Controlled human infections A report from the controlled human infection models workshop, Leiden University Medical Centre 4-6 May 2016

(1)

Conference report

Controlled human infections

A report from the controlled human infection models workshop, Leiden University Medical Centre 4–6 May 2016

Meta Roestenberg

^a,^⇑

, Annie Mo

^b

, Peter G. Kremsner

^c

, Maria Yazdanbakhsh

^a

aLeiden University Medical Centre, The Netherlands

bDivision of Microbiology and Infectious Diseases, National Institute of Allergy and Infectious Diseases, National Institutes of Health, USA

cUniversitätsklinikum Tübingen, Germany and Centre de Recherches Médicales de Lambaréné, Gabon

a r t i c l e i n f o

Article history:

Received 27 July 2017

Received in revised form 30 October 2017 Accepted 31 October 2017

Available online 20 November 2017

a b s t r a c t

The principle of deliberately infecting humans with infectious agents in a controlled setting, so-called controlled human infections (CHI), is not novel. Many CHI models have a long history and were established decades ago such as the intentional exposure to yellow fever and dengue performed in the 1900’s (Reed, 1902) [2]. In these times bioethics and scientific reasoning were in their infancy.

Nowadays, clinical trials are highly regulated and CHI are executed worldwide. Controlled human malaria infections and influenza infections are the two most frequently practiced. Others are experiencing a revi- val or are being carefully developed. Because CHI models test the efficacy of promising vaccine or drug candidates early in clinical development, they offer the potential to decrease the number of failing phase 2 and 3 trials, reducing risks for patients and saving costs and efforts.

In addition, CHI models provide unprecedented opportunities to dissect the physiological, immunological and metabolic changes that occur upon infection. However, it is clear that controlled infections require careful deliberation of safety, ethics, quarantine, scientific output and the production of infectious material.

An independent international workshop was hosted by the Leiden University Medical Centre in The Netherlands, bringing together clinical investigators, basic scientists, regulators, funders and policy mak- ers from 22 different countries to discuss the opportunities and challenges in CHI. The aim of the workshop was to discuss CHI as a tool to advance science, drug and vaccine development, share the challenges of establishing a CHI model with specific focus on neglected tropical diseases and the possibilities to transfer models to endemic sites. Noticeably, among the 128 participants were clinical investigators from ten different countries in Sub-Saharan Africa. An important dimension of the meeting was to give the floor to young established clinicians and scientists to voice their perspective on the future of CHI models.

1. Ethical considerations

During the first keynote lecture of the workshop, Hans-Jörg Ehni (University of Tübingen) gave an overview of historical CHI studies which are highly controversial experiments in view of the current ethical framework[2]. They were performed in times when informed consent was not yet established and the societal approval of the experiment was dependent on its outcome[3]. For example, Edward Jenner himself experimented on the son of his gardener and it was only for the success of these experiments that Jenner is considered the founder of vaccination[4]. However, those work-

ing on experimental yellow fever infections, who experimented on themselves as well as on prisoners and immigrants by injecting blood or by exposure to mosquito bites faced a different reality as severe cases developed and lives were lost[1]. As such, villains and heroes in historical experimental medicine can only be distin- guished in retrospect. Nowadays, with much more stringent bioethical demand, serious adverse events are increasingly rare in experimental medicine. Nevertheless, in an era where risk control and management are top priority in health care, novel CHI models can face scepticism.

For most CHI models, as for every clinical trial, absolute safety is a difficult if not an impossible claim to make. Ehni presented an updated categorization of pathogens relevant to CHI models as proposed by Grady and Miller [5] (Table 1), which provided a framework for subsequent ethical review of the individual models.

https://doi.org/10.1016/j.vaccine.2017.10.092

⇑Corresponding author.

E-mail address:M.Roestenberg@lumc.nl(M. Roestenberg).

Contents lists available atScienceDirect

Vaccine

j o u r n a l h o m e p a g e : w w w . e l s e v i e r . c o m / l o c a t e / v a c c i n e

(2)

3. What have we learnt so far

An overview of how the benefits have balanced the risks was reviewed by Meta Roestenberg (Leiden University Medical Centre), who emphasised the importance of CHI for progressing scientific insight into the pathophysiology of disease. Controlled human infections offer the opportunity to collect specimen in very early stages on the infection, which would not have been possible in field situations and may provide critical insight into mechanisms of disease or efficacy of novel vaccines. She highlighted the transition of controlled infection models into the 21st century and argued that the availability of highly sensitive diagnostic tools may increase safety of CHI experiments because they allow rapid diagnosis and early termination of a trial. Moreover, the increased use of genetically attenuated microbes will open up new opportunities for vaccine development, basic research and for alternative approaches establishing safer CHI by using attenuated microbes as challenge material. In addition, in depth analysis with–omics approaches or novel imaging techniques may provide important mechanistic data. Importantly, CHI provide new windows of opportunities to study the role of co-infection, diet, microbiome and previous exposure in an experimental setting.

Benjamin Mordmüller (University of Tübingen) discussed recent developments in CHI for malaria (CHMI). He stressed that the small sample size required, the inclusion of adults able to provide informed consent for the first efficacy assessment (in contrast to vulnerable groups such as children under natural exposure), and the increasing acceptance of these trials as surrogates of protection make CHI a highly valuable tool for vaccine and drug development (Fig. 1). CHMI can be performed using infected mosquitoes, blood stage parasites, or a purified, vialed and cryopreserved live sporozoite inoculum (a product from Sanaria Inc. – PfSPZ Challenge) [6,7]. The vialed sporozoites allow for multicentre trials which may be needed for advanced stage clinical development and even in areas where an insectary is not available[8]. In addition, several Plasmodium falciparum strains are available[9]and also Plasmod- ium vivax CHMI has been established by feeding mosquitoes directly on patient blood to generate infectious mosquitoes as challenge material[10]. Presently, CHMI is adapted to study transmission blocking interventions and larger CHMI studies are being performed to provide additional efficacy data which will add evi- dence of efficacy to conventional phase 2 data[11,12]. For such purposes, technology is transferred to endemic areas where life- long malaria-exposed volunteers reside. In addition, the use of large scale CHMI trials to substitute conventional phase 3 data may be imperative for the clinical development of, for example, a travellers’ malaria vaccine as the incidence of malaria in travellers is extremely low. The FDA approval of Vaxchora, a cholera vaccine,

for travellers’ market based only on CHI data has created a prece- dent for licensure based on CHI data alone[13].

4. Established controlled human infections: beyond malaria

Despite the diversity of microbes, the continuous careful balance between risk and scientific gain is a central theme for all CHI models. Peter Kremsner (University of Tübingen, Germany) started the discussion by listing the requirements of an ideal CHI (Table 2). He noted, however, that the risk analysis for considering CHI models should be based on available data rather than opinions.

With the possible occurrence of Guillain-Barré syndrome in Zika virus disease as an example, he stressed the importance of provid- ing a quantitative estimate of the risks for novel CHI models, preferably based on solid epidemiological data. Taking a very small risk of complications may be acceptable in certain cases if the benefit of such CHI model for product development and public health is imperative. Such risks cannot always be avoided, but may be mitigated by the trial setting (editorial note: the discussion on Zika virus CHI is ongoing[14].

5.1. Pneumococcal carriage

One novel CHI model investigates colonization with Streptococ- cus pneumonia. Invasive pneumonia or meningitis is a major killer of children under 5 years of age[15]and although pneumococcal Table 1

Categorization of pathogens by Ehni (University of Tübingen) in group A, B and C pathogens. Group A are the pathogens which are most safe. However, many infections will carry a small risk of severe or even serious adverse events and targeted therapeutics may not be available, such as in most systemic viral illnesses (group B).

These pathogens cause disease with small but potentially irreversible or serious risk.

The last category C pathogens cannot be used as controlled infection models, because they will lead to untreatable disease, intolerable symptoms or cause irreversible and serious damage. Of course, categorization of models can change as novel interventions become available.

Category Description Example

A Legitimate pathogens RhinovirusPlasmodium falciparumVibrio cholerae

B Limited risk pathogens Systemic viral illness Helicobacter pylori C Unacceptable risk

pathogens

HIVtetanusrabies

Fig. 1. Graphic representation of the number of product candidates and cumulative costs per development phase, before and after introduction of a CHI model. Light blue bars showing number of candidates at every stage of clinical development, dark blue the projected number after introduction of a CHI model. Light red line shows cumulative costs for the clinical development of one successful candidate, dark red line shows the projected cumulative costs after introduction of a CHI model. Arrows indicate cost reduction and shift towards early clinical testing if a CHI model is introduced. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Table 2

Characteristics of an ideal CHI model.

Safe 100% efficacious treatment, no sequelae Controlled cGMP infectious material

Detectable Highly sensitive, reliable and prompt diagnostic tool Fast Reasonable time window between patency and severe disease

(3)

conjugate vaccines are highly protective, an increase in non- vaccine serotypes occurs following vaccination campaigns. There- fore, several new vaccines for these serotypes are being developed.

To increase cost effectiveness and reduce time, a controlled pneumococcal colonization model has been developed. Daniela Ferreira (Liverpool School of Tropical Medicine, UK) presented her data on the pneumococcal colonization model. She showed that nasal instillation of bacteria leads to a roughly 50% colonization in healthy volunteers, lasting no more than one week[16]. Serious adverse events never occurred in more than 600 inoculations performed. Using this model of colonization, her team has tested the efficacy of different pneumococcal vaccines (including the 13- valent conjugate vaccine) showing 25–87% protection [17]. By nasal micro-sampling and bronchoscopy the local immune responses could be evaluated, revealing nasal granulocyte recruitment and an increased number of alveolar memory T-cells producing TNF and IL-17 in colonized volunteers in addition to antibody production [18]. Furthermore, the model has been critical for understanding the transition of colonization to invasive infection.

For example, individuals infected with respiratory viral infections had three times increased odds of becoming colonized[19], and live attenuated influenza vaccination increased the density of pneumococcus carriage in the nose. This valuable model will be used in future studies to focus on susceptibility of populations vulnerable to pneumococcal colonization such as the elderly.

5.2. Dengue

Beth Kirkpatrick (also on behalf of Anna Durbin, University of Vermont College of Medicine, USA) presented the state of the art in dengue CHI. With four billion people at risk of dengue infection and the life threatening nature of severe dengue, there is a large global effort to develop a safe dengue vaccine[20]. This active dengue vaccine development pipeline can benefit from a CHI model that will select appropriate candidates for downstream development. To decrease morbidity, genetically attenuated dengue strains have been developed in the context of a live attenuated dengue vaccine development programme[21]. Insufficiently attenuated strains failed as vaccines, but provided an opportunity for use in CHI. Whether the attenuated phenotype of the challenge strain also limits the extrapolation of dengue CHI models to the field, remains to be investigated. Kirkpatrick showed 100% infection rates (viremia) by the attenuated DEN2D30 virus in a CHI trial, in which 80% of volunteers developed a rash, but no fever[22]. A dengue vaccine candidate, TV003, showed complete protection against viremia in the DEN2D30 CHI model. The difference between dengue ‘‘disease” and ‘‘infection” models was discussed, whereby the outcome in the former is dictated by symptoms whereas the latter depends on the presence of the virus (as is the case for DEN2D³⁰⁾[22]. In addition, dengue CHI needs to be developed for other serotypes to investigate cross-protection. The need for accurate correlates of protection and a tool for down-selection of vaccine candidates early are imperative because phase III testing of a partly protective dengue vaccine may pose a risk of more severe disease through antibody dependent enhancement of viral replication. The CHI model would be valuable to understanding the mechanisms of antibody dependent enhancement and grasp the immunological basis of protection against this devastating viral infection.

5.3. Typhoid fever

The lack of reliable diagnostic tests and pathognomonic symptoms leads to poor control of infections with Salmonella typhi or paratyphi in resource poor settings. Novel medicines are needed as antimicrobial resistance of Salmonella is expanding. Malick

Gibani (Oxford University) shared data on CHI trials in typhoid fever. Although typhoid CHI were performed historically in Mary- land, his team is now updating the model to comply with current guidelines. A dose of 1-5x10⁴CFU of oral Salmonella typhi Quailes strain following bicarbonate solution caused typhoid disease in

65% of participants defined as fever of 38 °C for >12 h or a positive blood culture[23]. Novel vaccines such as the live attenuated MO1ZH09 (with mutations in the SPII gene) and the licensed oral live attenuated S. typhi Ty21a strain (Vivotif) have been tested using the CHI model and show modest protection of 13% and 35%, respectively[24]. Future studies will include testing of the capsular polysaccharide Vi preparation conjugated to protein carri- ers, which have shown improved immunogenicity in animal models. Currently another model for Salmonella paratyphi A is being developed in Oxford[25]. Salmonella paratyphi has accounted for an increasing fraction of cases of enteric fever worldwide, leading to increased investments in vaccine development whereby several candidates are now in preclinical and clinical development. The first CHI studies with Salmonella enterica serovar Paratyphi A (NVGH308 strain) show a 60% infection rate with 1–5 10³CFU [26]. Re-infection studies with typhoid versus paratyphoid strains are scheduled as well as vaccine trials. Nevertheless, in order to develop a more effective vaccine, a better understanding of the immune mechanisms is needed and to this end the CHI models provide unique opportunities[27].

5.4. Norovirus

Moving to viral gastrointestinal infections, the norovirus CHI model was discussed by Robert Atmar (Baylor College of Medicine, USA). Noroviruses are emerging as the leading cause of foodborne disease outbreaks[28]. While norovirus gastroenteritis is often a self-limiting disease, it can be the cause of significant morbidity and mortality among those with immature or weakened immune systems, such as children and the elderly. The first report of controlled infection of norovirus dates from 1947[29], after which many studies followed proving infectivity and confirming oral transmission routes [30–33]. The major limitation of norovirus CHI is the fact that the virus is difficult to culture and needs to be prepared directly from filtered stool. To ensure safety, the donor has to be screened for transmissible diseases such as syphilis, hepatitis viruses and HIV. As a result, the stocks of challenge material are limited, and participants in norovirus CHI studies are housed as inpatients during their symptomatic (infectious) phase until they do not produce any liquid stool or vomit. As soon as symptoms have subsided and the highly infectious phase of the disease has passed, participants are discharged with careful instructions as they may still be shedding virus. This procedure has successfully safeguarded against secondary infections. Using the norovirus CHI model, the importance of histo-blood group antigens (HBGA) for norovirus binding and infection was shown[34]. HBGA is a complex carbohydrate expressed on epithelial cells, and the Atmar group was the first to show a correlation between blocking anti- bodies to this molecule and clinical outcome[34]. Future work will focus on proof-of-concept studies for vaccines based on virus-like particles, development of antivirals, and the establishment of norovirus CHI for other strains.

5.5. Influenza

CHI for influenza is the second most frequently used CHI model nowadays and helped in understanding influenza immunology besides being used extensively for clinical evaluation of vaccines and drugs. Rob Lambkin-Williams (hvivo London, UK) presented mapping of T-cell responses before and after CHI with influenza H3N2 and H1N1 strains. He showed that CD4 + T-cells responding

(4)

to influenza internal proteins are associated with lower virus shedding and less severe illness[35]. A T-cell-based influenza vaccine (viral vectored nucleoprotein + matrix protein 1) was tested in a single vaccination schedule, and for the first time, clinical efficacy with less symptoms and decreased disease severity was seen upon CHI with influenza A/Wisconsin/67/2005[36,37]. Positive results were also obtained for a trivalent DNA vaccine encoding HA from influenza strain H1, H3 and B against CHI with influenza A/H3 Panama/2007/99 [38]. A comparison of influenza H3N2 CHI in younger (<45 y/o) versus older (46–65 y/o) non-vaccinated volunteers revealed longer duration of symptoms and viral shedding in the older participants (unpublished data). Several respiratory syn- cytial virus (RSV) CHI studies have also been carried out to test antiviral compounds (NCT02135614, NCT02673476). Building on the knowledge and resources from influenza and RSV CHI, future effort will also focus on CHI development with GMP production of infectious influenza virus challenge material using specific pathogen-free eggs and in specific high risk patients.

6. Controlled human infection models in development

The pneumococcus, typhoid, norovirus and respiratory viral CHI were all built on a foundation of historical experience gained over decades. The CHI models are now clearly propelling forward the field of infectious disease research and product development.

However, there is an array of neglected tropical diseases where important enabling steps to facilitate new developments are urgently needed but only can be achieved by substantial invest- ment in tools such as CHI. For example, despite the current worldwide distribution of drugs to control parasitic diseases such as hookworms and schistosomiasis, the burden of disease is high with 1758 thousand global disability-adjusted life years (DALYs) for hookworm and 2613 thousand DALYs for schistosomiasis in 2015 [39]. In addition, the limited number of available anthelminthic drugs is alarming as deworming programmes are the cornerstone of current control strategies and the possibility of developing resistance is a serious concern[40]. Additional interventions such as effective vaccines for parasite control are urgently needed[41].

6.1. Hookworm

With more than 400 million cases per year, hookworms are responsible for a vast amount of morbidity [39]. A hookworm CHI is currently being established by David Diemert (George Wash- ington University, USA) to rapidly assess efficacy of vaccine candidates for Necator americanus and for future testing of new drugs. L3 larvae from Nector americanus are the stages from the life cycle which are infectious to humans, as they are able to penetrate the host skin. Although several controlled infection studies with a limited number of L3 larvae have been undertaken before, consistent egg production and detection in stool has not been achieved[42–

47]. Currently, CHI infectious material consists of L3 larvae obtained from cultures of a chronic donor where HIV, HBV and HCV status is closely monitored and which are applied to the skin.

Preliminary data from a pivotal dose escalation study showed that an increase in dose from 25 to 50 L3 larvae paralleled increased reactogenicity, particularly a transient papulovesicular rash at the site of entry (NCT01940757). Fecal egg counts were highly vari- able, but they were positive in 9/10 subjects in the 50 larvae dose group. With the plan to further increase the infective dose to 75 larvae, achieving 100% infection of volunteers could be possible, yet the degree of adverse reactions in the acute phase will need to be carefully assessed. Interestingly, all adverse events were transient and hookworm infections beyond eight weeks were asymp- tomatic. Infections kept for a period of one year do not cause

anemia[42]and can be terminated by a curative regimen of alben- dazole. Most likely, the host response switching from Th1 to Th2 and more regulatory responses decreases reactogenicity of the gastrointestinal infection. With this model it is possible to study site- specific helminth associated immune responses that have remained elusive in humans, such as the switch between Th1, Th2 and regulatory responses.

6.2. Schistosomiasis

Meta Roestenberg (Leiden University Medical Centre, The Netherlands) presented efforts to develop a CHI for schistosomiasis. Cercaria, shed into the water from the intermediate watersnail host, are the infectious forms of the schistosome parasite. Roesten- berg discussed the challenges faced when producing infectious Schistosoma mansoni cercariae in Biomphalaria glabrata snails.

Because the schistosome eggs are responsible for the irreversible pathology in schistosome infections, the model will focus on single sex (male) infections. To differentiate between male and female cercariae, a multiplex real-time PCR was established. The availability of a highly sensitive circulating cathodic antigen test allows for accurate quantification of worm loads despite the lack of eggs[48].

The group in Leiden has recently started the first clinical trial, in which a four-step dose escalation from 10 cercaria to 100 cercaria will be performed to assess infectivity in groups of 10 volunteers in total (NCT02755324). The first results are expected in the last quarter of 2017. The drawback of the model would be the lack of any eggs, which would preclude testing of any anti-fecundity vaccines or drugs. Nevertheless, the model may prove to be useful for testing of vaccines or drugs targeted to stages spanning from larvae to adults. Although both controlled hookworm and schistosome infection models need further optimization and improvement, in keeping with the developments in the field of malaria, testing attenuated parasites through irradiation or genetic manipulation as effective vaccines may one day be feasible.

Currently another parasite CHI is also being explored and was presented by Beth Kirkpatrick. The Cryptosporidium CHI was pre- viously established primarily for basic studies of transmission and infectious doses in the 1990s, but was subsequently aban- doned [49]. A recent epidemiological survey revealed that cryptosporidium is one of top three causative agents of moderate- severe diarrhoea in children under two years of age. This infection has been associated with enteropathy and malnutrition, yet no vaccines and only one therapeutic drug are available [50]. Re- establishing the CHI model would provide a platform for rapid testing of candidate drugs and vaccines, as well as for investigating the importance of the host microbiome, pathophysiology, and possible public health impact of long-term shedding of the parasite. As of now, the infectious material will have to be produced by infection of gnotobiotic neonatal piglets and subsequent purification of challenge material from these animals. Knowledge learned from hookworm and norovirus infectious materials for CHI will certainly help for cGMP production of Cryptosporidium hominis. Many challenges remain and have to be addressed, such as the implementation of newly developed, highly sensitive PCRs to assess Cryptosporidium infection, careful definitions of pre-existing (partial) immunity to standardize outcomes, and concerns to prevent transmission.

7. Challenges and perspectives

Despite the heterogeneity of CHI models, many investigators have experienced similar challenges and hurdles in developing the models. In the final session, experiences and knowledge were shared among participants.

(5)

For some models, the production of infectious material for controlled infections and the regulatory perspectives have been chal- lenging. The establishment of master cells banks, rigorous quality assurance and quality control with high levels of standardization may be common for production of cGMP material. In addition, the recruitment, selection and screening of volunteers are considered critical steps of CHI trials. The use of visual aids and some- times knowledge tests would greatly enhance the accuracy and efficiency of a fully informed consent process. Furthermore, CHI trials require highly specialised medical care with 24/7 availability and quarantine facilities for some models. Quarantine criteria were considered a major concern, as it may be difficult to determine infectiousness based on clinical or microbiological criteria. Sharing of quarantine protocols and experiences among different CHI com- munities was highly encouraged. Other considerations, e.g., development and refinement of proper fit-for-purpose, preferably quantitative endpoints such as symptom scores, were deemed a priority area for many CHI models. Finally, ‘‘second stage” CHI models, i.e., the transfer of models to endemic regions or the inclusion of vulnerable populations were discussed. Because of the pos- sibly increased risk of adverse events in vulnerable populations, the transition of CHI models to such patients requires ethical debate on the balance between risk and benefit, and additional safety measures may be put in place. Nevertheless, the comparison of CHI in healthy volunteers as compared to vulnerable populations may be very useful as an intermediate step between CHI trials and field efficacy. Similarly, the transition of models to endemic sites may be needed to reassess efficacy in populations with different genetic background and pre-exposure. Efforts are currently ongoing to transfer models such as the pneumococcal carriage

model and the schistosome CHI model to African endemic sites.

Particularly for neglected tropical diseases transfer of models to endemic areas is essential to assess the potential efficacy of novel interventions in the target population in a very early stage of clinical development. However, transfer of CHI model does require capacity building not only of clinical sites but also of regulatory agencies and ethics committees.

8. Recommendations

The Controlled Human Infection Models Workshop at the Lei- den University Medical Centre provided an opportunity to discuss and exchange experiences in CHI trials. The workshop was instrumental in bringing together clinical scientists from different conti- nents working on viruses, bacteria and parasites in very different, highly demanding clinical environments. These scientists are responsible for safeguarding the continuous balance between health risk and potential future benefit through the development of new interventions and the increase in knowledge. Investigators expressed the need to more regularly engage in such open discussions and to establish a controlled human infections platform in which ‘‘best practices” can be shared. Discussions to develop such a platform, which will start as an online service, were initiated.

Such a platform could offer an opportunity to debate ethical aspects, share experiences and protocols (Fig. 2) and facilitate more efficient and faster sharing of safety data.

In addition, there was strong support for the use of CHI models for diseases with high morbidity or mortality. Solid epidemiological data may provide a basis for a risk assessment for new models.

Fig. 2. Challenges and Perspectives in CHI models. Because many challenges (problems) in CHI are shared between models, exchanging experiences, including ethical debate and establishing common guidelines can be instrumental. An online platform could facilitate the exchange and aid in disseminating information or organising follow-up meetings.

(6)

The provision of a framework for bioethical evaluation, as was sug- gested by categorization of pathogens in category A, B or C classes, can be useful in balancing risks and public health benefit. It should, however, be noted that such categorization is by no means defini- tive, e.g. dengue CHI can transition from a group B to A category by the development of attenuated strains and Schistosoma mansoni CHI was initially considered at C, but has developed to category A by the use of a monosexual infection.

To further dissect the mechanisms behind disease or infection, many investigators are now taking-omics approaches to charac- terise immunological effects and identify markers for infection or protection. In addition, efforts are increasing to obtain local tissue samples, e.g. respiratory mucosa, skin or rectal mucosa to provide a more detailed overview of immune responses at different anatom- ical sites. Progressing to big data analysis using detailed time trends was considered the most promising approach to dissect the mechanism of disease and immune response. Because this is as yet rather unexplored territory, sharing of data analysis pipeli- nes may prove to be useful in disseminating these important technologies in different CHI models.

None of the infection models discussed at the meeting (pneumococcus, influenza, norovirus, typhoid fever, cryptosporidium, hookworm, schistosomiasis, dengue and malaria) had been vali- dated for their capacity to correctly predict successful outcome of vaccine or drug candidates in an endemic or field setting. Never- theless, models such as malaria, dengue and influenza are already applied for downstream product development, whereby a failing proof-of-concept CHI trial aborted the development of several candidates. CHI models may be a tool to compensate for the increasingly complex and expensive path towards licensure of novel drugs or vaccines. In addition, the spread of antimicrobial resistance worldwide provides an urgent need to accelerate the development of novel interventions. Continuing the discussion on CHI safety, design and harmonisation is paramount to provide a framework for the ethical and scientific acceptance of the CHI trials and secure the ongoing use of this highly valuable tool.

Acknowledgements

The Controlled Human Infection Models Workshop would not have been possible without the invaluable support from the National Institute of Allergy and Infectious Diseases, part of the National Institutes of Health, USAID, European Vaccine Initiative, European & Developing Countries Clinical Trials Partnership and the Leiden University Medical Centre.

MR was supported by a VENI grant from ZONMW and a Gisela Thier fellowship from the LUMC. The authors have no conflicting financial interests.

References

[1]Reed W. Recent researches concerning the Etiology, Propagation, and Prevention of Yellow Fever, by the United States Army Commission. J Hyg (Lond) 1902;2:101–19.

[2]Ashburn PM, Craig CF. Experimental investigations regarding the etiology of dengue fever. 1907. J Infect Dis 2004;189:1747–83.

[3]Krugman S. The Willowbrook hepatitis studies revisited: ethical aspects. Rev Infect Dis 1986;8:157–62.

[4]Pead PJ. Benjamin Jesty: new light in the dawn of vaccination. Lancet 2003;362:2104–9.

[5]Miller FG, Grady C. The ethical challenge of infection-inducing challenge experiments. Clin Infect Dis 2001;33:1028–33.

[6]Epstein JE, Richie TL. The whole parasite, pre-erythrocytic stage approach to malaria vaccine development: a review. Curr Opin Infect Dis 2013;26:420–8.

[7]Sauerwein RW, Roestenberg M, Moorthy VS. Experimental human challenge infections can accelerate clinical malaria vaccine development. Nat Rev Immunol 2011;11:57–64.

[8]Richie TL, Billingsley PF, Sim BK, James ER, Chakravarty S, Epstein JE, et al.

Progress with plasmodium falciparum sporozoite (PfSPZ)-based malaria vaccines. Vaccine 2015;33:7452–61.

[9] Teirlinck AC, Roestenberg M, van d V, Scholzen A, Heinrichs MJ, Siebelink- Stoter R, et al. NF135.C10: a new Plasmodium falciparum clone for controlled human malaria infections. J Infect Dis 2013;207: 656–60.

[10] Arevalo-Herrera M, Forero-Pena DA, Rubiano K, Gomez-Hincapie J, Martinez NL, Lopez-Perez M, et al., Plasmodium vivax Sporozoite Challenge in Malaria-Naive and Semi-Immune Colombian Volunteers. PLoS One 2014;9:

e99754.

[11]Hodgson SH, Juma E, Salim A, Magiri C, Kimani D, Njenga D, et al. Evaluating controlled human malaria infection in Kenyan adults with varying degrees of prior exposure to Plasmodium falciparum using sporozoites administered by intramuscular injection. Front Microbiol 2014;5:686.

[12]Shekalaghe S, Rutaihwa M, Billingsley PF, Chemba M, Daubenberger CA, James E, et al. Controlled Human Malaria Infection of Tanzanians by Intradermal Injection of Aseptic, Purified, Cryopreserved Plasmodium falciparum Sporozoites. Am J Trop Med Hyg 2014.

[13] U.S.Food and Drug Administration. FDA approval of VAXCHORA, <http://www.

fda.gov/BiologicsBloodVaccines/Vaccines/ApprovedProducts/ucm508903.

htm>; 27-12-2016 [accessed 6-7-2017].

[14] News on Zikavirus challenge model, http://www.npr.org/sections/health- shots/2016/02/25/468120804/why-scientists-hope-to-inject-some-people- with-zika-virus, https://www.statnews.com/2017/02/28/zika-vaccine-ethics- panel/; 25-2-2017 [accessed 30-10-2017].

[15] WHO. WHO pneumonia fact sheet, <http://www.who.

int/mediacentre/factsheets/fs331/en/>; 1-9-2016 [accessed 6-7-2017].

[16]Ferreira DM, Jambo KC, Gordon SB. Experimental human pneumococcal carriage models for vaccine research. Trends Microbiol 2011;19:464–70.

[17]Collins AM, Wright AD, Mitsi E, Gritzfeld JF, Hancock CA, Pennington SH, et al. First human challenge testing of a pneumococcal vaccine. Double- blind randomized controlled trial. Am J Respir Crit Care Med 2015;192:853–8.

[18] Wright AK, Bangert M, Gritzfeld JF, Ferreira DM, Jambo KC, Wright AD, et al., Experimental human pneumococcal carriage augments IL-17A-dependent T- cell defence of the lung. PLoS Pathog 2013;9: e1003274.

[19]Glennie S, Gritzfeld JF, Pennington SH, Garner-Jones M, Coombes N, Hopkins MJ, et al. Modulation of nasopharyngeal innate defenses by viral coinfection predisposes individuals to experimental pneumococcal carriage. Mucosal Immunol 2016;9:56–67.

[20] WHO. Dengue fact sheet, <http://www.who.int/mediacentre/factsheets/fs117/

en/>; 6-7-2017 [accessed 6-7-2017].

[21] Dengue vaccine initiative. Live attenuated dengue vaccines, <http://

www.denguevaccines.org/live-attenuated-vaccines>; 6-7-2016 [accessed 6- 7-2017].

[22]Larsen CP, Whitehead SS, Durbin AP. Dengue human infection models to advance dengue vaccine development. Vaccine 2015;33:7075–82.

[23]Waddington CS, Darton TC, Jones C, Haworth K, Peters A, John T, et al. An outpatient, ambulant-design, controlled human infection model using escalating doses of Salmonella Typhi challenge delivered in sodium bicarbonate solution. Clin Infect Dis 2014;58:1230–40.

[24] Darton TC, Jones C, Blohmke CJ, Waddington CS, Zhou L, Peters A, et al., Using a human challenge model of infection to measure vaccine efficacy: a randomised, controlled trial comparing the typhoid vaccines M01ZH09 with Placebo and Ty21a. PLoS Negl Trop Dis 2016; 10: e0004926.

[25]McCullagh D, Dobinson HC, Darton T, Campbell D, Jones C, Snape M, et al.

Understanding paratyphoid infection: study protocol for the development of a human model of Salmonella enterica serovar Paratyphi A challenge in healthy adult volunteers. BMJ Open 2015;5:e007481.

[26]Dobinson HC, Gibani MM, Jones C, Thomaides-Brears HB, Voysey M, Darton TC, et al. Evaluation of the clinical and microbiological response to salmonella Paratyphi A infection in the first paratyphoid human challenge model. Clin Infect Dis 2017;64:1066–73.

[27]Fresnay S, McArthur MA, Magder L, Darton TC, Jones C, Waddington CS, et al.

Salmonella Typhi-specific multifunctional CD8+ T cells play a dominant role in protection from typhoid fever in humans. J Transl Med 2016;14:62.

[28] WHO. WHO estimates of the global burden of foodborne diseases, <http://

www.who.int/foodsafety/publications/foodborne_disease/fergreport/en/>; 3- 12-2015 [accessed 6-7-2017].

[29]Gordon I, Ingraham HS, Korns RF. Transmission of epidemic gastroenteritis to human volunteers by oral administration of fecal filtrates. J Exp Med 1947;86:409–22.

[30] Parrino TA, Schreiber DS, Trier JS, Kapikian AZ, Blacklow NR. Clinical immunity in acute gastroenteritis caused by Norwalk agent. N Engl J Med 1977;297:86–9.

[31]Wyatt RG, Dolin R, Blacklow NR, DuPont HL, Buscho RF, Thornhill TS, et al.

Comparison of three agents of acute infectious nonbacterial gastroenteritis by cross-challenge in volunteers. J Infect Dis 1974;129:709–14.

[32]Dolin R, Blacklow NR, DuPont H, Formal S, Buscho RF, Kasel JA, et al.

Transmission of acute infectious nonbacterial gastroenteritis to volunteers by oral administration of stool filtrates. J Infect Dis 1971;123:307–12.

[33]Steinhoff MC, Douglas Jr RG, Greenberg HB, Callahan DR. Bismuth subsalicylate therapy of viral gastroenteritis. Gastroenterology 1980;78:1495–9.

[34]Atmar RL, Bernstein DI, Lyon GM, Treanor JJ, Al-Ibrahim MS, Graham DY, et al.

Serological correlates of protection against a GII.4 Norovirus. Clin Vaccine Immunol 2015;22:923–9.

[35]Wilkinson TM, Li CK, Chui CS, Huang AK, Perkins M, Liebner JC, et al.

Preexisting influenza-specific CD4+ T cells correlate with disease protection against influenza challenge in humans. Nat Med 2012;18:274–80.

(7)

[36]Lillie PJ, Berthoud TK, Powell TJ, Lambe T, Mullarkey C, Spencer AJ, et al.

Preliminary assessment of the efficacy of a T-cell-based influenza vaccine, MVA-NP+M1, in humans. Clin Infect Dis 2012;55:19–25.

[37] Powell TJ, Peng Y, Berthoud TK, Blais ME, Lillie PJ, Hill AV, et al., Examination of influenza specific T cell responses after influenza virus challenge in individuals vaccinated with MVA-NP+M1 vaccine. PLoS One 2013;8: e62778.

[38]Jones S, Evans K, McElwaine-Johnn H, Sharpe M, Oxford J, Lambkin-Williams R, et al. DNA vaccination protects against an influenza challenge in a double- blind randomised placebo-controlled phase 1b clinical trial. Vaccine 2009;27:2506–12.

[39]GBD. DALYs and HALE Collaborators, Global, regional, and national disability- adjusted life-years (DALYs) for 315 diseases and injuries and healthy life expectancy (HALE), 1990–2015: a systematic analysis for the Global Burden of Disease Study 2015. Lancet 2015;2016(388):1603–58.

[40]Vercruysse J, Levecke B, Prichard R. Human soil-transmitted helminths:

implications of mass drug administration. Curr Opin Infect Dis 2012;25:703–8.

[41] Hotez PJ, Pecoul B, Rijal S, Boehme C, Aksoy S, Malecela M, et al., Eliminating the Neglected Tropical Diseases: Translational Science and New Technologies.

PLoS Negl Trop Dis 2016;10: e0003895.

[42]Croese J, Giacomin P, Navarro S, Clouston A, McCann L, Dougall A, et al.

Experimental hookworm infection and gluten microchallenge promote tolerance in celiac disease. J Allergy Clin Immunol 2015;135:508–16.

[43] McSorley HJ, Gaze S, Daveson J, Jones D, Anderson RP, Clouston A, et al., Suppression of inflammatory immune responses in celiac disease by experimental hookworm infection. PLoS One 2011;6: e24092.

[44]Feary J, Venn A, Brown A, Hooi D, Falcone FH, Mortimer K, et al. Safety of hookworm infection in individuals with measurable airway responsiveness: a randomized placebo-controlled feasibility study. Clin Exp Allergy 2009;39:1060–8.

[45]Feary JR, Venn AJ, Mortimer K, Brown AP, Hooi D, Falcone FH, et al.

Experimental hookworm infection: a randomized placebo-controlled trial in asthma. Clin Exp Allergy 2010;40:299–306.

[46] Daveson AJ, Jones DM, Gaze S, McSorley H, Clouston A, Pascoe A, et al., Effect of hookworm infection on wheat challenge in celiac disease–a randomised double-blinded placebo controlled trial. PLoS One 2011;6: e17366.

[47]Giacomin P, Zakrzewski M, Croese J, Su X, Sotillo J, McCann L, et al.

Experimental hookworm infection and escalating gluten challenges are associated with increased microbial richness in celiac subjects. Sci Rep 2015;5:13797.

[48]Corstjens PL, Nyakundi RK, de Dood CJ, Kariuki TM, Ochola EA, Karanja DM, et al. Improved sensitivity of the urine CAA lateral-flow assay for diagnosing active Schistosoma infections by using larger sample volumes. Parasit Vectors 2015;8:241.

[49]DuPont HL, Chappell CL, Sterling CR, Okhuysen PC, Rose JB, Jakubowski W. The infectivity of Cryptosporidium parvum in healthy volunteers. N Engl J Med 1995;332:855–9.

[50]Kotloff KL, Nataro JP, Blackwelder WC, Nasrin D, Farag TH, Panchalingam S, et al. Burden and aetiology of diarrhoeal disease in infants and young children in developing countries (the Global Enteric Multicenter Study, GEMS): a prospective, case-control study. Lancet 2013;382:209–22.