| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Perspective |
1 Department of Radiology, Stanford University, 470 University Ave., Palo Alto,
CA 94301.
2 Division of Pulmonary Medicine, University of California, San Francisco, San
Francisco, CA.
Received January 13, 2005; accepted after revision July 20, 2005.
Address correspondence to A. J. Yun
(ayun{at}stanford.edu).
OBJECTIVE. Many chronic diseases exhibit characteristic pulmonary distribution patterns, but the underlying biologic explanations remain elusive. On the basis of emerging evidence from systems biology, we propose that gradients of T helper immune function exist as an epiphenomenon of the hypoxic pulmonary vasoconstriction response. Regional variation of immune function may contribute to preferential distribution patterning of lung diseases.
CONCLUSION. The lungs represent but one example in which the distribution of immune function throughout the body may explain disease location. This hypothetic framework can apply to diseases outside the realm of pulmonary biology and illustrates the potential benefit of integrating advances in systems biology and medical imaging.
Keywords: adrenal gland • autonomic system • bacteria • cancer • chest • fibrosis • fungus • hypoxic • idiopathic pulmonary fibrosis • immunology • immunomodulation • lung • pulmonary sarcoidosis • sympathetic • systemic lupus erythematosus • systems biology • T helper • TH1 • TH2 • vasoconstriction • vasoconstrictor • virus • V/Q
Many chronic diseases exhibit characteristic distribution patterns. Such phenomena have undergone particular scrutiny in the lungs, where diagnosis by traditional roentgenology relies on pattern recognition. For instance, sarcoidosis shows an upper lobe predominance, whereas idiopathic pulmonary fibrosis (IPF) exhibits a lower lobe and peripheral predominance [1]. Such spatial distribution patterns aid in the differential diagnosis of radiographic findings. The association of specific diseases with particular locations has been delineated through empiric radiologic-pathologic correlation, but the biologic explanations underlying the phenomena remain elusive. However, recent evidence from molecular biology, physiology, and immunology has transformed the understanding of illnesses. An integrative or systems biology [2] approach to this newly acquired knowledge may enable radiologists to develop new models to explain the spatial variation of diseases. As an example of this emerging scientific opportunity, we propose a model in which spatial variations of autonomic and T helper (TH) balance play key roles in the distribution patterns of chronic pulmonary diseases. Although this model requires empiric validation, this paradigm could extend beyond the lungs and apply to the distribution of various diseases throughout the body.
TH1 and TH2 Immunity and the Autonomic System
Over the past two decades, the T helper subset model has emerged as a
working construct to explain the dichotomous nature of the immune system
[3]. CD4+ T cells
can be assigned to two distinct subsets based on mutually exclusive patterns
of cytokine secretion: TH1 cells and TH2 cells
[4-7].
Studies in both humans and animals have shown that TH1 cells
characteristically secrete interleukin-2 and interferon-
, whereas
TH2 cells secrete interleukin-4 and interleukin-5
[4-6].
TH1 immunity activates macrophages and drives cell-mediated immune
responses such as delayed-type hypersensitivity, and TH2 immunity
drives humoral immunity through the secretion of IgG and IgE
[3,
8]. Other cytokines, whether
secreted by T cells or not, can also be divided into those that preferentially
promote TH1 or TH2 immune functions
[3]. TH1- and
TH2-like polarizations of cytokine secretion patterns occur in
virtually all components of the human immune system, including CD8+
T cells, B cells, dendritic cells, mast cells, macrophages, eosinophils, and
natural killer cells [9].
The relationship between TH1 and TH2 immunity is viewed as adversarial in that the TH1 immune response in humans directly inhibits TH2 cytokines and vice versa [7]. Studies in murine models have shown that relative strengths of TH1 and TH2 immunity determine polarization to TH1 or TH2 bias [3], thus favoring cell-mediated or humoral immunity, respectively. Although both TH1 and TH2 functions participate in the host defense against pathogens and cancers, a TH1-biased environment is predominantly involved in defense against tumor cells and intracellular pathogens such as viruses, fungi, and mycobacteria [10]. By contrast, a TH2-biased environment is predominantly involved in defense against bacteria [10]. Given their reciprocal inhibition, the TH1 and TH2 immune responses and the human diseases associated with TH1 or TH2 polarization may mutually exclude each other [11-14].
|
through activation of a cAMP-dependent protein kinase A
pathway while stimulating TH2 cytokines
[22-24].
Adrenergic promotion of the cortisol pathway in humans also buttresses a shift
to TH2 bias
[25-27].
Vagal activation down-regulates TH2 inflammation and generally
dampens the immune response elicited by the sympathetic system
[28]. Spatial Variations of Autonomic and T Helper Balance
TH2 polarization and sympathetic bias may exist in the lower and peripheral lungs because of regional variations in pulmonary vascular tone mediated by the autonomic system. A pulmonary physiologic gradient of the ventilation-perfusion (V/Q) ratio exists in humans, with a lower V/Q ratio at the caudal regions of the lung compared with cranial areas [29-31], a situation attributed to gravitational effects on blood and other factors [32]. Studies in both human and animal models have shown that in regions of lower relative ventilation in comparison with perfusion, the peripheral autonomic reflex activates the hypoxic pulmonary vasoconstriction (HPV) response to maintain optimal gas exchange by actively decreasing perfusion to regions with low V/Q ratios [33, 34]. Evaluation of spatial heterogeneity of HPV in the lungs shows higher vasoconstriction in the caudal and dorsal areas of the lung [33]. In addition, a radial ventilation gradient has been noted with lower ventilation at the periphery compared with the hilar regions [35]. These findings suggest that the perfusion to these peripheral areas may be actively limited to optimize V/Q ratios. Pulmonary arterial perfusion studies also suggest that in small-diameter peripheral arteries as compared with large central arteries, there is an increased sensitivity of vasoconstrictor response to relative hypoxia. Studies in both animal and human models suggest that this sensitivity may limit access to high capacitance pulmonary capillary bed [36-41]. Partly because of this fundamental physiology, a relatively high HPV response has been observed in the lung periphery as compared with that of central regions [36-40] (Fig. 1A).
The higher degree of vasoconstriction in the caudal and dorsal areas and
periphery of the lungs may correlate with higher TH2 function in
these areas. Studies in canine models have shown that modulation of
sympathetic function affects the degree to which the HPV response occurs.
Although the HPV response is complex and is believed to involve redox sensing,
calcium and potassium influxes, and endothelial factors, there is also
extensive evidence that the HPV response involves activation of the
sympathetic nervous system
[42,
43]. Sympathetic stimulation
decreases compliance of human and animal pulmonary vasculature via
-adrenergic receptors
[43,
44]. On the other hand, vagal
stimulation promotes pulmonary vascular dilatation in both animals and humans
[44,
45].
Airway mucosal blood flow in humans is decreased with stimulation of
-adrenergic receptors
[46]. The hypoxic
vasoconstriction in obstructive sleep apnea in humans is likely due to
vascular remodeling in the pulmonary vasculature that is initiated by
sympathetic stimulation [47].
On the other hand, vagal stimulation promotes pulmonary vascular dilatation
[44,
45]. Although other
mechanisms, including hypoxia and free radicals, clearly participate in the
complex process of regulating vascular tone in the lungs, these mechanisms are
linked with sympathetic nervous system activity
[48-50].
Therefore, vertical and radial gradients of HPV are likely to be associated
with increasing sympathetic function in the cranial-to-caudal and
central-to-peripheral directions. Furthermore, because a higher sympathovagal
ratio is associated with greater TH2 polarization, TH
gradients may exist along similar axes, with relatively higher TH1
immunity situated in the upper and central lungs and relatively higher
TH2 immunity localized to the lower and peripheral lungs
(Fig. 1B).
Spatial Distribution of Sarcoidosis and IPF
The chronic human pulmonary diseases of sarcoidosis and IPF may follow a
spatial distribution based on preexisting immune gradients in the lungs in
accordance with their opposing TH balance profiles. Sarcoidosis is
associated with TH1-biased environments
[51-53].
Sarcoidosis appears to regress during TH2-polarizing circumstances,
such as infection with HIV
[54], and worsens when treated
with TH1-polarizing agents such as interferon
[55]. Individuals with higher
genetic levels of interferon exhibit greater susceptibility to sarcoidosis
[56]. On the other hand,
evidence from various studies with cytokine and pathologic specimens suggests
that IPF associates with TH2-polarized environments
[57,
58]. Specifically, the
inflammatory response in IPF appears to closely resemble a TH2-type
immune response [59]. Patients
with IPF manifest impaired production of interferon-, a TH1
cytokine [60]. The potential
antifibrotic properties of interferon- served as the mechanistic basis
of a recent multicenter trial evaluating its use in IPF, which revealed
survival benefit in a subgroup analysis
[61,
62].
The TH1/TH2 construct may help explain the preferential distribution of sarcoidosis in the upper and central lungs and IPF in the lower and peripheral lungs [63-65] (Fig. 1C). For patients with sarcoidosis, the relative dominance of TH1 immunity in the upper and central lungs may explain why the disease may start in those areas before progressing to other areas of the lung. Conversely, for patients with pulmonary fibrosis, the relative higher function of TH2 immunity in the lower and peripheral lungs may predispose to earlier involvement of those regions.
Spatial Distribution and TH2 Bias of Other Lung Diseases: The Role of Imaging
In situations in which other data paint a confusing picture, spatial distribution data based on medical imaging may clarify the biologic basis of a disease. For example, diseases such as systemic lupus erythematosus (SLE), chronic eosinophilic pneumonia, and pneumoconioses appear to show both TH1 and TH2 immune dysfunction in both human and murine models [66, 67]. Although the complex nature of the immune dysfunction in these diseases is acknowledged, the peculiar spatial distribution pattern of each disease based on imaging may help biologists define whether a disease is facilitated by a TH1- or TH2-dominant host environment. As might be expected in a condition manifesting autoantibodies [68, 69], a dysfunction of humoral immunity, SLE-related interstitial lung disease in humans generally involves the lower and peripheral lungs [1], suggesting a preferred distribution in TH2 environments [58, 70-74]. Scleroderma, another human collagen vascular disease associated with lower lobe and peripheral lung fibrosis, presents with increased skin expression of interleukin-4 [73], a TH2 cytokine. As might be expected from a condition manifesting abnormal IgE regulation, a TH2-mediated function, chronic eosinophilic pneumonia generally localizes to the peripheral lung [73, 75]. Although substantial evidence indicates that pneumoconioses are TH1-mediated diseases [76-78], some controversy still remains. In our proposed model, the upper lobe predominance of these conditions would further weigh the evidence toward these conditions favoring TH1-dominant environments [79, 80].
Although immune gradients based on sympathetic activation may be a contributing factor to the macrodistribution pattern of pulmonary disease, other factors almost certainly play independent roles. Regional variations in lung structure exist because of the underlying asymmetry of the bronchial tree. Airflow, lymphatic drainage, and penetration of septa do not distribute uniformly across the lung. Acute conditions such as infection may localize on the basis of oxygenation preferences of the pathogen, route of infection, bronchial anatomy, and comorbid pulmonary disorders. Properties of the particle may influence distribution of pulmonary disease related to environmental exposures [81]. The possibility remains that more granular microgradients between the bronchial mucosa and the pulmonary interstitium or vessels may override lungwide gradients. Underlying global TH balance may vary with age or with comorbid conditions [82]. In all cases, chronicity may enable more evanescent processes to resolve and enable immunologic forces to exert a greater degree of influence as to localization. Furthermore, some diseases may manifest a mixture of TH1 and TH2 features as part of their pathogenesis; and in other cases, the disease itself may exhibit significant mechanistic heterogeneity. Collectively, these factors introduce layers of noise such that empiric manifestations are likely to exhibit considerable variation. Thus, in diseases such as sarcoidosis and IPF, in which the immunologic evidence points more clearly toward either a TH1 or a TH2 bias, the distribution within the lungs will likely prove more definitive. Other pulmonary diseases that are more heterogeneous in terms of their immunologic profile may also exhibit a similar heterogeneity in their distribution.
Finally, some disease distributions may be explained simply by the heterogeneity of the architecture of the lung. The stasis of the lymph system in the apical portion of the lung may explain the predilection of tuberculosis for this region. That the apex of the lung also contains the largest alveoli that have the greatest tensile stress may explain the predilection for apical bullae formation in patients with Marfan syndrome [83, 84]. A more comprehensive understanding of the mechanistic basis of more pleiotropic diseases will require better integration of immunologic, radiologic, and pathologic findings.
Future Directions
Our proposed framework requires further empiric validation. Autonomic innervation is present throughout the lungs, and its functional anatomy could be further delineated with iodine-123-metaiodobenzylguanidine SPECT (MIBG SPECT), which identifies sympathetic innervation. Whether spatial variation of autonomic function generates gradients in pulmonary immune function remains to be investigated. As first shown in rodent models, the anatomic basis for autonomic modulation of TH balance in the lung [85] is supported by the recent discovery that the autonomic system directly and diffusely innervates bronchus-related lymphoid tissue, as it does lymphoid tissue throughout the body [15, 16, 86]. Spatial variation of TH balance could be confirmed with region-specific cytokine measurements or mRNA (messenger RNA) array experiments using cells from different parts of the lung to detect gene expression evidence of functional TH biases. Our model would receive further validation from the observation of a therapeutic response. We would expect corresponding shifts in pulmonary distribution patterns in conjunction with response. Once one confirms the robustness of the relationship between witnessed imaging and clinical presentations, monitoring such distribution pattern shifts may serve as an effective noninvasive method of tracking response to treatment. In addition, treatments with particular target specificities could be isolated and developed by observing their ability to alter or produce particular patterns of distribution.
Opportunities exist to extend this framework outside the pulmonary space. Many diseases exhibit eccentric distribution patterns in the body without satisfactory explanation. For instance, the adrenal gland is commonly involved with fungal, mycobacterial, and tumoral entities, which generally prefer TH2-biased environments, but not bacteria, which prefer TH1-biased environments. These propensities may represent epiphenomena related to adrenal production of steroids and catecholamines causing TH2 polarization. With access to vast amounts of in vivo empiric data, radiologists are in a unique position to observe many other repeating distribution patterns of disease. Using models constructed via a systems biology approach may benefit radiologists in how they think and practice.
Conclusions
These examples illustrate the potential role that imaging can play in the emerging field of systems biology. T helper immunology is an exciting systems-oriented field with enormous potential to elucidate the pathogenesis of diseases, but the early controversy and conflicting data are indicative of a paradigm at its embryonic stage. Our framework suggests that integration of medical imaging data into the paradigms of systems biology may play a critical role in advancing many areas of research, even in basic core science disciplines such as molecular biology and immunology. The tools enabling such integration include not only cutting-edge developments of molecular imaging techniques but also intelligent mining of the ongoing body of data collected through conventional imaging techniques.
Validation of our paradigm may prove useful for both systems biologists and medical imagers. Many eccentric distributions of chronic diseases as observed by medical imagers remain unexplained. Similarly, a mass of conflicting data regarding the TH immunity profiles of many diseases confounds biologists. Our model may play clarifying roles in both domains of scientific inquiry. We believe that medical imaging is positioned to become a substantial beneficiary of, and a contributor to, the emerging field of systems biology. We hope that the hypothetic framework presented in this article spurs discussion among radiologists and encourages greater adoption of a systems-oriented perspective that many believe will help change the way we think about diseases, their diagnosis, and their treatment.
References
-producing lung lymphocytes in mouse silicosis.
Am J Respir Cell Mol Biol 1999;20
: 813-824
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
), which exhibits
TH2 features, and of sarcoidosis (X), which exhibits TH1
features, correlate with proposed regions of relative TH2 and
TH1 bias (C). Th = TH = T helper.