Towards a paradigm shift in innate immunity—seminal work by Hans G. Boman and co-workers

Abstract

Four decades ago, immunological research was dominated by the field of lymphoid biology. It was commonly accepted that multicellular eukaryotes defend themselves through phagocytosis. The lack of lymphoid cells in insects and other simpler animals, however, led to the common notion that they might simply lack the capacity defend themselves with humoral factors. This view was challenged by microbiologist Hans G. Boman and co-workers in a series of publications that led to the advent of antimicrobial peptides as a universal arm of the immune system. Besides ingenious research, Boman ignited his work by posing the right questions. He started off by asking himself a simple question: ‘Antibodies take weeks to produce while many microbes divide hourly; so how come we stay healthy?’. This led to two key findings in the field: the discovery of an inducible and highly potent antimicrobial immune response in Drosophila in 1972, followed by the characterization of cecropin in 1981. Despite broadly being considered an insect-specific response at first, the work of Boman and co-workers eventually created a bandwagon effect that unravelled various aspects of innate immunity.

This article is part of the themed issue ‘Evolutionary ecology of arthropod antimicrobial peptides’.

1. Do fruit flies ever get sick?

In 1970, Hans G. Boman, initiated a collaboration with Bertil Rasmuson at Umeå University in the North of Sweden. Boman had been appointed Professor of Microbiology at the Medical Faculty in 1966 and Rasmuson became a Professor of Genetics at the establishment of the University the same year.

‘Do fruit flies ever get sick?’ Boman challenged his colleague.

‘Very rarely,’ Rasmuson replied.

‘Then they must have an efficient immune response!’

So they decided to investigate the nature of the immune response in the already established genetic model Drosophila melanogaster.

At the time, immunologists focused to a great degree on the adaptive arm of the immune system with a general consensus existing that the highly specific antibodies and immunological memory were keys to fight off infections. The use of insects to study general immunological processes was not regarded highly owing to their lack of an adaptive immune response, at least in its classical sense. Infection studies in invertebrates had to this point mainly been carried out in larvae from the caterpillar Bombyx mori and bees, with the main goal to combat pests that were having a great impact on the silk and honey industry [1]. The cellular immune responses of phagocytosis and encapsulation had since Metchnikow's discovery been accredited as a general feature of insect immunity, but were still on the descriptive level. Less was known about the humoral immune system. Fleming had early on demonstrated the activity of lysozyme, derived from human mucous and tears, against some bacteria. His subsequent finding of penicillin, which provides a much broader spectrum of activity, led to the era of antibiotics and perhaps a decreased interest in the endogenous humoral immunity in higher animals. The first steps were taken in the 1960s when lysozyme was detected in insects, and the first indications of inducible antimicrobial factors in the haemolymph were reported [2–4]. However, the nature of the response had yet to be convincingly demonstrated, and besides lysozyme, the factors remained to be identified.

Boman's interest in immune defence had arisen from personal communications with clinical colleagues. In response to his question ‘which are the greatest medical challenges of our time?’, it was often pointed out that tropical diseases such as malaria were among the greatest killers, of which efficient control measures were lacking. Because diseases as such are typically spread by insects and highly specific in terms of host and vector preference he felt that this was of fundamental immunological importance. However, he also realized the need to start out with a more simplistic system to grasp how insects protect themselves against microbes.

Around the same time, one of us (Faye, at the time Nilsson) was looking for a project as a last part of her undergraduate studies, and was delighted when Rasmuson asked her to initiate this new-born project. The first task was simply to monitor host survival and bacterial growth following injections of varying doses of log phase bacteria into adult flies. Three bacterial species were chosen, Pseudomonas aeruginosa, Escherichia coli and Aerobacter cloacae (later renamed Enterobacter cloacae). Growth in vivo was observed for all three but only P. aeruginosa managed to kill the flies (within 2 days), whereas the other two reached a steady state at 10⁴–10⁵ CFUs per fly after two days. This interesting finding invoked the idea of an ‘immunization’ trial using the non-pathogenic strains. Upon a secondary infection with E. cloacae (carrying another antibiotic marker to discriminate it from the primary infection), no colonies were initially observed and the assay had to be repeated at earlier time points to trace any viable bacteria (figure 1a) [5]. The ‘vaccination’ (nowadays referred to as priming) was also highly efficient against secondary infections with E. coli and P. aeruginosa. Hence, a potent immune response had been triggered by the primary challenge, the nature of which appeared non-specific, but discriminatory between self–non-self. The response was not dependent on cellular activity, because adding the secondary strain to the supernatant of the homogenate from the primary infected flies caused a bacterial drop by one order of magnitude within minutes (figure 1b). It was concluded that fruit flies possess (i) a potent humoral immune response that is (ii) inducible and causes immediate killing of bacteria in a second challenge; (iii) is non-specific/broad spectrum; and (iv) protects from low doses of the fly pathogen P. aeruginosa.

Figure 1.

(a) Bacterial counts in Drosophila homogenates at different time points following primary and secondary challenge with E. cloacae. (a) In vivo, primary challenge of E. cloacae PenG^R β1 (open circle, dashed line), secondary challenge with saline control (open circle, full line). Primary challenge with E. cloacae Nal^R β11 (black circles, dashed line) and secondary challenge E. cloacae Nal^R β11 (black circles, full line). (b) In vitro, E. cloacae Nal^R β11 added to homogenate from primary-challenged (black circle, full line) and primary-injected saline control flies (black circle, dashed line). Modified from Boman et al. [5] and Bisset [1].

2. The one-test-tube animal: the discovery of antimicrobial peptides

Despite the choice of Drosophila as a genetic model, it would at the time probably be a great challenge to design a screen for flies with immune gene mutations. Boman, with his background in biochemistry, suggested instead to focus on the purification of the humoral factors. Owing to its small size, Drosophila was not particularly suitable, and the need for a larger insect possessing a corresponding antibacterial response became apparent.

During his sabbatical in 1973 at the Massachusetts Institute of Technology (MIT) in Cambridge, Massachusetts, USA, Boman visited Carroll Williams, a highly recognized insect physiologist at Harvard University, who studied hormones involved in metamorphosis of Saturniid moths. These beautiful insects caught Boman's interest. Back in Umeå, he and his wife Anita introduced Samia cynthia and Hyalophora cecropia as ‘one-test-tube animals’ quoting Williams (figure 2a). Using the pupae of these two Saturniids, it was demonstrated that lepidopterans also possess an inducible immune response [7,8]. The main benefit of the pupal form was the ease by which it was injected, bled and stored in the fridge during the six month hibernation period (diapause) [9]. When brought into room temperature for an experiment, the metabolism remained low for up to a week while the response to bacterial infection was strongly and rapidly induced. In addition, a single H. cecropia pupa could yield up to 2 ml haemolymph, which significantly eased the biochemical extractions. Both large and fairly pure yields of immune factors could hence be obtained. With the use of radioactive amino acid labelling during the course of the challenge, peptides/proteins induced and excreted in the haemolymph were obtained through fractionation followed by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE), and were designated P1–P8 according to descending molecular weight [7,8]. In fact, the smallest but most potent compound isolated, P9, was later found initially to be migrating out of the gels. Two of the major immune proteins from H. cecropia, P4 and P5, were first purified merely on the basis of radioactive amino acid incorporation (see below) [10,11]. Antisera against these proteins were prepared at the time, but the antimicrobial properties were not clarified.

Figure 2.

(a) Picture of Hyalophora cecropia taken in the laboratory by IF. (b) Left panel: acidic gel electrophoresis of P7, P9a and P9b; right panel, distinctive clearance zones in the bacterial lawn grown on top of the gel [6]. (c) The secondary structure of cecropin (From HG Boman's memoirs. Enebyberg Anahit, 2002).

After 10 years in Umeå, Boman was appointed Professor of Microbiology at Stockholm University in 1976. He attracted new collaborators and group members, including Håkan Steiner, Dan Hultmark (at the time, a postdoctoral fellow and PhD student, respectively) and the group of Hans Bennich, biochemist and Professor of Immunology at Uppsala University. Together, they engaged in the purification of P1–P9 on the basis of their anti-microbial activities. For this purpose, Micrococcus luteus and lipopolysaccharide (LPS) mutants of E. coli K12 from Boman's earlier work were applied [11,12]. The LPS mutant strains, originally selected for phage resistance and found to have lost different components of the LPS–sugar core, appeared to have different sensitivity to the antimicrobial peptides (AMPs). Despite the advantages of H. cecropia, it was a tedious job, carried out in cold-rooms with large columns, ‘sticky proteins’ and low retrieval.

The first AMPs, P9 A-B and P7 (a lysozyme), were purified and published in 1980 [6]. In his work, Hultmark developed a simple yet elegant acidic gel assay, where clearance zones of a bacterial lawn grown on top of the gel reflected the activity of the antimicrobial compounds (figure 2b). The amino acid sequence was subsequently deciphered, and the P9 peptides were designated cecropin A and B [13]. The structure and putative mode of action was resolved to a certain extent by circular dichroism spectroscopy (CD spectra) and Nuclear magnetic resonance (NMR) [14,15]. The cecropins appear as a random structure in hydrophilic solutions but form two amphipathic α-helices with a three amino acid-long linker in-between in a hydrophobic surrounding (figure 2c). These types of peptides were recognized as particularly active in membrane bilayers, similar to melittin (a bee venom constituent). It was postulated that several cecropin molecules could disturb bacterial outer membranes and result in lysis in synergy with lysozyme, which acts on the peptidoglycan component of the cell envelope. During this period, six isoforms of P5 were also purified on the basis of antimicrobial activity and named attacins [16]. Chromatofocusing revealed two main groups, one with basic and one with neutral pI. Compared with the cecropins, the attacins had a narrower spectrum and were only active against E. coli, and a few of the investigated Gram-positives. Bennich and co-workers continued working on the mode of action and revealed that the neutral attacins downregulate transcription of several outer membrane proteins in E. coli [17]. A couple of years later, work by Hultmark and co-workers led to the cloning of Drosophila cecropins, which were later found to confer antifungal properties [18,19].

3. En route to a universal humoral response

Following the discovery of cecropin, the quest to identify peptide antibiotics in higher eukaryotes began to slowly gain interest after nearly a decade of dormancy. The era of whole-genome sequencing naturally led to the dismantling of a large array of AMPs orthologous to those originally characterized in H. cecropia. For example, the first defensins had already been discovered as effectors of phagocytosis in the 1960s, but it was first with the sequencing technique that they were renamed and classified as AMPs [20]. A few years later, the first insect defensins were discovered, which strongly supported the idea of a general humoral immune response [21,22].

Another breakthrough was made at the time by Michael Zasloff, a paediatrician at the University of Pennsylvania School of Medicine, who elegantly made use of the methods that were developed in Boman's laboratory to isolate AMPs from frogs. He observed that the African clawed frog Xenopus levis, from which he dissected out the eggs for RNA studies, survived such operations amazingly well. They were returned to a filthy aquarium without becoming infected or showing any signs of inflammatory response. Inspired by the findings in H. cecropia, he decided to search for similar peptides in frog skin. This led to the isolation and characterization of one of the first vertebrate AMPs (after the defensins), designated magainin (meaning ‘shield’ in Hebrew), and the cloning of two of the peptides, which were expressed in tandem in one transcript and cleaved after translation [23]. This was a breakthrough for the recognition of AMPs at the time. Thanks to Zasloff's promoting skills, the discovery gained worldwide public attention. Although magainins have not per se been found in any other species, the finding bolstered Boman's theory of peptide antibiotics as universal innate effectors.

4. A gut feeling: the characterization of vertebrate AMPs from pig intestine

Another important contribution by Boman was the idea that the animal gut must be able to produce AMPs to defend itself from the constant and vast exposure to microbes. In 1987, he contacted Professor Victor Mutt at Karolinska Institutet, an expert on gut endocrinology who had specialized in isolating hormonal peptides from pig intestine. Together with Jong-Youn Lee and Mats Andersson, Boman's group began the process of purifying peptides from the side fractions from Mutt's experiments [24]. In a dramatic moment, the first sequenced peptide turned out to be a cecropin, later named cecropin P1.

While initially considered a major breakthrough, several years passed without any other group succeeding in isolating mammalian cecropins. To banish all doubts, a long and agonizing process was initiated to clone cecropin P1, but without success. The mystery unravelled several years later, when the source was found to be Ascaris suum, a parasitic nematode that resides in the pig intestine [25]. Nevertheless, Boman continued his work on cecropins. A few years after his retirement, Boman moved to the Department of Microbiology, Tumor and Cell Biology (MTC), Karolinska Institutet. A collaboration with Katrin Pütsep and Staffan Normark (PhD student and former co-worker from Umeå, respectively) led to the finding that the ribosomal protein from clinical isolates of Helicobacter pylori displayed antibacterial activity that could be traced back to cecropin-like amino terminals of one of its ribosomal proteins [26]. In one sense, the story of cecropins was finally tied together, albeit in an unexpected manner.

Despite the cecropin P1 sidetrack, the partnership with Mutt proved highly fruitful. Among the key findings was NK-lysin (later found in humans and renamed granulysin) as well as a peculiar peptide containing 18 prolines and 10 arginines out of 39 amino acids, hence named PR-39 [27,28]. The mode of action of the latter differed from most AMPs discovered up to that point as it becomes internalized and inhibits DNA synthesis without disrupting bacterial membrane integrity [29]. Through the work of Zanetti and co-workers, a carrier protein, cathelin, was subsequently found to be connected via its N-terminal domain to several precursors of peptide antibiotics, including PR-39 [30,31]. Peptides that in their preform were bound to cathelin were subsequently commonly classified as cathelicidins. A race began to identify cathelin-bound peptides in humans. It ended in almost simultaneous publications by three independent groups, including Boman et al., all of which had discovered LL-37, the only human cathelicidin [32–34]. In a subsequent study, morbus Kostmann patients were found deficient of LL37 in saliva and plasma, confirming the vital importance of this AMP in humans [35].

5. From regulation to recognition

Towards the end of the 1980s, AMPs had been discovered in a variety of animals, but the underlying mechanisms behind their inducibility were still poorly understood. A great driving force was the theory of pattern recognition postulated by Charles Janeway in his famous presentation ‘Approaching the asymptote? Evolution and revolution in immunology’ at the Cold Spring Harbor Symposium [36]. He argued for the existence of non-clonal receptors for recognition of non-self, and illuminated the importance of an innate immune response. He also defined pathogen-associated pattern molecules (PAMPs) such as LPS from Gram-negative bacteria, lipoteichoic acid in Gram-positives or β-1,3 glucans in fungi, all of which constitute conserved, repetitive microbial structures known to induce rapid and rather unspecific responses in mammals. The theory led to a series of findings, briefly summarized below, which gradually made it evident that non-self-pattern recognition existed early in evolution.

(a). An immunoglobulin-like immune protein

In parallel with the first AMP purifications in 1979, Lee (a PhD student at the time), together with scientists in Umeå, initiated the task of cloning cDNA from AMP genes in H. cecropia using an elaborate procedure called RNA selection. As a postdoc with Gerard R. Wyatt, at Queens University, Canada, Faye demonstrated expression in vitro of immune proteins in H. cecropia fat body, today considered the main AMP-producing organ [37]. Back as a research fellow in Boman's laboratory at Stockholm University, she joined Lee. Using groups of cDNA clones translated in vitro and immunoprecipitated, they identified attacin (P5) and P4 as the most strongly induced proteins upon bacterial challenge [38]. Sequencing of the P4 cDNA revealed the first immunoglobulin superfamily protein that was part of an immune response in insects and was designated hemolin [39]. Several years later, it was demonstrated using X-ray crystallography that hemolin forms a horseshoe-shaped structure unlike any similar molecule known at the time (figure 3a) [41]. More recently, when neural adhesion molecules known as DSCAM had been discovered in Drosophila, it was revealed that the first four Ig domains fold in the same horseshoe structure as hemolin [42]. DSCAM could potentially be involved in immune recognition with great flexibility, because alternatively spliced Ig-domains enable the generation of an estimated 18 000 isoform-specific homophilic DSCAM receptor pairs [43]. A clear interaction between hemolin and LPS was found in 1997 [44]. This was followed up by Kanost et al., who demonstrated its binding to lipoteichoic acid [45]. Hemolin has been proposed to function as an opsonin and participate in phagocytosis, melanization, and to some extent antiviral processes [46,47]. However, while hemolin exists as a single gene in Lepidoptera, it is not found in Diptera, and its putative involvement in non-self-recognition is dubious. Despite an enduring potential, the exact role of hemolin as well as many other Ig-like proteins in insect immunity remains to be resolved.

Figure 3.

(a) The crystal structure of hemolin (courtesy from Xiao-Dong Su). (b) EMSA depicting the binding of CIF to κB-like sequences. Nuclear extracts from E. cloacae-injected and naive pupae were used. F = free κB oligonucleotide probe [40].

(b). The NF-κB regulatory elements

The cloning and sequencing of cDNA was naturally followed by the cloning of the corresponding genes. Crucial in this effort was the creation of a bacteriophage λ library over the H. cecropia genome, which was established by Lee. Partial N-terminal peptide sequences were used to deduce oligonucleotides that could be used to screen for corresponding cDNAs. These were in turn used to screen for the corresponding genes. The cDNAs of attacins and cecropins were first cloned in 1984 and 1985, respectively, and the first gene to become cloned was cecropin B [48]. When the sequence of the attacin gene locus was deciphered, it appeared to contain three genes in a row, two of which were functional, attacin A (neutral form) and attacin B (basic form), whereas the third was a pseudogene. Interestingly, a comparison of upstream promoter regions of the two functional genes revealed an identical glycine-rich and DNase-sensitive decameric sequence 50–60 bp upstream of the TATAA boxes. The missing link was retrieved in a symposium at Uppsala University. One of the invited speakers, David Baltimore, presented the κB element that had just been revealed in an intron of the IgG kappa chain gene and uncovered its basic structural interaction with NF-κB. Baltimore hypothesized that NF-κB was B-cell specific. However, looking at each other Faye and collaborators immediately became aware that this was the upstream elements found in the two attacin genes [48]. Back in the laboratory, they initiated a survey of the cloned and sequenced immune genes including those of Drosophila at the time supported their suspicion. In a series of papers, the evidence for the NF-κB-like factor being crucial in insect immune gene regulation was presented, which also led to the finding of the first insect NF-κB in H. cecropia named Cecropia immunoresponsive factor (CIF; figure 3b) [40,49–52].

These achievements had a bandwagon effect on the community, and the stage was set to unravel all the steps of the insect immunity from recognition to response (figure 4). As novel molecular biology techniques and genetic tools became more important for this task, many laboratories turned to Drosophila as the model of choice. Work was further eased by the establishment of immunocompetent cell lines [54,55]. Further indications of immune signaling conservation between moths and flies was a CIF homologous κB-binding factor revealed in LPS treated Drosophila cells and larvae shown to promote AMP expression [56]. Soon after, the Dorsal-related immunity factor (Dif) was cloned [57]. A few years later, Lemaitre et al. uncovered the requirement of the dorsoventral cassette, including the Toll receptor, for DIF-mediated AMP induction [57]. The same laboratory was early on applying mutant fly lines in survival screens to detect susceptibility to normally innocuous bacteria. This turned out to be a highly important tool to identify novel immune genes. Upon investigating mutants in the phenoloxidase cascade, they discovered by serendipity the role of immune deficiency (Imd) in the bacterial induction of AMPs, crucial for host resistance [58]. The antifungal Drosomycin was, however, in contrast to DIF mutants unaffected in Imd flies, indicating the presence of a second immune pathway. Further evidence for Imd belonging to a separate immune pathway, often acting in concert with the Toll pathway, was provided by Levashina et al. who demonstrated that double mutations of DIF and Imd are required to abrogate the induction of Metchnikowin, and Hultmark's group that discovered the additional NF-κB Relish [59,60].

Figure 4.

Schematic of the main findings by Boman and co-workers (orange boxes), others (blue boxes), or a combination of both (blue and orange boxes). The bomanin family was recently discovered and named in honour of Boman [53].

(c). Peptidoglycan recognition proteins

The receptor for Imd pathway activation was unravelled in a stepwise manner. The peptidoglycan binding property had previously been demonstrated, initially in Bombyx mori, where the interaction could be linked to activation of the phenoloxidase cascade [61]. Steiner and co-workers demonstrated its conservation from insects to humans, and were first to imply its link to NF-κB in both invertebrates and vertebrates [62]. Evidence was put forward a few years later, when soluble PGRP-SA was found to be the bona fide recognition receptor (later additional receptors were found) for Toll pathway activation, acting upstream of Toll [63]. Soon after, peptidoglycan recognition protein LC (PGRP-LC) was found to be required for Relish activation in the Imd pathway [64].

In retrospect, the unravelling of the signalling pathways was likely hampered to some degree by the use of mammalian pathogen-associated molecular patterns (PAMPs). As an example, LPS was originally used in the discovery of DIF and was subsequently found to be the ligand for human Toll-like receptor 4. The original assumption was, however, that the Drosophila Toll pathway should be specifically activated by Gram-positive bacteria and fungi, neither producing LPS. The belief that LPS was a stimulant of Drosophila humoral immunity was plausibly also supported by its interaction with hemolin. In cell line experiments, varying effects were regularly observed with LPS that often derived from fairly crude preparations. It was not until after the finding of the PGRPs for Toll and Imd pathways that peptidoglycans and not LPS were found to be the bacterial PAMPs in Drosophila [65]. The authors strongly demonstrated that LPS, even commercial and proclaimed pure, contained peptidoglycan contaminants, which could explain earlier findings.

6. Concluding remarks

The quest to unravel the humoral immune response initiated by Boman and co-workers opened new paths in the field of immunology. It has contributed to the understanding that the innate immune response is ancient and has evolved to maintain the integrity of different organisms on the Earth. To withstand fast-growing microbes, multicellular organisms use to some extent similar recognition receptors and in particular signalling pathways. These activate and control a variety of genes behind effector mechanisms and molecules directed against the intruder. The fact that innate immunity and its link to the adaptive counterpart became common knowledge is to a large extent contributed by research in insects. One key to Boman's success as a leading figure was his conviction that selecting important unresolved issues drives science forward and renders discoveries that can change an established paradigm. Several years into his 70s, he still determinedly shared this experience through teaching theoretical research methodology. His courses at Karolinska institutet were popular and well attended by students in Biomedical Research. Boman also cared for a creative surrounding in his department and was always engaged in both the technical and the artistic design of offices and laboratories. Last but not least, he chose to interact with people who were creative, original in thought and truly dedicated to research.

Authors' contributions

I.F. and B.L. wrote the manuscript together and approved the final version.

Competing interests

We have no competing interests.

Funding

We received no funding for this study.