Introduction
Human immunity, our natural defense against pestilence, which has underpinned two centuries of scientific research and clinical practice, is now perceived as a major force for both good and evil in medicine—'good' in the sense of its therapeutic and preventative potentials, and 'evil' in the sense that inflammatory responses underlie so many of the major chronic disease pathologies of the twenty-first century. Indeed, the welcome integration of immunology with myriad other disciplines is reflected at immunology meetings in the increasingly common speaker's disclaimer "I'm delighted to be invited to attend this meeting, although I am not an immunologist." The realization that inflammation is the source of many pathologies associated with major internal diseases, such as atherosclerosis and type 2 diabetes, fuels the audacious but plausible possibility that immunology, in the broad sense, may soon dominate clinical medicine and the biomedical research that supports it. For example, a recent article in Nature cited a strong association of obesity-related traits with genetically inherited differential expression of genes involved in the inflammatory and immune response1. Thus, the immunologist's understanding of stress recognition, cytokines, chemokines and cell trafficking may permit better understanding of disease etiology and allow relevant gene products to be developed as practical targets for clinical intervention. Indeed, such understanding will likely be key to functionally interpreting the flood of genetic data emanating from case-control consortium studies. Moreover, such knowledge may prove equally germane to major diseases of the developing world through, for example, the design of vaccine adjuvants and attempts to reduce crippling immunopathology.
In short, these should be rosy times for immunologists, as reflected by the steady growth in the combined academic, corporate and administrative workforce in immunology. This growth is spurred on by the incontrovertible evidence that immunology works in the clinic, for which the blockade of tumor necrosis factor in rheumatoid arthritis and other inflammatory pathologies is but one prime example. Given this, the question of what exactly has happened to human immunology research is frequently, and reasonably, posed. Sampled over six-month periods spanning 2007–2008, only 15% of 54 Nature Immunology research articles and 15% of 130 Journal of Experimental Medicine research papers were devoted solely to human immunology. Although the numbers rise slightly (to 26% and 21%, respectively) if one includes studies that combine human and animal experiments, the clear majority of papers in both journals—143 of 184—was devoted solely to animal model systems. Similarly, an equivalent six-month analysis of Nature Medicine revealed that only 16% (5 of 32) of immunology papers focused exclusively on human immunology, with the majority again devoted to animal modeling. The relative under-representation of 'high-profile' human immunology papers cannot be happening because we know all there is to know: for example, the natural human immune response to Plasmodium, with its curiously compromised memory response, is both unclear and deeply fascinating. Given that career progression is inevitably promoted by high-profile publications, there may be a bias in recruitment and promotion against researchers on the front line of delivering the promise of immunology to a global constituency. Habitual bias is usually self-amplifying.
From a more fundamental perspective, it has been publicly considered, by Mark M. Davis among others, that immunologists cannot as yet define a healthy human immune system. We clearly appreciate undesirable extremes, such as CD4+ cell counts below 500 cells per microliter; positivity for antibodies to neutrophil cytoplasmic antigens (ANCA); or a CD8+ T cell repertoire dominated by cytomegalovirus-reactive cells. Likewise, we can use major histocompatibility complex (MHC)-peptide multimers to measure the frequencies of defined antiviral, microbial or self-specificities that may be relevant to specific scenarios. Nonetheless, an immunologist would in most cases find it difficult to advise a clinician on how to describe a clean bill of immunological health to an enquiring layperson. In considering whether someone is well placed to mount antiviral, antimicrobial or antitumor responses, or conversely is predisposed to autoimmunity or allergy, we simply don't know what numbers of CD8+ T cells or 
T cells are optimal, either in the peripheral blood or in particular tissues; we don't know what constitutes 'healthy' levels of circulating cytokines; and we don't know what balance of memory, effector and naive T cells is appropriate. And this is just in the T cell arena!
More broadly, and to paraphrase Donald Rumsfeld, we even don't know what we don't know. Until we develop a definition of what constitutes a healthy, balanced immune system, immunologists will remain distant from cardiologists, hepatologists, dermatologists and neurobiologists (to name but a few) who can measure and describe organ function in general clinical practice. And yet the greatest share of immunology citations—another route to promotion and grant procurement—continues to go to those pioneers of immunology in animal models, primarily the mouse (http://esi.isiknowledge.com/rankdatapage.cgi). Is this not at odds with the powerful and broadly based momentum for translational research, provided by seemingly rich and diverse funding opportunities at a time of conspicuously low US National Institutes of Health RO1 funding for basic research? Does it not dangerously overemphasize various dogmas, such as the TH1-TH2 cell polarity, that may not be so directly applicable to human pathophysiology? Does it not alarmingly underemphasize the areas in which the human is not well represented by the mouse, such as the regulation of specific antimicrobial responses by vitamin D3? The current global credit crisis is proof that a set of extremely smart people can get things fundamentally wrong, probably because of several factors that quietly but assuredly combined to amplify a basic problem2. This brief Commentary addresses a number of issues that merit scrutiny as part of our obligation to ensure that we (immunologists) do not also get things fundamentally wrong, and that we create every opportunity for human immunology research to thrive. Unless this happens, the obligation of any government or charity to continue funding immunological investigation will be increasingly tested.
Basic research in mice and humans
Immunologists' empowerment to benefit human health draws largely on their impressive knowledge base, which was substantially derived using animal models. From the discovery of H2 histocompatibility by Gorer and Shell, through the fundamentals of lymphoid circulation, to the molecular basis of immunological diversity, the examples are too profound and too many to list. The choice of the mouse as the pre-eminent animal model is fully justified by the number and interventional nature of observations that may be made in this animal, coupled with its steady emergence as a facile and flexible genetic system. Any researcher effectively exploiting this system has the potential to advance fundamental knowledge for posterity and should not be denied that opportunity because of an inappropriate redirection of funding toward less conclusive research in the name of contemporary translational mores. As was noted in the Warry report on increasing the economic impact of the UK research councils3, "The Pharma sector wishes the Research Councils to support blue-skies high-quality research...." Moreover, there are enough disincentives in academic research without questioning and even eroding the motivation of a scientist thrilled by the prospect of truly incisive experiments. But it is quite wrong to imagine that all outstanding basic research is carried out in animal models. The profound advances made using human materials are likewise too many and too diverse to mention, but they run from the revolution in nosology initiated by the discovery of autoantibodies, through the biochemical dissection of immunoglobulins, to the structural elucidation of peptide-MHC recognition. This is not translational research, but because of its focus on human tissues, cells and molecules, this form of incisive basic research is inevitably one step closer to clinical translation (Fig. 1). Therefore, mechanisms should be found to promote more basic research using human materials.
Figure 1: Integration with integrity.
(a) In current common practice, the movement of basic knowledge moves rightward, through the numbered steps, toward ultimate application in medical practice. Examples of the skill sets that contribute to the respective steps are described, with the downward arrows reflecting the current scope of their commitment. (b) In our proposed common practice, by redistributing the commitment of defined skill sets across the steps, an improved integration can be achieved that may optimize and expedite the rightward progression.
Full size image (91 KB)First, basic and translational human immunology research require greater access to more cell types (Table 1), both for their intrinsic characterization and for the elucidation of their interactions with other cell types, for example by establishing organ cultures. This requires re-evaluation of ethical tissue-procurement regulations. Although the potential for abuse needs to be tightly restricted, current regulations may have overcompensated, with the result that precious and useful human material commonly goes (literally) to waste. One encounters instances where the ethical use of patient material is rendered futile by an inability to obtain ethical permission to analyze appropriate controls. Likewise, the use of human materials should be optimized by international agreement on standard protocols for cell and tissue isolation and preparation, augmented by the construction and availability of biobanks for community access. There is currently a danger of creating multiple parallel, insufficiently powered and diverse biobank developments that, though well intentioned, may deliver poor returns on the funding and energies expended4.
Basic understanding in human immunology will also benefit from wider application of humanized mice5. Again, the case can be made for open academic communication on protocols and approaches so that the datasets obtained may be more usefully cross-compared. One might assume that humanized mouse experiments are best pursued where strong expertise in both mouse and human immunology is available. However, remarkably few institutions can claim overlapping strengths in both fields, and the inevitable absence of regular cross-talk between experts in these respective areas seems unhelpful for either party. We are all too familiar with mouse immunologists scurrying about to find some extraordinarily rare disease that they believe phenocopies, and therefore justifies, their new transgenic model. One would think that such energies could be more profitably devoted to genuine issues in basic human immunology that may be amenable to innovative integrations of human and animal model research methods. In expanding human immunology in basic research departments and centers, recruitment and promotion committees may wish to better accommodate the challenges and recognize the merits of human immunology research methods as compared to those used in mouse research, a point considered in the next section.
Applied research in humans (and mice)
An enhanced capacity to establish basic biological mechanisms in human immunology should fuel a momentum to validate (or otherwise) those mechanisms in human health and disease. Such application of basic wet-lab research to improve the design and development of medicines forms one part (known as 'T1') of translational research, as reviewed by Woolf6. If tissue procurement and access are improved, if more consistent protocols are applied, and if a more coordinated input of broad basic research expertise is harnessed, then T1 might seem a relatively routine undertaking. However, this very assumption can create major problems in its failure to confront some habitual differences between animal model and human research.
One difference relates to the perceived representativeness of research findings. It is common (and possibly too easy) for basic researchers, editors, recruitment and promotion committees, and granting bodies to criticize human research (and researchers) because the data seem poorly powered, relating to a single cell line or to tissues from a relatively small group of patients: "If not all patients display a described phenomenon, how important can it be?" But seldom, if ever, is the same criterion applied to biological mechanisms elucidated using only a single strain of inbred mice, even though this is clearly equivalent to repeatedly sampling a single individual, and commonly one with an extreme yet experimentally favorable phenotype. Such a situation would be regarded as absurd if applied to human immunology.
To reduce this perceived double standard, and to helpfully narrow the gap between the basic and applied research cultures, translational research could perhaps begin by clarifying which observations made in mice apply across different strains (or even in genetically diverse outbred mice) and across single strains maintained on different diets, particularly given the influence of vitamins on immune function and the immune response to metabolic stress. This is not to say that the publication of fundamental advances of basic research teams need be held up while the generality of their findings is established. However, determining this generality might quickly identify those phenomena most likely to translate to heterogeneous human cohorts, and, in the case that modifier genes or the environment influence phenotype, the results may offer new conduits to improved understanding. To examine phenomena in different mouse strains and conditions poses few technical hurdles, but to facilitate it, institutions, funding agencies and regulators should review the costs (financial, regulatory and administrative) of rodent husbandry that currently conspire against the advancement of knowledge. As with human biobanks, one wonders whether costs could be reduced, and datasets made more robust and more comparable, through the use of a few central rodent facilities for sets of 'translational studies'. Such facilities might be eligible for large, peer-reviewed strategic grants, contingent on their establishing practical means of collaboration with accomplished basic research teams, so as to jointly expedite large-scale 'tests of generality'.
The challenge of data representativeness evokes a second issue that arises from the assumption that the same molecular skills used in making profound advances in basic research (in animal models and in humans) are all that are needed for high-quality T1 translation. This assumption is most overt in gene-profiling studies, in which one compares the behavior of tens of thousands of variables among a relatively small number of samples—for example, ten afflicted and ten healthy controls. The hazard—that if one scrutinizes enough variables, one is sure to find one (or even a group) that shows a correlation—is well known but largely ignored. It evokes a declaration by one of the characters of the nineteenth-century novelist Elizabeth Gaskell: "I have not been able to spell since I lost my tooth." But this hazard is genuine, as illustrated by recent controversies surrounding the reliability of putative biomarkers for solid tumors that were identified by microarray studies7. In animal models, puzzling correlations of gene sets with phenotype that a priori make little sense may be relatively easily re-examined, and often thrown out. But because this is much harder to achieve in humans, translational research methods need to pre-empt the problem. One approach is to more actively involve experts in biostatistics and randomized clinical trial formulation in the planning of studies. They can help safeguard experimental design (for example, in assuring independence of test and validation samples) and can emphasize the importance of describing precisely the nature, source, harvesting and storage methods for each sample that contributes to a dataset. In the theoretically infinite electronic world of supplementary data, there seems ample scope for such information to be provided, but by contrast to the detailed description of each patient in a clinical trial, it seldom is. Indeed, there are particularly strong concerns over the depth of this type of information provided by companies in relation to their own analyses of patients and patient materials. In sum, an enhanced integration of skill sets in molecular biology and trial design should be a key goal of translational research groups that are being assembled in medical centers across the globe, complemented by a regulated and explicit standard of publishing practice.
The preceding issues inevitably lead to a consideration of the appropriate scale of human immunology research. Ideally, the integration of molecular biology and trial design expertise should position the community to commence immunological phenotyping on the massive scale—comparable to that of the genotyping consortia—that will be required to develop useful and practical metrics for a healthy (and, by inference, unhealthy) immune system. Those metrics will form the passport for the community's entry into standard clinical practice. Funding such ambition requires that immunologists with the right combinations of skill sets proactively lobby funding bodies, politicians and patient advocacy groups, as was the case for the genetic case-control consortia studies that have had such a huge impact on medical research8. Ironically, the potential impact of those genetic studies has been worryingly hampered by a lack of progress in functional immunophenotyping. To cite just two examples, the genetic association of NOD2 with Crohn's disease was a landmark in complex disease research that, five years on, still awaits immunologists' general agreement on whether the mutation impairs or exaggerates an individual's response to microorganisms. Likewise, there is still some uncertainty as to whether overexpression of the polymorphic 'stress antigen', MICA, that is common in human cancer and inflammation should be viewed as promoting or suppressing immune responsiveness. In each case, the appropriate clinical strategies are 180° apart, depending on the conclusions obtained. Beyond laudable and ongoing basic research in these respective areas, it would seem that the most effective way to resolve such issues is to monitor immune function in large numbers of relevant individuals. By comparison to DNA analysis, the challenge of longitudinal live cell monitoring (possibly coupled with tissue imaging) is not trivial, but it must be accomplished.
Such an enterprise will require a level of public engagement that goes well beyond the focus on HIV/AIDS, 'bird flu' and biodefense. Immunologists, perhaps via the coordinated efforts of their academic societies, need to be imaginative in capturing the hearts, minds and thus the peripheral veins of the lay public. The potential rewards of immune monitoring in ameliorating symptoms of major diseases, in offsetting the development of autoimmune disease, and in fighting emerging infections need to be clearly and openly discussed, through all sorts of mechanisms and at all sorts of levels. Who, among the readers of Nature Immunology, has recently addressed a group of schoolchildren? "Bravo" to those who can answer "yes". In short, effective T1 translation requires high-quality efforts and skill sets that go beyond the simple extrapolation of basic science methods, and deserve recognition by recruitment, promotion and funding committees, among others.
Troubles in translating translation
According to the terminology reviewed by Woolf6, a major component of translational research ('T2') focuses on how to extend advances in research and medicine to the people. Much of this centers on health services research, but it also should include the development, operation and review of clinical research facilities (CRFs). These are the sites where advances in basic research may be integrated with refined clinical datasets so as to take new ideas into 'proof-of-principle' tests of clinical intervention and monitoring. Such CRFs seem to be shooting up in many academic medical centers, fuelled in part by directed funding initiatives (http://www.wellcome.ac.uk/Achievements-and-Impact/Initiatives/UK-biomedical-science/Clinical-Research-Facilities/WTD003492.htm). Appropriately, these seem to be of great interest to the pharmaceutical industry, which perceives them as unique means to obtain detailed information germane to target validation and drug action under highly controlled conditions.
Nonetheless, it remains very difficult to interest industry in early-stage ideas for treatment. Big Pharma seems obsessed with big diseases and with extending the broadest possible application of new drugs in their earliest years of patent protection. Given the hundreds of millions of dollars it takes to get that far, one can hardly blame them. Likewise, so long as the clinical germination of a new idea, born in academia, depends on commercial spin-outs or in-licensing, it risks fatal collision with the detailed aspects of commercial viability that are the understandable preoccupations of corporate boardrooms. Hence, obtaining funding for investigator-initiated trials, particularly for new targets that may require expensive, newly formulated Good Manufacturing Practice (GMP)-grade material, remains a major disincentive for academic researchers, even when assisted by growing university infrastructures for regulatory, legal and commercial support. Indeed, such structures may actually create additional obstacles, as universities' insistence on patenting discoveries at the time of the earliest observation often precludes dissemination of data and delays collaboration until a relatively late stage in a process. Thus, one needs to consider mechanisms for greatly facilitating small clinical trials that clearly establish the validity of a target in a defined disease scenario, even if the pharmacology initially used to manipulate it is not commercially viable. Perhaps there is an obligation for industry to contribute (based on an activity-related index) to a 'superfund' that supports a series of such activities selected through independent peer review. Along those lines, GlaxoSmithKline made an initial multimillion-pound pledge to fund a 'global medical excellence cluster' (GMEC) that was recently formed in the UK and was very much modeled on US initiatives. Only time will tell, however, whether this structure can effectively address the most pressing gaps in the early phases of drug discovery and development. Moreover, in instances of major, broad-based industrial-academic collaboration, academic researchers probably need clear assurance that facilitated translation aims to bring broad benefits and not simply to advance the agenda of one or a few companies whose liaison the researchers did not themselves develop.
Conservative commercial perspectives create additional hurdles for human immunology research. Because the very roots of immunology lie in vaccination and the development of lifelong protection, be it against infectious diseases, malignancy, graft rejection or autoimmunity, true success in immunology is only realized in the long term. Therefore, there is a justifiably intense focus on identifying short-term biomarkers of long-term prognosis. But clearly, such biomarker validation requires long-term studies (preferably as part of mass immunophenotyping) that fatally conflict with alternative commercial opportunities that can be realized much more quickly. Again, a dedicated private-public fund for long-term studies may be an appropriate vehicle to reduce this obstacle to longitudinal charting of human immune responses. The consideration of policy options such as 'superfunds', windfall-tax levies and long-term longitudinal patient monitoring would benefit from the involvement of bodies not obviously representing the interests of a single party. Again, academic societies spring to mind, with their obligation to represent both corporate and academic members, to act in the best interests of immunology and to operate above the (understandable) parochialism of individual universities. In this regard, we note the rapid and potentially powerful development of the Federation of Clinical Immunology Societies (FOCIS).
Troubles in training
A further hurdle in T2 that academia ought to be able to surmount is a serious anachronism in clinical immunology training. Although Jean-Laurent Casanova has imaginatively considered that the majority of the population may suffer from one form or another of primary immunodeficiency9, clinical immunology remains an unglamorous, orphan clinical specialty, commonly perceived as focusing on a tiny number of patients suffering from obscure diseases. By contrast, autoimmune and chronic inflammatory diseases, from which over 1 in 20 of the population will suffer at some stage of their life, affect endocrine organs, joints, vessels and nerves, and are currently managed by the appropriate clinical specialist; yet the 'cure' for these disorders, when it comes, will be immune-based. Likewise, the responsibility for applying and monitoring new protocols in tumor immunotherapy rests largely with oncologists, who in most cases cannot claim critical expertise in the relevant area of contemporary immunology.
Starkly illustrating specific limitations in training, in 2008 only 6 of 236 entrants (2.5%) into academic clinical fellowship posts in England and only 1 of 85 entrants (1.2%) as clinical lecturers will be in immunology. A further consideration is whether this represents either the only or the optimal training route through which to plug the gaps in our clinical immunological repertoire, or whether we should be striving to achieve a much higher level of immunological cognizance and training among all medical and dental graduates and postgraduates, so as to prepare them for the important role that human immunology will play in their careers. In the UK, the Department of Health, via the National Institute for Health Research that it recently spawned, is promoting a far greater involvement in research design and implementation for clinicians, rather than concentrating on a few enlightened clinical academics (who themselves often see very few patients). Early indications suggest that including greater numbers of 'jobbing clinicians' in the discovery process is extremely beneficial. This goal will ultimately be realized by improving science education for all medical students. Unlike in the United States, in many countries medicine is taught as a bachelor's degree, and in such instances there can be a woeful inattention to the basics of scientific method and investigation. For all the limitations of studies on inbred mice noted above, the rigor that their uniformity has brought to fundamental research needs to be appreciated by all medical students. A more scientifically literate and 'involved' clinical work force collaborating with active research groups holds the key to developing incisive biomarkers that distinguish separate etiologies of phenotypically similar patients, as in, for example, lupus attributable to B cell dysregulation and to T cell dysregulation. Such an approach will certainly expedite clinical trials and improve their quality in countries willing to invest in it, a point that will hopefully be recognized by the pharmaceutical industry and by politicians alike.
Summary
The goal of this article has been to raise the profile of some key issues germane to promoting research in human immunology. Only when those and other issues are carefully considered by appropriate groups can concrete recommendations be made. This notwithstanding, we are keen to contribute a few policy pointers (Table 1). We reiterate that human immunology will suffer if constraints are placed on basic research, in terms of both content and structure. The reality is that basic immunological research already has a proven track record, which is something that cannot yet be claimed for translational research. Nonetheless, statements like this one that pay habitual lip service to the role of basic research are only helpful if supported by proposals for how basic research can best promote translation. We believe that the use of funds to coerce, overtly or tacitly, the conversion of basic researchers to translational researchers will most often not be helpful. Rather, the resources should be used to drive the effective integration of respective skill sets across the spectrum of endeavor (Fig. 1). Basic research can be effectively undertaken using human materials, creating a direct platform for translation. Indeed, some of the tensions between basic and translational research might be lessened if such a platform encouraged translational initiatives to be facilitative, rather than prescriptive, vis-à-vis basic research programs. In that regard, universities and funding bodes alike need to recognize that effective translation will require the integration of different expertise across the stepwise progression of research (Fig. 1).
Integration should be at a local and global level. It would be quite inappropriate for an accomplished research team to feel obliged (for example, by ill-thought-through funding imbalances or institutional pressures) to shift its focus from the incisive study of nominal antigen responses in mice to, for example, the study of adaptive human responses to a microbe, chosen simply to 'tick the translational box'. But it is not inappropriate for the team to be deeply conscious of the importance and the difficulties of understanding the human adaptive response to an infectious disease with substantial morbidity and mortality, such as dengue. We consider that funding support for groups studying such problems in human immunology might be contingent on demonstrated and ongoing consultation with named experts in a relevant area of basic animal-model research. In turn, basic science funding may be contingent on a group's active involvement in such 'translational consultation' (Fig. 2). Whereas grant forms currently ask an applicant to describe details of their public engagement, they might equally usefully request details of consultative involvement in front-line studies of human immunology. Active consultation, as opposed to larger-scale collaboration, could bring researchers together without jeopardizing the integrity and trajectory of their respective research programs.
Figure 2: Incentivized consultation.
The funding agencies supporting basic research and applied research, respectively, each request that their investigators consult with each other on specified projects. This allows specialists to maintain their areas of expertise while exchanging information in areas of proven relevance to the human immunology of infectious diseases and internal diseases, respectively, affecting both the developing and developed world.
Full size image (58 KB)Policy innovations such as this can inevitably be construed as "investing in 'proven people'." Although this is common practice in industry, it remains a sensitive issue in academia. Nonetheless, it may provide a means to pilot various forms of focused collaboration, via specific programs and/or training schemes developed by funding agencies, in consultation with the research community. Some of these collaborations, such as human genetics and functional human immunology, have been referred to above. One short-term achievement might be a substantially improved gene-ontology annotation that provides meaningful functional immunology data relevant to thousands of gene entries. Other collaborations, using funding as an incentive, might integrate advanced pathogen classification—as undertaken at the US Centers for Disease Control or the UK Health Protection Agency—with studies of immune responsiveness, of pathogen virulence and of tissue-pathology. Respecting separate skill sets and then getting key representatives of each to work together at various levels (Figs. 1 and 2) seems a practical approach that preserves the best prognosis for incisively advancing knowledge in human immunology.
Note: The authors welcome feedback on the issues raised by this Commentary, no matter what degree of disagreement with us this may reveal.
