This study reveals the presence and distribution of lymphatic vessels in peripheral compartments of human normal lungs and shows for the first time that in advanced COPD the most manifest increase in lymphatic numbers takes place in the alveolar compartment. This study further shows that in advanced COPD lymphatic vessels exhibit an altered phenotype characterized by up-regulation of CCL21 and D6. Considering the role of lymphatics in transporting leukocytes, these data suggest that lymphatic vessel changes may have important roles in the immunopathology of patients with severe stages of COPD.
The morphological evaluation of lymphatic vessel changes in the present study revealed further that the absolute numbers of bronchiolar and arterial wall lymphatic vessels are increased in lungs of patients with severe COPD compared with control subjects. However, normalization to tissue area did not reveal a similar increase in lymphatic vessel numbers. These results suggest that in these compartments tissue remodelling, such as thickening of the airway and arterial adventitia, is accompanied by a proportionally increased formation of lymphatic vessels. In bronchioles lymphatic vessel changes may also involve formation of more folded lymphatic vessels since the accumulated lymph endothelial length per area unit was increased in this compartment. The most clear lymphatic vessel changes were, however, observed in the alveolar parenchyma, a compartment where under normal conditions the lymphatics are relatively sparse in numbers.
The present type of morphological lymph vessel changes, which were most pronounced in advanced stage COPD, is likely to have several implications for the immunopathology and pathophysiology observed in this disease. As COPD becomes severe, inflammatory processes are significantly amplified . Thus, the increased numbers of lymphatic vessels in advanced COPD could result from an increased demand for clearance of senescent leukocytes and extravasated fluid from the inflamed peripheral lung regions. The observed increased numbers of lymphatic vessels in this and a recent publication  may also reflect an increased demand for transporting activated immune cells from the sites of inflammation to draining lymph nodes. In support of this, this study reveals that the lymphatic endothelial expression of the chemokine CCL21 is increased in all peripheral lung compartments of patients with severe COPD. Among the major cells expressing the receptor for CCL21 are antigen-activated dendritic cells [20, 37]. In vitro studies have shown that although CCR7 expression on dendritic cells may decrease in response to tobacco smoke, their capacity to migrate towards CCL21 is preserved . Thus, the increased expression of CCL21, which is critical for the entry of immune cells into the lymphatic vessels, may facilitate the migration of activated dendritic cells  and enhance antigen presentation in draining lymph nodes. In addition to CCL21, our data also show that the lymphatic expression of D6 was increased in advanced COPD. Recent studies have suggested that lymphatic D6, which scavenges inflammatory CC chemokines, controls the flow of fluid and migration of appropriate immune cells to lymph nodes . Previous studies have also demonstrated that D6 is expressed on leukocytes . Interestingly, patients with COPD have increased percentage of D6-positive alveolar macrophages  indicating that D6 may generally be up-regulated in COPD lungs. Even though D6 is mainly expressed in lymphatic endothelium, this study confirms previous observations  that D6 is not expressed in all lymphatic vessels. It has been proposed that the lymphatic expression of D6 is up-regulated by pro-inflammatory cytokines . It is likely that this mechanism is partly responsible for the presently observed up-regulation in COPD lungs. Interestingly, the mechanisms of up-regulation may also differ between the anatomical regions of the lung. For example, in the bronchioles almost 100% of the lymphatic vessels expressed D6 already in control lungs. Hence, in this compartment the increase in total lymph vessels-associated D6 in COPD may mainly be a consequence of a general increase in vessel numbers. On the other hand, in the alveolar parenchyma where only around 40% of the lymphatic vessels expressed D6 under baseline conditions, the up-regulation of D6 immunoreactivity in COPD lungs was also caused by a significant increase in the proportion of vessels expressing D6 (which in COPD was increased to around 80%). A similar increased proportion of positive alveolar lymphatic vessels was also observed for CCL21. Taken together, the combined action of increased lymphatic expression of D6 and CCL21 and a general increased density of lymphatic vessels in COPD lungs is likely to increase the capacity for regulated trafficking of leukocytes from the distal parts of the lung to draining lymph nodes.
This study further confirmed a close proximity between lymphatic vessels and ectopic lung lymphoid aggregates [12, 43]. In addition to transporting leukocytes from the lung to draining lymph nodes, lymphatic vessels may efficiently transport leukocytes, as well as antigens, to local lymphoid aggregates in COPD lungs. Since we observed CCL21-immunoreactive lymphatic vessels within lymphoid aggregates, one could also speculate that these lymphatic vessels may offer an important route for leukocytes to exit from the lymphoid aggregates. Similar to lymph nodes, the lymphoid aggregates in COPD lungs contain lymphocytes, follicular dendritic cells and germinal centres [11, 13, 43] and are, thus, capable of initiating adaptive immune responses locally in the lung .
The molecular mechanism behind the increased number of lymphatics in COPD is currently unknown. It is of note, however, that immune cells such as macrophages may contribute to lymphatic vessel formation by their production of lymphangiogenic factors, including VEGF-C and VEGF-D [32, 45, 46]. Pro-inflammatory cytokines, many of which are up-regulated in COPD, such as TNF-α, can also regulate lymphatic vessel growth [47, 48].
At a more upstream level respiratory infections may also induce the formation of lymphatics. Elegant studies in animal models of chronic airway inflammation have demonstrated that lymphangiogenesis, as well as remodelling of lymphatic vessels, occur extensively in lungs infected with Mycoplasma pulmonis[32, 49]. Although other pathogens are generally associated with COPD, the majority of patients with advanced COPD suffer from recurrent infections of the lower respiratory tract  and this may be one of the driving forces for the presently observed peripheral lymphatic vessel changes in this patient category. In any case, an expanded lung lymphatic system is likely to result in a faster clearance of pathogens. Thus, the increased number of lymphatic vessels in major lung compartments in COPD may be a double-edged sword; a tool for accelerated adaptive immune responses in response to infections, and a structural basis for an aggravated inflammation in COPD lungs.
It is of note that once established an expanded lymphatic system may persist for considerable time . Interestingly, after Mycoplasma pulmonis infection, remodelled blood vessels, in contrast to lymphatics, normalized readily once the underlying inflammation resolved . This observation suggests that remodelling of the lymphatic system should be expected to be long lasting and relatively resistant to anti-inflammatory treatment. This notion is supported by the fact that lymphatic vessel alterations due to infections are steroid resistant . In this study lymphatic vessel changes were most pronounced in patients who received inhaled corticosteroids indicating that also in human an expanded lymphatic system may not normalize readily upon steroid treatment. Future studies in larger COPD cohorts are, however, warranted to elucidate the detailed effects of steroids on newly formed lung lymphatics.
Although lymphatic vessel changes have been demonstrated in several lung diseases [33, 51], it remains unknown if de novo formed lymphatics are fully functional in terms of interstitial fluid clearance. For example, in vivo animal studies have shown that newly formed lymphatics in inflammation have an altered structural phenotype that could impair clearance of fluid [52, 53]. Whether or not the increased number of lymphatics in COPD lungs sufficiently compensates increased plasma leakage in, for example, an exacerbation remains to be investigated.
As indicated by this study, the role, extent and type of lymph vessel change may differ between the different compartments of the lung. Indeed, our data suggest that the relative increase in numbers and activation differed between bronchioles, vessels and the alveolar parenchyma. Unfortunately, our study material did not contain enough pleural material to justify a meaningful quantitative analysis in this compartment, which normally is rich in lymphatic vessels. Nevertheless, we did analyse other microenvironments such as patchy fibrotic lesions that can be observed in advanced COPD. Confirming other studies [27, 28] our study revealed a marked increase of lymphatic vessels in fibrotic lesions. Interestingly, while newly formed lymphatic vessels normally develop from existing ones through sprouting, in fibrosis some of the newly formed lymphatic vessels lack connection to existing lymphatics and may thus have a different growth pattern [27, 28]. Whether similar mechanism for growth of lymphatic vessels are active in the fibrotic lesions that develop in advanced stages of COPD remains to be investigated.
Although conflicting data exist regarding the presence of lymphatic vessels in alveolar parenchyma [27, 28, 54], alveolar podoplanin-immunoreactive lymphatic vessels have been detected in normal human lungs [36, 55]. Also in this study we could observe lymphatic vessels in the alveolar parenchyma. However, it cannot be excluded that some of the alveolar lymphatic vessels may be associated with pulmonary blood vessels not visible by the 2D-projections provided by conventional histological sections. Also, in our 2D analysis it cannot be excluded that the increased numbers of lymphatic vessels may to some extent be caused by the folding of vessels into complex 3D-structures. The overall orientation of blood vessels and bronchioles was in the present study, however, random and equal among the study groups. Therefore, the presently observed differences in lymphatic vessel changes are unlikely a result from biased tissue orientation.
In this study patients with suspected lung cancer were included. This may be relevant information when investigating lung lymphatic vessels as lymphangiogenesis is of importance in tumour metastasis . Lung tissue specimens used in this study were, however, obtained as far away from the tumour as possible in order to minimize the risk of cancer as potential contributor to the expanded lymphatic system revealed in COPD. Two factors strongly argue that this is not the case. Firstly, the patients of our non-COPD control groups had similar types of solid tumours, suggesting that all differences in the COPD-group are truly COPD related. Secondly, increased lymphatic vessel numbers were detected in patients with very severe COPD who lacked any history of cancer and where the tissue was collected in association with lung transplantation. As most histological human studies, another limitation of this study was the relatively small study groups, a factor that we in this study have tried to compensate by exploring several lung regions, each containing multiple lung compartments, in each study subject.