOL. 57: 297–307.

116. Gossett DR, Weaver WM, Mach AJ, Hur SC, Tse HTK, et al. (2010) Label-free cell separation and sorting in microfluidic systems. Anal Bioanal Chem 397: 3249–3267. Available: http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=2911537&tool=pmcentrez&rendertype=abstract. Accessed 21 January 2014.

117. Tomlinson MJ, Tomlinson S, Yang XB, Kirkham J (2013) Cell separation: Terminology and practical considerations. J Tissue Eng 4: 2041731412472690. Available: http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=3578272&tool=pmcentrez&rendertype=abstract. Accessed 26 March 2014.

118. Lenshof A, Laurell T (2010) Continuous separation of cells and particles in microfluidic systems. Chem Soc Rev 39: 1203–1217.

85

Available: http://www.ncbi.nlm.nih.gov/pubmed/20179832. Accessed 20 March 2014.

119. Linda Johansson (2009) Acoustic manipulation of particles and fluids in microfluidic systems. Acta Universitatis Upsaliensis. Digital Comprehensive Summaries of Uppsala Dissertations from the Faculty of Science and TEchnology 641.

120. Di Carlo D, Irimia D, Tompkins RG, Toner M (2007) Continuous inertial focusing, ordering, and separation of particles in microchannels. Proc Natl Acad Sci U S A 104: 18892–18897. Available: http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=2141878&tool=pmcentrez&rendertype=abstract.

121. Di Carlo D (2009) Inertial microfluidics. Lab Chip 9: 3038–3046. Available: http://www.ncbi.nlm.nih.gov/pubmed/19823716. Accessed 19 February 2014.

122. Zola H, Health C (2001) Monoclonal Antibodies  : Diagnostic Uses. ENCYCLOPEDIA OF LIFE SCIENCES. Nature Publishing Group. pp. 1–9. Available: www.els.net.

123. Saleem M, Kamal M (2008) Monoclonal antibodies in clinical diagnosis  : A brief review application. 7: 923–925.

124. Borrebaeck CAK (2000) immunoassays to protein chips. 5699: 379–382.

125. Sullivan MJO (n.d.) Immunoassays. Principles of Immunopharmacology. pp. 171–181.

126. Wu AHB (2006) A selected history and future of immunoassay development and applications in clinical chemistry. Clin Chim Acta 369: 119–124. Available: http://www.ncbi.nlm.nih.gov/pubmed/16701599. Accessed 19 August 2013.

127. Lodish H, Berk A, Zipursky S, Matsudaira P, Baltimore D, et al. (2000) Molecular Cell Biology. In: Freeman W., editor. New York. Available: http://www.ncbi.nlm.nih.gov/books/NBK21665/.

86

128. Silhavy TJ, Kahne D, Walker S (2010) The bacterial cell envelope. Cold Spring Harb Perspect Biol 2: a000414. Available: http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=2857177&tool=pmcentrez&rendertype=abstract. Accessed 23 May 2014.

129. Sethu P, Anahtar M, Moldawer LL, Tompkins RG, Toner M (2004) Continuous flow microfluidic device for rapid erythrocyte lysis. Anal Chem 76: 6247–6253. Available: http://www.ncbi.nlm.nih.gov/pubmed/15516115.

130. Chen X, Cui D-F (2009) Microfluidic devices for sample pretreatment and applications. Microsyst Technol 15: 667–676. Available: http://link.springer.com/10.1007/s00542-009-0783-8. Accessed 22 April 2014.

131. Sethu P, Moldawer LL, Mindrinos MN, Scumpia PO, Tannahill CL, et al. (2006) Microfluidic isolation of leukocytes from whole blood for phenotype and gene expression analysis. Anal Chem 78: 5453–5461. Available: http://www.ncbi.nlm.nih.gov/pubmed/16878882.

132. Raetz CRH, Whitfield C (2002) Lipopolysaccharide endotoxins. Annu Rev Biochem 71: 635–700. Available: http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=2569852&tool=pmcentrez&rendertype=abstract. Accessed 8 August 2013.

133. Hitchcock PJ, Leive L, Mäkelä PH, Rietschel ET, Strittmatter W, et al. (1986) Lipopolysaccharide nomenclature--past, present, and future. J Bacteriol 166: 699–705. Available: http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=215174&tool=pmcentrez&rendertype=abstract.

134. Hwang K-Y, Lim H-K, Jung S-Y, Namkoong K, Kim J-H, et al. (2008) Bacterial DNA sample preparation from whole blood using surface-modified Si pillar arrays. Anal Chem 80: 7786–7791. Available: http://www.ncbi.nlm.nih.gov/pubmed/18729477.

135. Hwang K-Y, Jeong S-Y, Kim Y-R, Namkoong K, Lim H-K, et al. (2011) Rapid detection of bacterial cell from whole blood: Integration of DNA sample preparation into single micro-PCR chip. Sensors Actuators B Chem 154: 46–51. Available:

87

http://linkinghub.elsevier.com/retrieve/pii/S0925400509008703. Accessed 30 June 2014.

136. Lee J, Jeong KJ, Hashimoto M, Kwon AH, Rwei A, et al. (2014) Synthetic Ligand-Coated Magnetic Nanoparticles for Micro fl uidic Bacterial Separation from Blood: 1–5.

137. Xia N, Hunt TP, Mayers BT, Alsberg E, Whitesides GM, et al. (2006) Combined microfluidic-micromagnetic separation of living cells in continuous flow. Biomed Microdevices 8: 299–308. Available: http://www.ncbi.nlm.nih.gov/pubmed/17003962. Accessed 2 April 2014.

138. Yung CW, Fiering J, Mueller AJ, Ingber DE (2009) Micromagnetic-microfluidic blood cleansing device. Lab Chip 9: 1171–1177. Available: http://www.ncbi.nlm.nih.gov/pubmed/19370233. Accessed 22 August 2013.

139. White AG (2012) Selective recognition of bacteria utilizing zinc(II)Dipicolylamine conjugated far-red fluorescent probes. University of Notre Dame.

140. White AG, Fu N, Leevy WM, Lee J, Blasco MA, et al. (2011) Optical Imaging of Bacterial Infection in Living Mice using Deep-Red Fluorescent Squariaine Rotaxane Probes. 21: 1297–1304. doi:10.1021/bc1000998.Optical.

141. Leevy WM, Serazin N, Smith BD (2010) Optical imaging of bacterial infection models: 1–13. doi:10.1016/j.ddmod.2007.07.001.Optical.

142. Leevy WM, Johnson JR, Lakshmi C, Morris J, Marquez M, et al. (2006) Selective recognition of bacterial membranes by zinc(II)-coordination complexes. Chem Commun (Camb) 1: 1595–1597. Available: http://www.ncbi.nlm.nih.gov/pubmed/16582990. Accessed 4 June 2014.

143. Leevy MW, Labert TN, Johnson JR, Morris J, Smith BD (2010) Quantum Dot Probes for Bacteria Distinguish Escherichia coli. 1: 2331–2333. doi:10.1039/b803590c.Quantum.

88

144. Lee J-J, Jeong KJ, Hashimoto M, Kwon AH, Rwei A, et al. (n.d.) Supporting Information Synthetic ligand-coated magnetic nanoparticles for microfluidic bacterial separation from blood: 1–22.

145. One shot Stbl3 chemically competent E.coli (n.d.). Life Technol. Available: http://www.lifetechnologies.com/order/catalog/product/C737303. Accessed 15 July 2014.

146. Klena J, Zhang P, Schwartz O, Hull S, Chen T (2005) The Core Lipopolysaccharide of Escherichia coli Is a Ligand for the Dendritic-Cell-Specific Intercellular Adhesion Molecule Nonintegrin CD209 Receptor. 187: 1710–1715. doi:10.1128/JB.187.5.1710.

147. Kang JH, Super M, Yung CW, Cooper RM, Domansky K, et al. (2014) An extracorporeal blood-cleansing device for sepsis therapy. Nat Med. Available: http://www.ncbi.nlm.nih.gov/pubmed/25216635. Accessed 15 September 2014.

148. Opal SM, Scannon PJ, Vincent JL, White M, Carroll SF, et al. (1999) Relationship between plasma levels of lipopolysaccharide (LPS) and LPS-binding protein in patients with severe sepsis and septic shock. J Infect Dis 180: 1584–1589. Available: http://www.ncbi.nlm.nih.gov/pubmed/10515819.

149. Marshall JC, Walker PM, Foster DM, Harris D, Ribeiro M, et al. (2002) Measurement of endotoxin activity in critically ill patients using whole blood neutrophil dependent chemiluminescence. 6.

150. Prohm C, Tröltzsch F, Stark H (2013) Optimal control of particle separation in inertial microfluidics. Eur Phys J E Soft Matter 36: 118. Available: http://www.ncbi.nlm.nih.gov/pubmed/24154721. Accessed 22 July 2014.

151. Mach AJ, Di Carlo D (2010) Continuous scalable blood filtration device using inertial microfluidics. Biotechnol Bioeng 107: 302–311. Available: http://www.ncbi.nlm.nih.gov/pubmed/20589838. Accessed 6 August 2013.

89

152. Hou HW, Bhagat AAS, Lee WC, Huang S, Han J, et al. (2011) Microfluidic Devices for Blood Fractionation. Micromachines 2: 319–343. Available: http://www.mdpi.com/2072-666X/2/3/319/. Accessed 19 March 2014.

153. St-Louis P (2000) Status of point-of-care testing: promise, realities, and possibilities. Clin Biochem 33: 427–440. Available: http://www.ncbi.nlm.nih.gov/pubmed/11074234.

154. Price CP (2001) Clinical review Point of care testing: 1285–1288.

155. Chin CD, Chin SY, Laksanasopin T, Sia SK (2013) Point-of-Care Diagnostics on a Chip. In: Issadore D, Westervelt RM, editors. Biological andMedical Physics, Biomedical Engineering. Biological and Medical Physics, Biomedical Engineering. Berlin, Heidelberg: Springer Berlin Heidelberg. pp. 3–22. Available: http://link.springer.com/10.1007/978-3-642-29268-2. Accessed 17 January 2014.

156. Yager P, Domingo GJ, Gerdes J (2008) Point-of-care diagnostics for global health. Annu Rev Biomed Eng 10: 107–144. Available: http://www.ncbi.nlm.nih.gov/pubmed/18358075. Accessed 15 August 2013.

157. Leggieri N, Rida a, François P, Schrenzel J (2010) Molecular diagnosis of bloodstream infections: planning to (physically) reach the bedside. Curr Opin Infect Dis 23: 311–319. Available: http://www.ncbi.nlm.nih.gov/pubmed/20592531. Accessed 11 March 2014.

158. Niemz A, Ferguson TM, Boyle DS (2011) Point-of-care nucleic acid testing for infectious diseases. Trends Biotechnol 29: 240–250. Available: http://www.ncbi.nlm.nih.gov/pubmed/21377748. Accessed 13 August 2013.

159. Dineva MA, Mahilum-Tapay L, Lee H (2007) Sample preparation: a challenge in the development of point-of-care nucleic acid-based assays for resource-limited settings. Analyst 132: 1193. Available: http://xlink.rsc.org/?DOI=b705672a. Accessed 15 August 2013.

160. Hartman MR, Ruiz RCH, Hamada S, Xu C, Yancey KG, et al. (2013) Point-of-care nucleic acid detection using nanotechnology.

90

Nanoscale 5: 10141–10154. Available: http://www.ncbi.nlm.nih.gov/pubmed/24057263. Accessed 17 January 2014.

161. Olasagasti F, Ruiz de Gordoa JC (2012) Miniaturized technology for protein and nucleic acid point-of-care testing. Transl Res 160: 332–345. Available: http://www.ncbi.nlm.nih.gov/pubmed/22683416. Accessed 3 September 2013.

162. Cordray MS, Richards-Kortum RR (2012) Emerging nucleic acid-based tests for point-of-care detection of malaria. Am J Trop Med Hyg 87: 223–230. Available: http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=3414556&tool=pmcentrez&rendertype=abstract. Accessed 8 August 2013.

163. Ahmad F, Hashsham S a (2012) Miniaturized nucleic acid amplification systems for rapid and point-of-care diagnostics: a review. Anal Chim Acta 733: 1–15. Available: http://www.ncbi.nlm.nih.gov/pubmed/22704369. Accessed 13 August 2013.

164. Weigl B, Domingo G, Labarre P, Gerlach J (2008) Towards non-and minimally instrumented, microfluidics-based diagnostic devices. Lab chip 8: 1999–2014. doi:10.1039/b811314a.Towards.

165. Nanosphere (n.d.) Nanosphere, Verigene systems. Available: http://www.nanosphere.us/products/verigene-system. Accessed 1 August 2014.

166. Buchan BW, Ginocchio CC, Manii R, Cavagnolo R, Pancholi P, et al. (2013) Multiplex identification of gram-positive bacteria and resistance determinants directly from positive blood culture broths: evaluation of an automated microarray-based nucleic acid test. PLoS Med 10: e1001478. Available: http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=3699453&tool=pmcentrez&rendertype=abstract. Accessed 1 August 2014.

167. Buchan BW, Ginocchio CC, Manii R, Cavagnolo R, Pancholi P, et al. (2013) Supporting information: full study protocol including inclusion and exclusion criteria, BC-GP test method, and routine

91

culture method used in this study. Plos Med 10. Available: http://www.plosmedicine.org/article/info%3Adoi%2F10.1371%2Fjournal.pmed.1001478#s6.

168. Raich T, Powell S (2015) Identification of Bacterial and Fungal Pathogens from Positive Blood Culture Bottles: A Microarray-Based Approach. In: Mancini N, editor. Sepsis, Methods in Molecular Biology. Methods in Molecular Biology. New York, NY: Springer New York, Vol. 1237. pp. 73–90. Available: http://link.springer.com/10.1007/978-1-4939-1776-1. Accessed 30 October 2014.

169. Lefferts J a, Jannetto P, Tsongalis GJ (2009) Evaluation of the Nanosphere Verigene System and the Verigene F5/F2/MTHFR Nucleic Acid Tests. Exp Mol Pathol 87: 105–108. Available: http://www.ncbi.nlm.nih.gov/pubmed/19573527. Accessed 19 November 2014.

170. Pierce VM, Elkan M, Leet M, McGowan KL, Hodinka RL (2011) Comparison of the Idaho Technology FilmArray System to Real-Time PCR for Detection of Respiratory Pathogens in Children. J Clin Microbiol 50: 364–371. Available: http://jcm.asm.org/cgi/doi/10.1128/JCM.05996-11. Accessed 9 June 2014.

171. Altun O, Almuhayawi M, Ullberg M, Ozenci V (2013) Clinical evaluation of the FilmArray blood culture identification panel in identification of bacteria and yeasts from positive blood culture bottles. J Clin Microbiol 51: 4130–4136. Available: http://www.ncbi.nlm.nih.gov/pubmed/24088863. Accessed 12 March 2014.

172. Dineva MA, Mahilum-Tapay L, Lee H (2007) Sample preparation: a challenge in the development of point-of-care nucleic acid-based assays for resource-limited settings. Analyst 132: 1193. Available: http://xlink.rsc.org/?DOI=b705672a. Accessed 17 January 2014.

173. Reddington K, Tuite N, Minogue E, Barry T (2014) A current overview of commercially available nucleic acid diagnostics approaches to detect and identify human gastroenteritis pathogens. Biomol Detect Quantif 1: 3–7. Available: http://linkinghub.elsevier.com/retrieve/pii/S2214753514000035. Accessed 19 November 2014.

92

174. Ward C, Stocker K, Begum J, Wade P, Ebrahimsa U, et al. (2014) Performance evaluation of the Verigene® (Nanosphere) and FilmArray® (BioFire®) molecular assays for identification of causative organisms in bacterial bloodstream infections. Eur J Clin Microbiol Infect Dis. Available: http://www.ncbi.nlm.nih.gov/pubmed/25311986. Accessed 20 November 2014.

175. Poxton IR (1995) Antibodies to lipopolysaccharide. J Immunol Methods 186: 1–15. Available: http://www.ncbi.nlm.nih.gov/pubmed/7561138.

176. Gronow S, Brade H (2001) Invited review: Lipopolysaccharide biosynthesis: which steps do bacteria need to survive? J Endotoxin Res 7: 3–23. Available: http://ini.sagepub.com/cgi/doi/10.1177/09680519010070010301. Accessed 11 September 2013.

177. Siegel S a, Evans ME, Pollack M, Leone a O, Kinney CS, et al. (1993) Antibiotics enhance binding by human lipid A-reactive monoclonal antibody HA-1A to smooth gram-negative bacteria. Infect Immun 61: 512–519. Available: http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=302758&tool=pmcentrez&rendertype=abstract.

178. Mascelli M a, Frederick B, Ely T, Neblock DS, Shealy DJ, et al. (1993) Reactivity of the human antiendotoxin immunoglobulin M monoclonal antibody HA-1A with lipopolysaccharides from rough and smooth gram-negative organisms. Infect Immun 61: 1756–1763. Available: http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=280762&tool=pmcentrez&rendertype=abstract.

179. Hildner WK (1988) Immunolabeling of Lipopolysaccharide. 56: 2760–2762.

180. Bowden R a, Cloeckaert a, Zygmunt MS, Bernard S, Dubray G (1995) Surface exposure of outer membrane protein and lipopolysaccharide epitopes in Brucella species studied by enzyme-linked immunosorbent assay and flow cytometry. Infect Immun 63: 3945–3952. Available: http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=173554&tool=pmcentrez&rendertype=abstract.

93

181. Hackstadt T (1990) The role of lipopolysaccharides in the virulence of Coxiella burnetii. Ann N Y Acad Sci 590: 27–32. Available: http://www.ncbi.nlm.nih.gov/pubmed/2378455.

182. Monreal D, Grillo MJ, Gonza D, Marín CM, Miguel MJ De, et al. (2003) Characterization of Brucella abortus O-Polysaccharide and Core Lipopolysaccharide Mutants and Demonstration that a Complete Core Is Required for Rough Vaccines To Be Efficient against Brucella abortus and Brucella ovis in the Mouse Model. 71: 3261–3271. doi:10.1128/IAI.71.6.3261.

183. Chwalek M, Lalun N, Bobichon H, Plé K, Voutquenne-Nazabadioko L (2006) Structure-activity relationships of some hederagenin diglycosides: haemolysis, cytotoxicity and apoptosis induction. Biochim Biophys Acta 1760: 1418–1427. Available: http://www.ncbi.nlm.nih.gov/pubmed/16875785. Accessed 5 August 2014.

184. Walker JM (2010) Immunocytochemical Methods and Protocol. third edit. Oliver C, Jamur MC, editors New York, NY: Humana Press. doi:10.1007/978-1-59745-324-0.

185. Bennett-guerrero E, Mcintosh TJ, Robin G, Snyder DS, Gibbs RJ, et al. (2000) Preparation and Preclinical Evaluation of a Novel Liposomal Complete-Core Lipopolysaccharide Vaccine Preparation and Preclinical Evaluation of a Novel Liposomal Complete-Core Lipopolysaccharide Vaccine. doi:10.1128/IAI.68.11.6202-6208.2000.Updated.

186. Galanos C, Lüderitz O, Westphal O (1971) Preparation and properties of antisera against the lipid-A component of bacterial lipopolysaccharides. Eur J Biochem 24: 116–122. Available: http://www.ncbi.nlm.nih.gov/pubmed/4944989.

187. Di Padova FE, Brade H, Barclay GR, Poxton IR, Liehl E, et al. (1993) A broadly cross-protective monoclonal antibody binding to Escherichia coli and Salmonella lipopolysaccharides. Infect Immun 61: 3863–3872. Available: http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=281087&tool=pmcentrez&rendertype=abstract.

188. Greisman SE, Johnston C a. (1997) Review: Evidence against the hypothesis that antibodies to the inner core of lipopolysaccharides

94

in antisera raised by immunization with enterobacterial deep-rough mutants confer broad-spectrum protection during Gram-negative bacterial sepsis. Innate Immun 4: 123–153. Available: http://ini.sagepub.com/cgi/doi/10.1177/096805199700400206. Accessed 25 January 2013.

189. Caroff M, Karibian D (2003) Structure of bacterial lipopolysaccharides. Carbohydr Res 338: 2431–2447. Available: http://linkinghub.elsevier.com/retrieve/pii/S000862150300466X. Accessed 6 September 2013.

190. Marvin HJ, ter Beest MB, Witholt B (1989) Release of outer membrane fragments from wild-type Escherichia coli and from several E. coli lipopolysaccharide mutants by EDTA and heat shock treatments. J Bacteriol 171: 5262–5267. Available: http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=210360&tool=pmcentrez&rendertype=abstract.

191. Pelletier C, Bourlioux P, van Heijenoort J (1994) Effects of sub-minimal inhibitory concentrations of EDTA on growth of Escherichia coli and the release of lipopolysaccharide. FEMS Microbiol Lett 117: 203–206. Available: http://www.ncbi.nlm.nih.gov/pubmed/8181724.

192. Merck Millipore (n.d.) Turk’s solution. Available: http://www.merckmillipore.com/SE/en/product/Türk’s-solution,MDA_CHEM-109277. Accessed 11 August 2014.

193. Holst O (2011) Bacterial Lipopolysaccharides. In: Knirel YA, Valvano MA, editors. Bacterial Lipopolysaccharides. Vienna: Springer Vienna. pp. 21–40. Available: http://link.springer.com/10.1007/978-3-7091-0733-1. Accessed 11 September 2013.

194. Bellstedt DU, Human P a, Rowland GF, Van der Merwe KJ (1987) Acid-treated, naked bacteria as immune carriers for protein antigens. J Immunol Methods 98: 249–255. Available: http://www.ncbi.nlm.nih.gov/pubmed/2437205.

195. Weibull E, Antypas H, Kjäll P, Brauner A, Andersson-Svahn H, et al. (2014) Bacterial nanoscale cultures for phenotypic multiplexed antibiotic susceptibility testing. J Clin Microbiol. Available:

95

http://www.ncbi.nlm.nih.gov/pubmed/24989602. Accessed 6 August 2014.

196. Mohan R, Mukherjee A, Sevgen SE, Sanpitakseree C, Lee J, et al. (2013) A multiplexed microfluidic platform for rapid antibiotic susceptibility testing. Biosens Bioelectron 49: 118–125. doi:10.1016/j.bios.2013.04.046.