Detection of Enteroviruses in Tissue Samples

(1)

MAARIT OIKARINEN

Methods and applications in type 1 diabetes

ACADEMIC DISSERTATION To be presented, with the permission of

the Board of the School of Medicine of the University of Tampere, for public discussion in the Small Auditorium of Building M,

Pirkanmaa Hospital District, Teiskontie 35, Tampere, on June 10th, 2015, at 12 o’clock.

UNIVERSITY OF TAMPERE

(2)

MAARIT OIKARINEN

Detection of Enteroviruses in Tissue Samples

Methods and applications in type 1 diabetes

Acta Universitatis Tamperensis 2069 Tampere University Press

Tampere 2015

(3)

ACADEMIC DISSERTATION

University of Tampere, School of Medicine Finland

Reviewed by

Docent Arno Hänninen University of Turku Finland

Professor Ilkka Julkunen University of Turku Finland

Supervised by

Professor Heikki Hyöty University of Tampere Finland

Docent Sisko Tauriainen University of Jyväskylä Finland

Cover design by Mikko Reinikka

Acta Universitatis Tamperensis 2069 Acta Electronica Universitatis Tamperensis 1562 ISBN 978-951-44-9839-8 (print) ISBN 978-951-44-9840-4 (pdf )

ISSN-L 1455-1616 ISSN 1456-954X

ISSN 1455-1616 http://tampub.uta.fi

Suomen Yliopistopaino Oy – Juvenes Print Tampere 2015

Distributor:

verkkokauppa@juvenesprint.fi https://verkkokauppa.juvenes.fi

The originality of this thesis has been checked using the Turnitin OriginalityCheck service in accordance with the quality management system of the University of Tampere.

(4)

Abstract

Type 1 diabetes (T1D) is a disease in which insulin-producing beta cells in the islets of Langerhans in the pancreas are destroyed by an autoimmune process. The events leading to the development of the disease are complex and still partially unclear. A long preclinical phase defined by the presence of the diabetes-associated autoantibodies in blood precedes the onset of the disease. Genetic predisposition determines the susceptibility to develop T1D, but also environmental factors, of which enterovirus infections are one of the most evident candidates, play an important role. Numerous epidemiological studies have shown the connection between enteroviruses and T1D, and the detection of these viruses directly in the pancreas and other organs of diabetic patients has given new evidence of the role of enteroviruses in disease development. The aim of this study was to develop new immunohistochemical and in situ hybridization methods for enterovirus detection, and to use these methods for the analysis of tissue samples of diabetic and prediabetic subjects to further assess the role of these viruses in the process leading to T1D.

A panel of immunohistochemical and in situ hybridization methods was set-up for the detection of enteroviruses in tissue samples. Immunostaining with nine different antibodies or antibody combinations as well as in situ hybridization were optimized for formalin-fixed paraffin-embedded tissues, and immunostaining with seven different antibodies or antibody combinations for frozen tissues. The optimization and validation of the methods, as well as the evaluation of the ability of the hybridization probe and each antibody to detect different enterovirus serotypes, was carried out in enterovirus- infected cultured cells. Most antibodies as well as in situ hybridization worked well and proved to be a suitable tool for enterovirus detection in tissue samples.

The presence of enteroviruses and certain immunological markers was analyzed using immunohistochemical and in situ hybridization methods and RT-PCR from a prediabetic child who had accidentally died and who had been repeatedly positive for islet cell antibodies. The morphology of the pancreas was normal, and the only marker of a possible beta cell-damaging process was the strong production of beta2-microglobulin in the islets. The child was also positive for enterovirus in the pancreatic islets in

(5)

immunohistochemical staining with two different enterovirus-specific antibodies;

however, in situ hybridization assay was negative. All follow-up serum samples taken preceding his death turned out to be enterovirus-negative in RT-PCR.

The presence of enteroviruses in the small intestine of T1D patients was analyzed in two separate studies. In the first study, small intestinal biopsies of twelve T1D patients and ten controls were analyzed by immunohistochemistry with enterovirus VP1 protein- specific antibody and in situ hybridization. Nine of the patients (75 %) and one of the control subjects (10 %) were enterovirus positive in either in situ hybridization or immunohistochemistry (p=0,004). The presence of the virus was also confirmed by RT- PCR in some of the cases. In the second study, the same result was confirmed in a larger cohort which included intestinal biopsies from 39 T1D patients and 41 controls, as well as from 40 celiac disease patients. 74 % of the T1D patients were enterovirus-positive in in situ hybridization compared to 29 % of the control subjects (p<0,001), and the same trend was also seen in RT-PCR. The presence of enterovirus in in situ hybridization was also associated with the detection of inflammation markers in the duodenal mucosa.

In conclusion, methods for enterovirus detection in tissue samples were set-up and optimized. The panel included several antibodies or antibody combinations and in situ hybridization assay which can be used for the detection of enteroviruses in both formalin-fixed paraffin-embedded and frozen tissue samples. These methods were applied in the analysis of pancreatic tissue of a prediabetic child, as well as in the examination of small intestine biopsy samples of T1D patients. The results suggest that enteroviruses can be detected in the pancreas and duodenum of T1D patients. These findings further strengthen the hypothesis of a pathogenic role of enteroviruses in the development of type 1 diabetes.

(6)

Tiivistelmä

Tyypin 1 diabetes (T1D) johtuu haiman insuliinia tuottavien beetasolujen tuhoutumisesta immunologisten mekanismien kautta. Tapahtumaketju, joka johtaa sairastumiseen, on monimutkainen ja vielä osittain epäselvä. Taudin puhkeamista edeltää pitkä oireeton vaihe, jonka merkkinä verenkierrossa esiintyy autovasta-aineita beetasolun rakenteita kohtaan. Geneettinen alttius määrittelee taipumuksen sairastua T1D:een, mutta myös ympäristötekijät ovat tärkeässä osassa taudin synnyssä. Enteroviruksia on pitkään epäilty taudin syntyyn myötävaikuttaviksi tekijöiksi. Useat epidemiologiset tutkimukset ovat osoittaneet T1D:n ja enterovirusten välisen yhteyden, ja virusten löytyminen diabeetikoiden haimasta ja muista kudoksista on vahvistanut epäilyjä enterovirusten osuudesta taudin synnyssä. Tämän tutkimuksen tarkoituksena oli kehittää immunohistokemiaan ja in situ -hybridisaatioon perustuvia menetelmiä enterovirusten osoittamiseen sekä soveltaa näitä menetelmiä diabeetikoiden ja esidiabeetikoiden kudosnäytteiden analysointiin.

Immunovärjäys yhdeksällä eri enterovirusvasta-aineella tai -vasta-aineyhdistelmällä sekä in situ -hybridisaatio viruksen RNA-genomiin sitoutuvalla koettimella optimoitiin formaliinilla fiksoiduille ja parafiiniin valetuille enteroviruksella infektoituja soluja sisältäville leikkeille. Vastaavasti immunovärjäys seitsemällä eri vasta-aineella tai vasta- aineyhdistelmällä optimoitiin jääleikkeille, ja menetelmien kyky tunnistaa eri enterovirustyyppejä selvitettiin. Useimmat vasta-aineet ja vasta-aineyhdistelmät sekä in situ -hybridisaatio toimivat hyvin ja osoittautuivat hyväksi työkaluksi enterovirusten osoittamiseen tämän tyyppisistä näytteistä.

Enterovirusten ja eräiden immunologisten markkereiden esiintyvyyttä tutkittiin näillä immunohistokemiallisilla ja in situ -hybridisaatiomenetelmillä sekä RT-PCR- menetelmällä diabeetikoiden ja esidiabeetikoiden kudosnäytteistä. Haimanäyte tutkittiin lapselta, jolla oli todettu diabetekseen liittyviä saarekesoluvasta-aineita. Haiman morfologia oli normaali, ja ainoa merkki mahdollisesta immunologisesta reaktiosta oli beta2-mikroglobuliinin runsas tuotanto haiman saarekkeissa. Lapsen haimanäyte oli enteroviruspositiivinen kahdella eri enterovirusvasta-aineella testattuna, mutta in situ

(7)

-hybridisaatio antoi negatiivisen tuloksen. Myös lapsesta otetut seeruminäytteet olivat enterovirusnegatiivisia RT-PCR-analyysissä.

Virusten esiintyvyyttä tyypin 1 diabeetikoiden ohutsuolessa selvitettiin kahdessa eri tutkimuksessa. Ensimmäisessä tutkimuksessa kahdentoista diabeetikon ja kymmenen verrokin ohutsuolibiopsianäyte värjättiin enteroviruksen VP1-proteiiniin kohdistuvalla vasta-aineella ja viruksen genomin esiintymistä tutkittiin in situ -hybridisaatiolla. Kaiken kaikkiaan yhdeksän diabeetikkoa (75 %) ja yksi verrokki (10 %) olivat positiivisia joko immunohistokemialla tai in situ -hybridisaatiolla (p=0,004). Viruksen läsnäolo vahvistettiin RT-PCR:llä osassa tapauksista. Jälkimmäisessä tutkimuksessa sama tulos toistui suuremmassa aineistossa, johon kuului 39 diabeetikkoa ja 41 verrokkia sekä 40 keliaakikkoa. 74 % diabeetikoista oli enteroviruspositiivisia in situ -hybridisaatiolla verrattuna 29 %:iin verrokeista (p<0,001), ja sama suuntaus oli nähtävissä myös RT- PCR:llä. Positiivisuus in situ -hybridisaatiossa oli myös yhteydessä tulehdusreaktioon suolen limakalvolla.

Yhteenvetona voidaan todeta, että tutkimuksessa kehitettiin onnistuneesti uusia immunohistokemiallisia ja in situ -hybridisaatiomenetelmiä enterovirusten osoittamiseen soluja kudosnäytteistä. Menetelmiä sovellettiin tyypin 1 diabeetikoiden ohutsuolinäytteiden tutkimiseen, ja niiden avulla osoitettiin ensi kertaa enterovirusten esiintyminen diabeetikoiden ohutsuolen limakalvolla. Lisäksi saatiin merkkejä siitä, että virus voi esiintyä haiman saarekkeissa jo diabeteksen oireettoman esivaiheen aikana.

Nämä löydökset tukevat aiempia havaintoja enterovirusten ja tyypin 1 diabeteksen yhteydestä.

(8)

1 List of original publications

The study is based on the following original publications referred to in the text by their Roman numerals I-IV:

I Oikarinen M, Tauriainen S, Penttilä P, Keim J, Rantala I, Honkanen T, Hyöty H (2010): Evaluation of immunohistochemistry and in situ hybridization methods for the detection of enteroviruses using infected cell culture samples. J Clin Virol 47(3):

224-8

II Oikarinen M, Tauriainen S, Honkanen T, Vuori K, Vasama-Nolvi C, Oikarinen S, Verbeke C, Blair GE, Rantala I, Karhunen P, Ilonen J, Simell O, Knip M, Hyöty H (2008): Analysis of pancreas tissue in a child positive for islet cell antibodies.

Diabetologia 51(10):1796-802

III Oikarinen M, Tauriainen S, Honkanen T, Oikarinen S, Vuori K, Kaukinen K, Rantala I, Mäki M, Hyöty H (2008): Detection of enteroviruses in the intestine of type 1 diabetic patients. Clin Exp Immunol 151(1): 71-5

IV Oikarinen, M, Tauriainen, S, Oikarinen, S, Honkanen, T, Collin, P, Rantala, I, Mäki, M, Kaukinen, K, Hyöty, H (2012): Type 1 diabetes is associated with enterovirus infection in gut mucosa. Diabetes 61(3): 687-91

(13)

2 Abbreviations

A549 carcinomic human alveolar basal epithelial cells ATCC American Type Culture Collection

CAR coxsackievirus-adenovirus receptor CAV coxsackie A virus

CBV coxsackie B virus

MHC major histocompatibility complex COX-2 cyclooxygenase-2

CTLA cytotoxic T-lymphocyte antigen DAB diaminobenzidine

DAF decay-accelerating factor DIG digoxigenin

DIPP the Finnish Diabetes Prediction and Prevention Study

EV enterovirus

FFPE formalin-fixed paraffin-embedded GADA glutamic acid decarboxylase antibody GMK green monkey kidney cells

HEL-7 human fibroblast cells

HeLa carcinomic human cervix epithelial cells HLA human leukocyte antigen

HPeV human parechovirus HRP horse radish peroxidase HSP heat shock protein IA-2A islet antigen 2 antibody IAA insulin autoantibody

IAR/IA-2 protein tyrosine phosphatase ICA islet cell antibody

ICAM-1 intracellular adhesion molecule 1 IDIN IRF7-driven inflammatory network IEL intraepithelial lymphocyte

IFIH1 interferon induced with helicase C domain 1

IFN interferon

Ig immunoglobulin

IHC immunohistochemistry

IL interleukin

IRF7 Interferon regulatory factor 7 ISH in situ hybridization

JDFU Juvenile Diabetes Foundation unit MA104 vervet monkey kidney cells

(14)

NBT/BCIP nitro-blue tetrazolium and 5-bromo-4-chloro-3'-indolyphosphatase NCR non-coding region

NK cells natural killer cells

NOD mouse non-obese diabetic mouse

nPOD Network for Pancreatic Organ Donors with Diabetes OAS1 2'-5'-oligoadenylate synthetase 1

OCT optimal cutting temperature

PEVNET Persistent Virus Infection in Diabetes Network PKR protein kinase R

PTPN22 protein tyrosine phosphatase, non-receptor type 22

PV poliovirus

RT room temperature

RT-PCR reverse transcriptase polymerase chain reaction

T1D type 1 diabetes

TEDDY The Environmental Determinants of Diabetes in the Young study TRIGR Trial to Reduce IDDM in the Genetically at Risk study

TRITC tetramethylrhodamine ZnT8 zinc transporter-8

(15)

3 Introduction

Type 1 diabetes (T1D) is an autoimmune disease in which insulin-producing beta cells in the islets of Langerhans in the pancreas are destroyed. The preclinical phase preceding the onset of the clinical disease usually lasts for a relatively long period of time ranging from some months to several years. This subclinical phase of the disease can be detected by the appearance of diabetes-associated autoantibodies in blood.

The presence of such antibodies, particularly when multiple autoantibodies against two or more beta cell autoantigens appear, predicts the development of clinical T1D (1).

The detailed mechanisms of the beta cell damaging process have remained unknown.

According to current knowledge, the clinical symptoms appear when approximately 80-95 % of the original beta cell mass has been lost. However, the process is probably not linear and varies between individuals. In addition, individual variation in the initial beta cell mass can influence its progression (2).

The incidence of T1D is continuously increasing in most countries and is particularly high in Finland. Susceptibility to develop T1D is genetically determined and several genes have been identified which are associated with the risk of T1D (2). However, since only a small fraction of people with this genetic risk eventually develop T1D, environmental factors are believed to play an important role in the pathogenesis of T1D. For example, some dietary factors, such as cow’s milk protein and vitamin D intake, as well as some viruses, have been linked to the disease.

Enteroviruses have been connected to the development of T1D in numerous studies carried out during the last decades (3). Enteroviruses are known to have the ability to infect pancreas and pancreatic islets in animals and in man, and it is possible that such infection can damage beta cells directly by cytolytic effects or by triggering the autoimmune process eventually leading to beta cell damage and T1D.

Epidemiological studies have supported the association between enteroviruses and T1D; both enteroviral genome and antibodies against enteroviruses have been found more frequently in blood of T1D patients than in healthy controls. In addition to

(16)

epidemiological studies, the direct detection of enteroviruses in the pancreas has supported the causal relationship.

The purpose of this study was to develop new methods for the detection of enteroviruses in tissue samples and to apply these methods for enterovirus detection in different tissues of T1D and prediabetic patients.

(17)

4 Review of the literature

4.1 Enteroviruses

4.1.1 Classification, structure and replication

Enteroviruses belong to the family of Picornaviridae which consists of 17 genera:

enterovirus, cardiovirus, aphthovirus, hepatovirus, parechovirus, erbovirus, kobuvirus, teschovirus, sapelovirus, senecavirus, tremovirus, avihepatovirus, aquamavirus, cosavirus, dicipivirus, megrivirus and salivirus. Picornaviruses are small non-enveloped icosahedral RNA viruses with a diameter of about 30 nm. Virus particles are composed of a protein capsid and a positive-stranded RNA genome (Fig. 1). The capsid consists of structural proteins VP1, VP3 and VP0, or alternatively VP1, VP2, VP3 and VP4 in those picornaviruses in which VP0 undergoes a maturation cleavage during the final stages of the assembly leading to a formation of VP2 and VP4. VP1-3 form the outer shell of the virion and contain the main antigenic sites, whereas VP4 is located in the inner surface of the capsid. (4)

Figure 1. The structure of an enterovirus. The virus particle consists of a single-stranded positive- sense RNA genome surrounded by a capsid formed by four structural proteins VP1-VP4.

(The figure is presented at http://viralzone.expasy.org/)

(18)

Human enteroviruses are grouped into four species, human enteroviruses A (coxsackievirus A2–8, 10, 12, 14 and 16, and enterovirus 71, 76 and 89-91), human enteroviruses B (coxsackievirus A9, coxsackievirus B1–6, echovirus 1–7, 9, 11–21, 24–27 and 29–33, and enterovirus 69, 73-75, 77-88, 93, 97, 98, 100, 101, 106 and 107), human enteroviruses C (coxsackievirus A1, 11, 13, 17, 19–22 and 24, poliovirus 1–3, and enterovirus 95, 96, 99, 102, 104 and 105), and human enteroviruses D (enterovirus 68, 70 and 94). In addition, rhinoviruses are currently classified as members of the enterovirus genus. Species are grouped on the basis of sequence similarity in the VP1 genome region, and so far over 100 different enterovirus serotypes and over 150 rhinovirus serotypes have been identified. The number of identified serotypes is constantly increasing due to identification of new serotypes by ongoing sequencing studies of clinical virus isolates. Enteroviruses have a single- stranded positive sense RNA genome consisting of approximately 7000-7500 nucleotides (Fig. 2). A single open reading frame encodes for a long polyprotein which is during translation further cleaved into capsid proteins (VP1-4) and non- structural proteins (2A-C, 3A-D) which take part for example in RNA synthesis and protein processing. The genome has a highly structured and conserved 5’ noncoding region which is covalently linked to a protein called VPg, and a short 3’ noncoding region linked to a polyA tail. (4, 5)

Figure 2. The structure of enterovirus genome. (Adapted from

http://old.sinobiological.com/Biological-Characteristics-of-Enterovirus-71-EV71-a- 6404.html)

Enteroviruses can infect several human cells and tissues. The primary replication of the virus occurs in the lymphoid tissue of pharynx and small intestine, and during the viremic phase the virus can spread to various organs, such as the heart, central nervous system and the pancreas. The first step in the replication cycle is the attachment of the virus to its cellular receptor. Enteroviruses use a variety of different receptors to enter various cell types and the receptor-binding properties vary between

(19)

different virus types. These receptors include intracellular adhesion molecule 1 (ICAM-1), decay-accelerating factor (DAF), integrins ( 1, 3), poliovirus receptor and coxsackievirus-adenovirus receptor (CAR), but also other cell membrane molecules such as heparin sulphate can influence viral entry. After the entry into the cell, the virion is uncoated and the released RNA is directly translated in the cytoplasm to produce all the viral proteins required for viral replication and assembly. RNA synthesis takes place in replication vesicles where the positive- stranded RNA is copied through a negative-stranded intermediate with the help of non-structural proteins. Viral RNA assembles with structural proteins to form infectious virus particles which are released from the cell. (4, 6)

Enteroviruses cause in most cases acute infections, but they are also able to establish persistent infections in their host, which has been demonstrated both in cell cultures (7-9) and animal models (10-12). There is also some indication of persistent infections in humans, although direct evidence is still lacking (post-polio syndrome, dilated cardiomyopathy, chronic fatigue syndrome). Persistent replication of different strains of CBV3 and CBV4 has also been shown in cultured human pancreatic islet cells (13, 14). The interactions between the virus and the cell are critical for the development of persistence where the viral replication cycle is switched from lytic to a less destructive one. The lytic potential of the virus may be reduced due to mutations in the virus. Specific deletions have been described in the 5’-noncoding region in virus variants which establish such persistent infection (15). Such deletion can compromise the replication and protein synthesis of the virus leading to slow- grade persisting infection which can hide from the immune surveillance system. In addition,mutated amino acid residues in VP1 are suggested to be involved in receptor recognition or binding, which may promote viral persistence (16). During an acute enterovirus infection the amount of positive-stranded RNA is 50-100 times higher than negative-stranded RNA (14), whereas in persistent infection the amount of positive and negative-stranded RNA in infected cells is equal (17). It has been suggested that some chronic diseases in humans, such as dilated cardiomyopathy, post-polio syndrome and T1D, might be caused by a persistent enterovirus infection (18). Terminally deleted virus variants have also been detected in a patient with dilated cardiomyopathy (19). A persistent enterovirus infection in the pancreatic islets is suspected to play a part in T1D, and CBVs are able to establish a persistent infection in pancreatic ductal cell culture (20) and in mouse pancreas (21).

(20)

4.1.2 Enterovirus diseases

Enteroviruses are one of the most common pathogens across the world causing a wide range of different diseases, varying from mild asymptomatic infections to systemic, sometimes fatal multiorgan infections. Enteroviruses can be transmitted by both fecal-oral and respiratory routes, the former being predominant in areas with poor sanitary conditions and a lower standard of hygiene. Enterovirus infections in populations may be either sporadic or epidemic, and the epidemiology varies according to serotype, time, area and disease. The climate has a crucial impact on the incidence of enterovirus infections, as in temperate climates, such as the Nordic countries, the infections follow clear seasonality, most of the cases being reported during summer and early autumn, whereas in tropic countries the incidence is more constant throughout the year. (22)

Enterovirus infections are common in all age groups. They are usually asymptomatic or cause only mild flu-like symptoms. They can also cause eye infections, acute otitis media, hand, foot and mouth disease and more severe diseases including myocarditis, meningitis, encephalitis, pancreatitis, hemorrhagic systemic infection in newborns and paralysis. (22, 23) Enterovirus infections may also play a role in chronic diseases such as dilated cardiomyopathy (24), chronic fatigue syndrome (25) and T1D (3, 26, 27).

It is known that host factors regulate the outcome of enterovirus infection. Male gender and young age increase the risk of severe enterovirus disease (22). Infections in patients with humoral immunodeficiencies can be more severe or prolonged (28, 29). Some dietary factors also play a role, for example the deficiency of selenium is associated with severe coxsackievirus-induced myocarditis (30). On the other hand, breastfeeding is shown to protect against enterovirus infections (31), which is most probably due to protection by maternal antibodies present in breast milk. Animal models have shown that the host genes which regulate the innate immune response against enteroviruses have a strong effect on the course of the infection (32, 33).

(21)

4.2 Type 1 diabetes

4.2.1 Autoimmune process

Type 1 diabetes is an autoimmune disease where insulin-producing beta cells in the pancreatic islets of Langerhans are destroyed. A pioneering study published in 1965 by Gepts showed the reduction of beta cells and the infiltration of inflammatory cells, termed insulitis, in the pancreas of T1D patients (34). Recent studies have pointed out that this insulitis is usually mild and affects only a small fraction of the islets (35).

According to current knowledge, the clinical symptoms appear when approximately 80-95 % of the original beta cell mass has been lost (36). The preclinical phase usually lasts for several months or even years, showing considerable individual variation, and the actual mechanisms of the beta cell destruction still remain unknown. The appearance of diabetes-associated autoantibodies ICA (islet cell antibodies), IAA (insulin autoantibodies), GADA (glutamic acid decarboxylase antibody), IA-2A (islet antigen 2 antibody) and ZnT8 (zinc transporter-8 autoantibodies) against pancreatic antigens in peripheral blood is the first sign of an ongoing autoimmune process (37- 39). IAA are most frequently the first autoantibodies seen in young children who develop autoimmunity to islet cells, which suggests that insulin could be among the most important autoantigens in T1D (40, 41). The presence of two or more autoantibodies strongly predicts the development of T1D in a genetically susceptible child (1). In addition to autoantibodies, autoreactive T cells against beta cell proteins have been detected in the peripheral blood of T1D patients (42).

It has been difficult to study the local immunological process in the pancreas since the retroperitoneal location of the organ makes it difficult to take samples from living patients. Most human studies have been done on autopsy materials when the tissues have been exposed to post-mortem changes. Recently, new studies have been started aiming at collecting fresh tissues from brain dead cadaver organ donors (nPOD

(http://www.jdrfnpod.org/) and PEVNET

(http://www.uta.fi/med/pevnet/index.html) studies). A feasibility study was originally carried out in Finland (43). Mouse models have been used to study the mechanisms of beta cell autoimmunity (mostly in NOD mouse model) but it is not clear how well these animal models reflect the pathogenesis of human T1D. Overall, the results from human studies suggest that CD8+ cytotoxic T cells and macrophages are the most abundant immune cells in insulitis; CD4+ helper T cells and B

(22)

lymphocytes are also present but in lower numbers (44, 45). Increased expression of class I MHC molecules in pancreatic islets, and occasionally in vascular endothelium, is an early marker in the process leading to T1D and it is believed to precede insulitis (46, 47). Besides, some studies have also reported overexpression of class II MHC molecules in the pancreas of recent-onset T1D patients, but distinct from class I MHC, it is restricted to beta cells (47, 48). However, the detailed mechanisms by which the adaptive immune system damages beta cells in T1D are not known. In addition, the innate immune system may also play a role since interferon-alpha and protein kinase R (PKR) are shown to be expressed in the pancreatic islets of T1D patients as a sign of an activation of innate immune system (49, 50).

4.2.2 Genetic susceptibility

The susceptibility to develop T1D is genetically determined. Approximately 60 different genes have been shown to modulate the risk for the disease, the major determinants being the highly polymorphic human leukocyte (HLA) genes of the major histocompatibility complex (MHC), accounting for 42 % of the familial inheritance of the disease (51). The risk associated with T1D is defined by a particular combination of predisposing and protecting alleles in HLA class II DR and DQ gene regions. DRB1 and DQB1 genes are the strongest risk components, but other MHC genes may modify their influence as well. The mechanisms of HLA-associated susceptibility and protection events are complex and remain presently unknown. The molecules encoded by HLA-DR and -DQ genes are thought to play a central role in the activation of autoreactive T cells which destroy beta cells, and the specific structure of the molecules is probably the key determinant in their interaction with T cells. (52) However, it is also possible that the effect of HLA genes is mediated by their effects on immune responses against external agents such as viruses.

In addition to the HLA region, up to 50 other genes have been identified to be associated with T1D. These include insulin gene locus (INS) (53, 54), CTLA locus (55) and PTPN22 gene (56). Rare variants of recently found IFIH1 gene, which codes for an intracellular receptor of viral dsRNA, are also strongly protective against T1D (57, 58). In addition, T1D-associated polymorphisms in PTPN22 gene regulate innate immune responses against enteroviruses (59).

(23)

Despite the influence of several genes on child’s susceptibility to develop T1D, genetic factors are evaluated to account only for approximately 50 % of the overall disease susceptibility (60), and the studies in genetically identical twins have shown the disease concordance to be only about 40 % (61, 62). In addition, less than 10 % of the individuals having HLA risk genes develop clinical diabetes (63). Among newly diagnosed patients, the amount of individuals carrying neutral or protective HLA genotypes has increased while the risk genotypes have become less common (64).

These facts propose that the role of environmental factors has increased in the pathogenesis of T1D during the past decades.

4.2.3 Environmental factors

Several environmental factors have been connected to the initiation and acceleration of T1D. Some general features of the disease support the role of environment in the pathogenesis. For example, the incidence of T1D has been increasing rapidly throughout the world (65). In addition, marked differences in the incidence of T1D in different geographical regions have been reported (66), and even neighboring populations which have a similar genetic background can have significantly different incidence of T1D (67). Migration studies also support the role of environmental factors (68). The Asian children who migrated to UK had an increased incidence of T1D compared to that seen in Asia (69-71). In Finland, the prevalence of T1D is similarly high in both Finnish children and Somali children (72). These changes in the incidence rate and in the geographical locations suggest that the genetic determinants are not exclusively contributing to the disease. The current knowledge is that environmental factors trigger an autoimmune process leading to the destruction of beta cells in genetically susceptible individuals (73). Different combinations of genetic and environmental factors may be needed to trigger the disease in different individuals (74). Environmental factors may also have opposing effects, some increasing the risk of T1D while some others can reduce the risk (75).

A number of dietary factors have been linked to the pathogenesis of T1D. The duration of breastfeeding and the incidence of T1D have been shown to have an inverse correlation (31, 76), and the age of exposure to supplementary milk feeding is related to the risk of T1D (77). Specifically, the role of cow’s milk proteins including bovine serum albumin (78, 79), -lactoglobulin (80), beta casein (81), and insulin (82) have been implicated as potential triggers of the disease. A large

(24)

intervention trial (TRIGR) which is based on the elimination of early exposure to cow’s milk proteins is in progress and pilot trial has provided evidence supporting possible beneficial effect of this intervention (83). However, the recent findings from the main TRIGR study showed that the use of hydrolyzed casein formula (which contains no intact proteins) does not reduce the incidence of diabetes-associated autoantibodies compared to conventional cow’s milk-based formula in children at genetic risk for T1D (84). Gluten intake has also been studied as a potential risk factor for T1D but a recent clinical trial studying the effect of delayed exposure to gluten did not reduce the risk of T1D (85, 86). Vitamin D supplementation has been associated with a reduced risk for T1D (87). Some toxins, such as N-nitroso compounds which are formed from dietary nitrites, have been linked to the development of T1D (88). Virus infections have also been associated with T1D.

Suspected viruses include rotavirus (89), mumps virus (90), cytomegalovirus (91), rubella virus (92, 93), and most convincingly certain enteroviruses (26).

However, even if the role of environmental factors seems to be important in the pathogenesis of T1D, none of the suspected factors has been proved to be causally related to the disease. This reflects the complex nature of the disease and the difficulties to identify the critical environmental factors among all the others. In addition, such factors may have mutual interactions and different factors may play a role in different subgroups of T1D patients.

4.2.4 Diabetes and the gut

The gastrointestinal tract has an important role in the development of many immune- mediated diseases. Besides absorbing nutrients, intestinal epithelial cells also contain Toll-like receptors and other signaling mechanisms for the transduction of inflammatory signals from the intestinal lumen. In fact, the gut-associated lymphoid tissue constitutes the largest lymphoid organ in the human body. Accumulating evidence suggests that the gut has an important role in the pathogenesis of T1D.

Intestinal microbiota, permeability of the intestinal mucosal barrier and mucosal immunity have all been linked to T1D (94). Environmental factors can modify the normal gut microbiome or change the gut cytokine milieu. Such an effect is typical for example for gastrointestinal infections and antibiotic treatments. (95, 96)

(25)

Previous studies have demonstrated that the small intestine of T1D patients shows enhanced immune activation. Specifically, the expression of HLA class II molecules (HLA-DR and HLA-DP), intracellular adhesion molecule 1 (ICAM-1), 7 integrin and some cytokines (IL-4, IL-1 , IFN- ) has been observed to be expressed in higher levels in small intestinal biopsy samples of T1D patients (97, 98). In addition, high numbers of intraepithelial CD3 and cells, as well as CD25 cells in lamina propria, have been reported in T1D patients (99, 100). It has also been demonstrated that children with T1D show increased expression of matrix metalloproteinases and increased amounts of apoptotic cells in intestinal mucosal compared to children without T1D (101). The significance of immune activation in the pathogenesis of T1D is not known. However, it is possible that the lymphocytes which are activated in the gut mucosa can spread into the pancreas, since the pancreas and the gut share common homing receptors for lymphocytes (102-104).

Growing evidence has shown gut microbiome to play an important role in the development of T1D by modelling mucosal immunity (94, 105, 106). Animal studies have demonstrated that altered gut microbiome is strongly associated with T1D and that a so-called protective gut flora might delay or even prevent the development of T1D (107-111). Human studies have shown that altered gut microbiome is strongly related to beta cell autoimmunity and T1D (112-115) and especially the development of healthy microbiome in the early childhood has been shown to be important (116- 118), although the mechanisms by which the gut microbiome interacts with host immunity and changes beta cell autoimmunity remain unclear.

The possible role of gut in the pathogenesis of T1D has also been supported by observations showing that T1D correlates with other autoimmune diseases such as the celiac disease (119). Both diseases are associated with HLA-DR3 and HLA-DQ2 genotypes, which support the idea of a common pathogenetic component. Gluten may indeed have a causative role in diabetes, as altered intestinal immune response to gluten has been reported in T1D (120). However, a recent follow-up study did not find any outcome of an early gluten exposure to the development of T1D-associated autoimmunity (121). It has also been suggested that some patients with celiac disease may have subclinical insulitis in the pancreas and that intestinal inflammation may stimulate beta cell-specific immune response measured by diabetes-associated autoantibodies. (96)

(26)

4.3 The link between enterovirus infections and type 1 diabetes

4.3.1 Epidemiological studies

The connection between enterovirus infections and T1D has gained support from several epidemiological studies. The studies carried out by Gamble and Taylor in 1969 described this connection for the first time. They showed that the seasonality of T1D and enterovirus infections show a similar pattern; both show a peak in autumn (122). They also found that enterovirus antibodies were more frequent in patients with T1D compared to control subjects (123). The seasonal pattern of T1D onset has later been confirmed by other studies and it has also been observed at the onset of the subclinical process; children participating in the prospective birth cohort study DIPP turned positive for diabetes-associated autoantibodies most frequently during the season when enterovirus infections are most common (124). In the USA, T1D incidence was increased in persons under 20 years of age after a CBV5 epidemic which began in 1983 (125), and a clear relationship was seen in children 18 years old or younger between T1D and enterovirus IgM positivity (126). After that, several studies have been published which have measured antibodies against enteroviruses in T1D patients and control subjects. Altogether, these studies have showed variable results some of which indicating a risk association while some failed to find such a link. (60, 127). Besides enterovirus antibodies, enteroviral genome has been found in blood of T1D patients in many cross-sectional studies using RT-PCR. These studies are summarized in Table 1. A very recent meta-analysis (128) systematically reviewed 33 T1D prevalence studies from 1990 to 2010 and found a significant association between enterovirus infections and both T1D-related autoimmunity and clinical T1D.

(27)

Table 1. Summary of retrospective case-control studies showing enteroviral RNA in the blood of type 1 diabetic patients.

Patients Controls

Country Positive N Positive N Reference

UK 64 % 14 4 % 45 (129)

UK 27 % 110 5 % 182 (130)

France 42 % 23 0 % 27 (131)

France 38 % 56 0 % 37 (132)

Sweden 50 % 24 0 % 24 (133)

Australia 30 % 206 4 % 160 (134)

Japan 38 % 61 3 % 58 (135)

Germany 20-36 % 50+47 2 % 50 (136)

Cuba 16-27 % 34+32 0-3 % 68+64 (137)

Netherlands 2 % 10 0 % 20 (138)

Prospective studies based on a follow-up of initially healthy individuals have given additional evidence for the link between T1D and enterovirus infections. The first prospective studies were done in Finland (139, 140), which showed that enteroviruses were more common in children who later developed T1D than in control children.

In addition, enterovirus infections seemed to cluster to the time when islet antibodies appeared. Additional prospective studies in the Finnish population have confirmed these findings (141-144, 144-147). It has also been suggested that enterovirus infections during pregnancy may as well increase the risk of T1D in the offspring although their risk effect seems to be relatively weak (139, 148-150). However, two prospective studies showed no association between enterovirus infections and T1D

(28)

(151, 152). These two negative studies were based on the detection of enterovirus RNA from stool samples while the studies showing a risk association were based on the detection of the virus or virus antibodies from longitudinal serum samples. Thus, it is possible that the enterovirus-T1D association is linked to invasive infection where the virus spreads to blood. In addition, these studies have not distinguished different enterovirus serotypes from each other, and the high overall frequency of enterovirus infections could mask the possible effect of some specific viral subtypes.

Recent studies have suggested that the risk-association between enteroviruses and T1D can only be mediated by some specific enterovirus types. A large prospective study in Finland found that among the 41 different serotypes studied only the group B coxsackieviruses showed this association (153). This study was based on the measurement of neutralizing antibodies from follow-up serum samples of children who later turned positive for multiple autoantibodies or developed clinical diabetes.

This finding was also validated in a case-control study among newly diagnosed patients recruited in different European populations (154). This group of enteroviruses included six different serotypes, and one of them (coxsackivirus B1) showed an association with T1D while the others attenuated the risk effect of CVB1 in a manner suggestive of immunological cross-protection (153). CBVs have been connected to T1D also in previous studies and fatal CBV infections have been associated with cell destruction in the pancreatic islets (155). CBVs have also been isolated from the pancreas of T1D patients (156, 157). In addition to CBVs, some other species of enteroviruses have been linked to induction of beta cell autoimmunity as well (158). Thus, based on the current evidence it is possible that only certain specific enterovirus types are able to cause T1D, and CBV group of enteroviruses are the main suspects. In this scenario it would be important to use assays which can specifically detect these virus types in T1D and control subjects.

An association between coxsackievirus B infection and elevated levels of IFN- , a marker of a virus infection, has been reported in blood of T1D patients (159). IFN-

expression in the beta cells of T1D patients has also been detected using immunohistochemistry and IFN- mRNA measurements (49, 160). Genome-wide gene expression studies from peripheral blood have also indicated that innate immune system is activated during a few months long time period preceding the first detection of autoantibodies (161, 162). Interestingly, an excess of respiratory infections has been reported during this same time period in prospective studies (163,

(29)

164), and enterovirus infections have been diagnosed in this time window more frequently than in control children (153).

Some studies have suggested that enterovirus infections could also contribute to the increasing incidence of type 1 diabetes. Enterovirus infections have become less frequent during the last decades (165) and at the same time the incidence of T1D has increased. In addition, enterovirus infections were found to be less frequent in Finland where the incidence of T1D is the highest in the world, compared to the frequency of enterovirus infections in other European countries with lower incidence of T1D (165, 166). These findings led to the introduction of polio hypothesis, which refers to the similar situation in polio epidemics which started to increase at the end of the 19th century when poliovirus infections simultaneously decreased (167). This phenomenon might be caused by the first poliovirus infections appearing at an older age when maternal antibodies are not any longer protecting the child. According to this hypothesis, the protection from enterovirus infections in neonatal period has decreased because of the shift in the infections to an older age. Another hypothesis possibly linking enterovirus infections and the increase in T1D incidence is the hygiene hypothesis. According to this hypothesis, enterovirus infections have a beneficial effect by down-regulating the immune responses by a bystander suppression mechanism (168, 169).

4.3.2 Detection of enteroviruses in the pancreas of type 1 diabetic patients

In addition to epidemiological studies, an association between enteroviruses and T1D has been evaluated by searching for enteroviruses in the pancreatic tissue of T1D patients. This work has been based on the hypothesis that enteroviruses may infects beta cells and remain detectable in the pancreas for such a long period of time that it can be found in at least some patients at the time of clinical diagnosis of T1D.

Obviously, in case of viral persistence the virus should be detectable for longer periods of time and at different stages of the diabetic process assuming that enough virus proteins and RNA are produced. Thus, these studies mainly look for possible viral persistence. This means that the possible absence of the virus in T1D patients would not necessarily exclude the role of infections which could have occurred earlier and started the beta cell damaging process but the virus had already disappeared by the time of clinical diagnosis of T1D.

(30)

Immunohistochemistry (IHC) and in situ hybridization (ISH) are the most widely used methods for the detection of enterovirus in tissue samples. IHC is based on specific antibodies against enterovirus proteins, whereas ISH uses a probe designed to hybridize with enteroviral genome. Human pancreatic tissue has usually been obtained from autopsy or organ donation, although laparoscopic pancreatic biopsies have occasionally been performed (170-172). Histological staining methods have been developed for the detection of enteroviruses in both formalin-fixed paraffin- embedded (FFPE) and frozen samples. These methods enable the determination of the exact localization of virus replication and, using immunofluorescence double- staining with specific antibodies against enterovirus and pancreatic islet hormones (insulin, glucagon and somatostatin), the localization of virus in different islet cell types can also be studied. The studies carried out so far have demonstrated enterovirus proteins and its genome in the pancreas of T1D patients. However, relatively few studies have been done so far and most of them have employed only formalin-fixed tissues making it difficult to confirm the presence of the virus using molecular methods such as PCR. These histological methods have also been applied in studies showing enteroviruses in myocarditis and cardiomyopathies as well as in some other diseases.

The most frequently used antibody in enterovirus detection in immunohistochemistry is the commercial monoclonal antibody clone 5-D8/1 (Dako). This antibody was developed in 1987 (173) and it recognizes a conservative group-specific epitope in enteroviral VP1 capsid protein (174). It works well in FFPE tissues and reacts with a wide range of enterovirus serotypes in cell culture (175). On the other hand, it has been shown that this antibody fails to detect several CAV and echovirus serotypes as well as EV68-71 serotypes (176-178). In addition, the clone 5-D8/1 has been reported to cross-react with host proteins which are present in the pancreatic tissue such as HSP60/65 and IA-2 (179, 180) and to bind to uninfected human cardiomyocytes (181) as well as vascular smooth muscle and centroacinar cells in the exocrine pancreas (182). Recently, the possible cross-reactivity of this antibody with pancreatic islet cell proteins has been studied extensively (183, 184). Altogether, this widely used antibody seems to react with several different enterovirus types.

However, it may also recognize some host proteins if used in non-optimal conditions, and therefore the staining conditions should be optimized for different tissues and the presence of the virus should be confirmed by other methods.

(31)

Besides the clone 5-D8/1, there are other commercially available antibodies designed for enterovirus detection. The Enterovirus Screening Set (Merck Millipore, Darmstadt, Germany) includes four species-specific antibody combinations Coxsackievirus B Blend, Echovirus Blend, Enterovirus Blend and Poliovirus Blend, as well as Pan-Enterovirus Blend which is a mixture of two monoclonal antibodies.

Pan-Enterovirus Blend reacts with all enterovirus serotypes but according to the manufacturer also cross-reacts with some non-enteroviral species. This antibody combination is reported to be very sensitive, since it detected all 41 tested enterovirus serotypes in cell culture, but also showed cross-reactivity with seven out of ten tested non-enteroviruses (178). In addition, some studies have used in-house antibodies which are not commercially available but produced in different laboratories and targeting different enterovirus proteins.

Non-isotopic ISH applications have today become more and more popular and replaced methods based on radioactive reagents. For example digoxigenin-labelled probes are commonly used in enterovirus detection. The probes are generally designed to hybridize with a conserved sequence which is common to all known enterovirus serotypes, but also species-specific probes have been used (185, 186).

ISH is a challenging method due to its many different stages, and it is always a kind of a compromise to get a strong enough positive signal without gaining distracting background staining.

The first isolation of enterovirus from the pancreas of T1D patient occurred in 1979 from a 10-year-old boy who died of diabetic ketoacidosis. The virus was typed to be CBV4 by neutralization with CBV4-specific antibody, and it caused diabetes after being inoculated into a mouse (156). A similar case involving CBV5 was reported in 1980 (187) but in that case the virus was isolated from the stools. However, even earlier a report showing enterovirus in human pancreatic tissue using histological staining methods was published in 1976 when CBV antigen was found by immunofluorescence in the pancreatic islets of a 5-year-old girl who had an acute T1D. In a neutralization test, a high antibody titer against CBV4 was found in serum of this child. (188). In 1985 CBV3 antigen was found in the pancreatic islets of a 10- day-old baby with generalized CBV3 infection using immunofluorescence assay (189). In one other study, enterovirus tests from the pancreas of an 8-year-old child after the onset of T1D proved negative, indicating possible inter-individual differences in the role of enteroviruses in T1D development (190).

(32)

Following these case reports, more extensive studies have been published showing enterovirus proteins and RNA in the pancreatic tissue of patients with T1D and patients who died of fatal enterovirus infections using immunohistochemical and in situ hybridization methods. Enterovirus capsid protein VP1 was found in the pancreatic islets in all of seven infants who died of coxsackieviral myocarditis and who had insulitis, and VP1 almost exclusively localized in insulin-positive cells. (191).

Likewise, VP1 was found in pancreatic acinar cells in three of five patients with myocarditis using an advanced technique with antigen retrieval and EnVision system (192). Enteroviral RNA was found predominantly in pancreatic islets in five out of nine myocardial patients using in situ hybridization (185), and later in seven out of twelve patients with a fatal enterovirus infection and in four out of 65 patients with T1D (186). More recently CBV4 was found in the pancreas of three of the six recent- onset T1D patients using immunohistochemistry, electron microscopy and virus isolation, and double-staining with insulin and glucagon revealed that the infection was specific for beta cells (157). Enterovirus was also found in the pancreatic islets in 61 % of recent-onset T1D patients and only in 6 % of pediatric controls using immunohistochemical staining with anti-VP1 antibody (182). In very recent studies enterovirus was found in the pancreas of patients with fulminant T1D, which is a subtype of T1D that results from an acute and almost complete destruction of islet cells at the onset of the disease (193, 194).

The studies showing enteroviruses in human pancreatic tissue are summarized in Table 2. Altogether, viral proteins has been detected by VP1 staining in the pancreas of a large proportion of T1D patients, and in some studies the presence of a virus has been confirmed by in situ hybridization or virus isolation. The virus signal has mainly come from beta cells. The fact that the virus is present in the islets also in systemic enterovirus infections demonstrates that enteroviruses have a natural tropism to endocrine cells in the pancreas.

(33)

Table 2. Summary of studies showing enterovirus in human pancreatic tissue.

Tissue Patients Positive for EV Method Reference

Pancreas T1D 1/1 IHC (188)

T1D 1/1 VI (156)

Fatal EV infection 1/1 IHC (189)

Myocarditis 7/21 IHC (191)

Myocarditis 5/9 ISH (185)

Fatal EV infection, T1D 11/77 ISH (186)

T1D 3/6 IHC, EM, VI (157)

T1D 44/72 IHC (182)

Fulminant T1D 3/3 IHC (193)

Fulminant T1D 1/3 ISH (195)

EM = electron microscopy, IHC = immunohistochemistry, ISH = in situ hybridization, VI = virus isolation In addition to pancreas, enteroviruses have also been detected in other tissues using immunohistochemical and in situ hybridization techniques. Several studies have reported enteroviruses in the heart of patients with myocarditis. Enterovirus infection has also been connected to chronic fatigue syndrome. Table 3 summarizes the studies showing enteroviruses in other tissues.

(34)

Table 3. Summary of studies showing enterovirus in other human tissues.

Tissue Patients Positive for EV Method Reference

Heart n.a. n.a. ISH (196)

Fatal EV infection 1/1 IHC (189)

Stomach Chronic fatigue syndrome 135/165 IHC (199) IHC = immunohistochemistry, ISH = in situ hybridization

4.3.3 Mechanism of enterovirus-induced beta cell destruction

The mechanisms by which enteroviruses may contribute or cause destruction of beta cells are not fully understood. Animal models on virus-induced diabetes suggest that viruses can infect and destroy beta cells directly or induce an immune-mediated process leading to beta cell damage (200, 201). Direct viral effects could be mediated by induction of apoptosis or necrosis in infected cells, as well as impaired function of beta cells (shut-off of cell functions is typical for enterovirus infection). One possibility is that enterovirus infection leads to an autoimmune process and destruction of beta cells by so-called bystander activation mechanism (202). This theory implies that local virus infection in the pancreatic islets maintains inflammation and creates proinflammatory milieu leading to activation of antigen presenting cells and autoreactive lymphocytes. (203, 204). Another mechanism by which enteroviruses could cause beta cell destruction is molecular mimicry, when autoimmune attack results from immunological cross-reactivity induced by similar structures in the virus and beta cell autoantigens (205). A well-known model is the

(35)

sequence similarity discovered in the enteroviral non-structural protein 2C and islet autoantigen GAD65 (a common amino acid sequence PEVKEK) (206, 207). Other possible molecular mimicry epitopes are shared between enterovirus VP1 protein and VP0 protein precursor and host IAR/IA-2 tyrosine phosphatase and heat chock protein 60/65 (179, 180).

Studies with isolated human pancreatic islets have demonstrated that the effect of virus replication in the islet cells is serotype and strain-dependent, since some of the viruses cause severe morphological changes and cell death as well as functional impairment, whereas some serotypes and isolates are less destructive (208-210).

These islet cell experiments have demonstrated that most of the studied enteroviruses infect predominantly beta cells, although some enteroviruses seem to infect both beta and alpha cells (14, 209, 211).

The genetic background of the host influences the development of beta cell damage and diabetes in response to a viral infection. Several genes which are associated with T1D are known to have a role in antiviral immune responses. The HLA genes which modulate the risk of T1D are associated with immune response against certain viruses including enteroviruses (212) and can also modulate the course of virus infections (213-215). Thus, it is possible that the presentation of critical viral peptide to T cells by diabetes–associated HLA molecules is an important step in the pathogenesis and may either lead to induction of immunological cross-reactivity to homologous sequences in beta cell proteins or regulate the quantity and quality of antiviral responses and immune protection against the virus. In addition, the polymorphism of the IFIH1 gene which has a role in innate immune system by recognizing double- stranded enteroviral RNA, has recently been associated with T1D (57, 216), and it has been shown that certain rare alleles of the gene protect from T1D (58). This suggests that antiviral innate immune responses may be important regulators of beta cell damage, possibly by their effect on virus-induced inflammation or on immune protection against the virus. In addition to IFIH1, another gene linked to T1D and having a role in viral defense is the OAS1 gene which encodes an enzyme that binds to and degrades endogenous and viral double-stranded RNA, and consequently restricts viral replication and promotes cell death. This enzyme may damage beta cells through RNaseL-mediated degradation of cellular RNA. (217). In addition, polymorphisms in the IRF7-driven inflammatory network (IDIN) genes are associated with increased risk of T1D (218). One recent study demonstrated that one

(36)

activation of innate immunity and its diabetes susceptibility genotype is associated with a low innate immune response against the virus (219). In addition, a recent study demonstrated that the capacity of the virus to induce innate immune responses was depended on virus strain and its interactions with plasmacytoid dendritic cells (220).

In summary, several different mechanisms could mediate virus-induced damage in beta cells. The mechanisms are not mutually exclusive and they may operate in different stages of the process. For example, part of them may be important in immune protection against the virus (e.g. innate immunity), part could determine if immunological cross-reactivity is induced (e.g. HLA type), and part could modulate viral interactions with beta cell (e.g. serotype of the virus determines receptor binding). Further studies are needed to find out which of these scenarios fits best with the pathogenesis of T1D.

(37)

5 Aims

The aims of this study were

1. To develop immunohistochemical and in situ hybridization methods to detect enterovirus proteins and RNA in formalin-fixed paraffin-embedded samples and frozen tissue samples (I) and

2. To use these methods to investigate possible presence of enteroviruses, viral antigen and RNA, in tissue samples collected from T1D and non-diabetic subjects (II-IV).

(38)

6 Materials and methods

6.1 Collection of tissue samples from study subjects

6.1.1 Tissue samples from a child positive for islet cell autoantibodies (II)

The material in this study comprised sera and pancreatic tissue from a boy taking part in the Diabetes Prediction and Prevention (DIPP) study in Finland. The DIPP study follows children with increased HLA-associated risk for T1D from birth, and blood samples are taken at 3 to 12 month intervals and analyzed for the presence of diabetes-associated autoantibodies (221). This child tested positive for islet cell antibodies (titres 4-8, Juvenile Diabetes Foundation units [JDFU]) in five sequential serum samples taken over a period of 13 months preceding his accidental death at the age of 36 months, but he was negative for other T1D-associated autoantibodies (IAA, GADA, IA-2A). Tissue samples were taken during a routine autopsy at Tampere University Hospital. One 1,5 cm³ sample was available from one part of the pancreas. The sample was fixed in formalin, embedded in paraffin and cut in 5 µm serial successive sections onto microscopic slides. The child had the HLA- DQB1*02/*0302 genotype which confers increased genetic risk for T1D. The study was approved by the National Authority for Medicolegal Affairs in Finland.

6.1.2 Small intestine biopsy samples (III and IV)

In Report III, small intestine biopsy samples from 12 T1D patients and 10 control subjects, and in Report IV, from 39 T1D patients, 40 celiac disease patients and 40 control subjects, were collected during the years 1995-2000 at the Department of Gastroenterology, Tampere University Hospital, Finland. T1D had been diagnosed in all T1D patients and all of them were on insulin treatment. A follow-up sample (taken 1 year after the initial sample) was available from three T1D patients. Celiac disease was diagnosed by a positive endomysial antibody result and by small-bowel mucosal villous atrophy and crypt hyperplasia in duodenal biopsy. Celiac disease was

(39)

diagnosed at this time-point and all patients were still on a normal gluten-containing diet at the time of the biopsy. All control subjects underwent gastroscopy due to unspecific gastrointestinal symptoms. Morphological analyses of small bowel mucosal biopsies indicated normal gut mucosa in all control subjects. The characteristics of all study subjects are described in Tables 4 and 5.

Table 4. Study subjects of Report III.

T1D* Controls

Subjects (N) 12 10

Males/females (N) 2/10 3/7

Median age in years (range) 30 (18-53) 54 (23-71)

Median age at diabetes diagnosis in years (range) 17 (5-51)

Median duration of diabetes in years (range) 13 (0-51)

**Exact information from only 10 T1D patients available. Others were diagnosed early in childhood.

(40)

Table 5. Study subjects of Report IV.

T1D* Celiac disease Controls

Subjects (N) 39 40 41

Males/females (N) 11/28 13/27 16/25

Median age in years (range) 43 (18-63) 41 (18-75) 47 (23-76)

HLA DQ2 and/or DQ8 (%) 57 100 34

Median age at diabetes diagnosis in years (range)** 18 (4-51)

Median duration of diabetes in years (range)** 20 (0-38)

*16 of the T1D patients also had celiac disease.

**Exact information from only 23 T1D patients available. Others were diagnosed early in childhood.

Study subjects of Report III are also included in the study material of Report IV.

For in situ hybridization and immunohistochemistry, biopsy samples were fixed in formalin and embedded in paraffin, after which they were cut in 5 µm sections onto microscopic slides. For RT-PCR, unfixed samples were stored frozen in OCT medium at -70ºC. FFPE samples were available from all study subjects while frozen samples were available from four subjects in Report III and 86 in Report IV.

The study protocol was approved by the ethical committee of the Tampere University Hospital. All subjects gave their written informed consent.

(41)

6.2 Methods

6.2.1 Enterovirus-infected cell lines (I)

Green monkey kidney cells (GMK and Vero), vervet monkey kidney cells (MA104), human fibroblast cells (HEL-7), carcinomic human cervix epithelial cells (HeLa) and carcinomic human alveolar basal epithelial (A549) cells were grown to a monolayer and infected with different enterovirus serotypes and control viruses as shown in Table 6. Infected cells were incubated at 37 C until a cytopathic effect was seen in about 30-60 % of the cells. Cells were then harvested and either frozen in liquid nitrogen or fixed in formalin and embedded in paraffin. 5 µm cryostat or paraffin sections were cut onto Superfrost Plus microscopic slides (Menzel-Glaser, Braunschweig. Germany). The cells and viruses were obtained from the American Type Culture Collection (ATCC), except for the serotype echovirus 3 which was isolated in our laboratory (222), and the GMK cells which were obtained from National Institute for Health and Welfare, Helsinki, Finland (a gift from Dr. Merja Roivainen).

Table 6. Infected cells used for the development of immunohistochemical stainings and in situ hybridization.

Cells Infected with

GMK cells CBV1, CBV2, CBV3, CBV4, CBV5, CBV6, CAV9, echo3, echo6, echo9, echo11, E71, PV3

Vero cells CAV16

A549 cells echo30, HPeV1

MA104 cells rotavirus

HeLa cells adenovirus

HEL-7 cells cytomegalovirus

(42)

6.2.2 In situ hybridization (I, II, III and IV)

In Reports I, II, III and IV, the presence of enteroviral genome in FFPE sections was analyzed using in situ hybridization. An enterovirus-specific oligonucleotide probe (sequence from 5’ to 3’ GAA ACA CGG ACA CCC AAA GTA GTC GGT TCC GCT GCR GAG TTR CCC RTT ACG ACA) was designed to hybridize with the conserved, group-common sequence in the 5’ non-coding region of the enteroviral genome to detect all known enterovirus types (according to sequences in Gene bank). The probe was 3’ end-labeled with digoxigenin using the DIG oligonucleotide tailing kit (Roche Diagnostics Ltd, Welwyn Garden City, U.K).

The in situ hybridization method was optimized using CBV3-infected and mock- infected cell culture and tissue samples from mice infected by CBV3. The hybridization was performed using manufacturer’s instructions (DIG System for In Situ Hybridization, Roche Diagnostics Ltd) except for the DIG labeling reaction which was carried out using 10 pmol of the probe. The amount of the probe in the hybridization cocktail was 250 ng (1 ng/µl), and the hybridization time was 3 h. The binding of the probe was revealed by anti-DIG antibody, which was conjugated with alkaline phosphatase. This enzyme together with its substrate NBT/BCIP yields an insoluble purple precipitate, which was detected using a light microscope.

6.2.3 Immunohistochemical stainings (I, II, III and IV)

6.2.3.1 Antibodies

In Reports I, II, III and IV, the following monoclonal antibodies were used for the immunohistochemical stainings: anti-enterovirus 5-D8/1 (for enterovirus VP1 protein), anti-HSP60 (heat shock protein 60), anti-Fas (apoptosis), anti-CD3 (T lymphocytes), anti-CD68 (macrophages), anti-Ki-67 (cell proliferation), anti-CD34 (endothelium of blood vessels), anti-CD56 (natural killer cells), anti-CD23 (dendritic cells), anti-COX-2 (cyclooxygenase-2), anti-Leu-4 (CD3+ intraepithelial lymphocytes (IELs)), anti- F1 ( + IELs), anti-T-cell receptor ( + IELs), anti-HLA-DR (HLA-DR expression), fluorescein anti-human IgA, and anti-transglutaminase 2. The following polyclonal antibodies were used: anti-CBV3, anti-CAV16, anti-echovirus 11, anti-CAR, anti-insulin, anti-glucagon, anti-somatostatin and anti-beta-2-

(43)

microblobulin (HLA class I antigen). In Report I, the following monoclonal antibody combinations were used: Pan-Enterovirus Blend, Coxsackievirus B Blend, Echovirus Blend and Poliovirus Blend. In Reports II and IV, the following fluorescein secondary antibodies were used for the immunofluorescent double-stainings: rabbit Alexa-Fluor® 488, swine TRITC and anti-IgG. All the antibodies, their clones, dilutions and manufacturers are shown in Table 7.

In-house antibodies against CBV3, CAV16 and echovirus 11 were produced in The Department of Virology, University of Turku, Turku, Finland, by immunizing rabbits by highly purified heat-treated CVB3, CVA16, and echovirus 11, respectively (collaboration with Professor Timo Hyypiä). The IgG fraction of rabbit hyperimmune sera was purified by fast protein liquid chromatography using a protein A column (Pharmacia Fine Chemicals, Uppsala, Sweden). In-house antibody against CAR was produced in the Department of Histopathology, St. James’s University Hospital, Leeds, UK, by immunizing rabbits with recombinant CAR protein (collaboration with Dr. Caroline Verbeke). The IgG fraction of the rabbit serum was purified on a Protein A affinity column (Amersham-Pharmacia, Little Chalfont, UK).

Detection of Enteroviruses in Tissue Samples - Methods and applications in type 1 diabetes

MAARIT OIKARINEN

Detection of Enteroviruses in Tissue Samples

Methods and applications in type 1 diabetes

MAARIT OIKARINEN

Detection of Enteroviruses in Tissue Samples

Methods and applications in type 1 diabetes

Abstract

Tiivistelmä

Contents

1 List of original publications

2 Abbreviations

3 Introduction

4 Review of the literature

4.1 Enteroviruses

4.1.1 Classification, structure and replication

4.1.2 Enterovirus diseases

4.2 Type 1 diabetes

4.2.1 Autoimmune process

4.2.2 Genetic susceptibility

4.2.3 Environmental factors

4.2.4 Diabetes and the gut

4.3 The link between enterovirus infections and type 1 diabetes

4.3.1 Epidemiological studies

4.3.2 Detection of enteroviruses in the pancreas of type 1 diabetic patients

4.3.3 Mechanism of enterovirus-induced beta cell destruction

5 Aims

6 Materials and methods

6.1 Collection of tissue samples from study subjects

6.1.1 Tissue samples from a child positive for islet cell autoantibodies (II)

6.1.2 Small intestine biopsy samples (III and IV)

6.2 Methods

6.2.1 Enterovirus-infected cell lines (I)

6.2.2 In situ hybridization (I, II, III and IV)

6.2.3 Immunohistochemical stainings (I, II, III and IV)