Nonclassical Monocytes Promote Edema in Lung Allografts from Traumatic Brain Injury Donors

To the Editor:

Head trauma remains one of the most common causes of brain death among human lung donors. Unfortunately, traumatic brain injury (TBI)-associated lung dysfunction precludes a large proportion of such lung donor offers from clinical transplantation (1). Even those lungs from TBI donors with acceptable gas exchange prior to transplantation demonstrate increased susceptibility for primary graft dysfunction, the predominant cause of early post–lung transplant mortality. However, the mechanisms through which TBI leads to donor lung injury and subsequent primary graft dysfunction remain unclear (2). In this letter, we report that immunological mechanisms driven by intravascular CCR2⁻ nonclassical monocytes (NCM) may play a role in the pathogenesis of pulmonary edema associated with TBI.

Using our previously validated murine model of TBI (3) (Figures 1A and 1B), we found that TBI was associated with the development of pulmonary edema (Figure 1C). Multicolor flow cytometry demonstrated a significant increase in intravascular pulmonary Ly6c^low NCM and extravascular natural killer (NK) cells in lungs of mice undergoing TBI (Figures 1D and E1 in the data supplement). In contrast, other myeloid cell populations, including neutrophils and classical monocytes, remained unaltered. Human lungs from TBI donors similarly demonstrated increased intravascular CD14^lowCD16^high NCM, despite being rigorously perfused for clinical transplantation (Figures 1E and E2).

Figure 1.

Nonclassical monocytes (NCM) promote lung edema from traumatic brain injury (TBI). Mouse brains were harvested for determination of (A) morphology and (B) neurological severity score, from sham and TBI mice (n = 4–5). (C) Lungs were harvested for determination of pulmonary edema by wet-to-dry ratio measurement following pharmacological and genetic depletion of the different subset of monocytes (n = 3–7). (D) Number of NCM and NK cells following TBI in mice (n = 3). (E) The ratio of NCM/AM in the lungs isolated from human non-TBI and TBI donors. Data expressed as cell count per alveolar macrophage (AM) to standardize across patients (n = 10). (F) NCM isolated from mouse lung secrete TNF-α in response to HMGB1 (n = 4–5). All graphs show means ± SD. Graph on (C) was analyzed by one-way ANOVA followed by Tukey’s post hoc test. All other graphs were analyzed by two-tailed unpaired t test. *P < 0.05, **P < 0.01, and ****P < 0.0001. B/L = bilateral; HMGB1 = high-mobility group box-1 protein; NK = natural killer; n.s. = nonsignificant.

To determine whether monocytes played a role in TBI-induced pulmonary edema, we administered clodronate liposomes, or control PBS liposomes, to deplete both classical monocytes and NCM (Figures E3A and E3B), 24 hours before inducing TBI, as previously described (3, 4). Pulmonary edema was prevented in mice receiving clodronate liposomes but not in those injected with control PBS liposomes (Figure 1C). However, pharmacological depletion of classical monocytes, which constitutively express CCR2, through the use of anti-CCR2 antibodies (Figures E3C and E3D) had no effect on TBI induced pulmonary edema (Figure 1C). The NCM are dependent on the orphan nuclear factor Nur77, which is genetically deleted in Nr4a1^−/− mice. These Nr4a1^−/− mice did not develop pulmonary edema following TBI (Figure 1C). In contrast, depletion of the NK cells using anti-NK1.1 antibodies did not confer protection against pulmonary edema (data not shown).

We next investigated whether lungs from human TBI donors were susceptible to allograft dysfunction. To specifically determine the function of the transplanted lung, we collected blood from the point of confluence of the superior and inferior pulmonary veins and analyzed the arterial oxygen pressure (Pa_O2)/fraction of inspired oxygen (Fi_O2) ratio. We chose to perform this analysis at 30 minutes following reperfusion, as the chest cavity of the recipient is still open at that point and the pulmonary veins of the lung allograft are accessible. TBI donor lungs had a significantly decreased Pa_O2/Fi_O2 ratio (Figure 2A). We modeled these findings in murine single left lung transplantation using lungs from TBI or non-TBI donors. At 24 hours, the hilum of the right lung was clamped and the gas exchange of the transplanted left lung analyzed. Similar to the humans, TBI lungs from murine donors had significantly reduced Pa_O2/Fi_O2 ratio (Figure 2B). As expected, the Pa_O2/Fi_O2 ratio did not decrease when TBI lungs were from Nr4a^−/− mice donors (Figure 2B). Neutrophil counts were increased in the blood of TBI when compared with the non-TBI group (percent CD45⁺: TBI, 6.13 ± 1.12 vs. non-TBI, 26.3 ± 11.1; P < 0.01), and although both groups experienced an expected neutrophil influx in the lungs resulting from ischemia–reperfusion injury, the neutrophil counts were similar in TBI and non-TBI groups at 24 hours (TBI: 1.6 × 10⁶ ± 0.7 × 10⁶ cells vs. non-TBI: 1.4 ± 0.9 × 10⁶ cells; P = 0.8). Collectively, these findings suggest that TBI is associated with the recruitment of NCM, which promote the development of pulmonary edema. However, unlike lung transplant ischemia–reperfusion injury (5), the TBI-associated pulmonary edema and post–lung transplant dysfunction are perhaps not dependent on the ability of NCM to recruit neutrophils into the lung.

Figure 2.

TBI is associated with worse allograft function following both murine and human lung transplantation. Allograft function measured by Pa_O2/Fi_O2 ratio at 24 hours after transplant of human (n = 10) (A) and murine (B) lungs (n = 5–12) is shown. (A) Following human lung transplantation, the function of the allograft was analyzed by obtaining the blood directly from the pulmonary vein of the allograft. (B) For murine lung transplantation, the blood was obtained from the aorta after the hilum of the native right lung was clamped for 5 minutes and the host was on 100% Fi_O2. Graphs show means ± SD (A) analyzed by two-tailed unpaired t test and (B) by one-way ANOVA followed by Tukey’s post hoc test. *P < 0.05. Pa_O2/Fi_O2 = arterial oxygen pressure/fraction of inspired oxygen.

Although the pathogenesis of pulmonary edema remains poorly understood, the prevailing hypothesis suggests that sympathetic overstimulation and catecholamine storm resulting from the traumatic brain injury can lead to generalized vasoconstriction (6). This causes elevated left atrial pressure and pooling of blood into the pulmonary circulation, resulting in increased pulmonary hydrostatic pressure. The observation that supports a predominant role of the sympathetic system is that stellate gangliotomy, but not bilateral vagotomy, prevents neurogenic pulmonary edema in mice (7, 8). Increased permeability that may be attributed to the stimulation of α- or β-adrenergic receptors by sympathetic overactivity or cytokine release has also been shown to be necessary for pulmonary edema (9). Here, we report a complementary immunological mechanism mediated by intravascular NCM that may contribute to the permeabilization of endothelium.

The mechanisms through which NCM are recruited to the lungs and mediate pulmonary edema need to be further investigated. However, one possibility might be related to the release of HMGB1 (high-mobility group box-1 protein), an important damage-associated molecular pattern, by injured neurons after TBI (10). HMGB1 increases the expression of ICAM (intercellular adhesion molecule-1) on the endothelium via RAGE axis and promotes the binding of circulating NCM to the lung endothelium through the receptor LFA-1 (lymphocyte function-associated antigen 1) (11). Intriguingly, we found that HMGB1 also promotes the secretion of TNF-α by the freshly isolated NCM (Figure 1F), which is known to promote the macromolecular flux through the paracellular route of the vasculature by opening intercellular gaps in endothelium (12). This may explain the development of pulmonary interstitial edema without the recruitment of neutrophils in lungs from TBI donors. If proven to be true, this hypothesis may be clinically relevant because several TNF-α blockers, such as infliximab, etanercept, adalimumab, certolizumab pegol, and golimumab, have been approved by the Food and Drug Administration for clinical use. We have reported that the secretion of cytokines by the NCM is dependent of TLR (Toll-like receptor) signaling. Subsequent investigation is also necessary to identify the specific TLRs involved in the activation of the NCM, which may also provide clues to the pathogenesis of lung injury in the setting of remote organ damage. Finally, as clodronate is a first-generation bisphosphonate, depletion of intravascular NCM could be possible through the use of next-generation bisphosphosphonates to attenuate TBI-associated lung injury.

In conclusion, our data suggest that monocytes promote the development of pulmonary edema in response to TBI and predispose to allograft dysfunction after lung transplantation. Given the preponderance of human lung donors with TBI, treatment directed toward ameliorating lung injury associated with TBI could potentially improve lung utilization and outcomes following lung transplantation.