Primary graft failure and chronic lung allograft dysfunction (CLAD) limit lung transplant long–term outcomes. Various lung diseases have been correlated with surfactant protein (SP) expression and polymorphisms. We sought to investigate the role of SP expression in lung allografts prior to implantation, in relation to posttransplant outcomes. The expression of SP-(A, B, C, D) mRNA was assayed in 42 allografts. Posttransplant assessments include pulmonary function tests, bronchoscopy, broncho-alveolar lavage fluid (BALF) and biopsies to determine allograft rejection. BALF was assayed for SP-A, SP-D in addition to cytokines IL-8, IL-12 and IL-2. The diagnosis of CLAD was evaluated 6 months after transplantation. Lung allografts with low SP-A mRNA expression prior to implantation reduced survival (Log-rank p < 0.0001). No association was noted for the other SPs. Allografts with low SP-A mRNA had greater IL-2 (p = 0.03) and IL-12 (p < 0.0001) in the BALF and a greater incidence of rejection episodes (p = 0.003). Levels of SP-A mRNA expression were associated with the SP-A2 polymorphisms (p = 0.015). Specifically, genotype 1A1A0 was associated with lower SP-A mRNA expression (p < 0.05). Lung allografts with low levels of SP-A mRNA expression are associated with reduced survival. Lung allograft SP-A mRNA expression appears to be associated with SP-A gene polymorphisms.
broncho alveolar lavage fluid
chronic lung allograft dysfunction
chronic obstructive pulmonary disease
diffuse alveolar damage
Enzyme-linked immunosorbent assay
forced expiratory volume in 1 second
interstitial lung disease
multi organ failure
polymerize chain reaction
primary graft dysfunction
restriction fragment length
surfactant protein A, B, C, D
Within the past three decades, lung transplantation has become an increasingly applied therapy for end-stage lung disease. However, clinical outcomes are affected by a variety of stressors including donor brain death, graft preservation, ischemia/reperfusion injury, rejection, medication toxicity, acute/chronic infections and gastro-esophageal reflux with aspiration [1-3].
Pulmonary surfactant and the related proteins (surfactant protein, SP) serve as one of the first defense mechanisms the lung can mount against various insults. Surfactant phospholipids, in addition to lowering the alveolar surface tension, serve as a part of the physical mucosal barrier . SP-A and SP-D play important roles in the innate host defense and may serve as cross–talk proteins between the innate and the adaptive immune systems . SP-B and SP-C play important roles in surfactant function and viscoelasticity, and, along with SP-A, are involved in the phospholipid homeostasis . A genetic regulation of the quality and quantity of SP has been suggested by the association of SP-A and SP-D variants with various lung diseases: respiratory distress syndrome, idiopathic pulmonary fibrosis, emphysema and lung cancer [6-13].
We hypothesized that donor lung SP-(A, B, C, D) mRNA expression prior to implantation is associated with lung transplantation outcomes. We studied SP mRNA expression of the donor lung prior to implantation, in association with the development of chronic lung allograft dysfunction (CLAD)  and survival. The association between SP-A mRNA expression and SP-A1/SP-A2 gene polymorphisms was also explored. SP-A and SP-D proteins were measured in the posttransplant surveillance broncho-alveolar lavage fluid (BALF), as well as IL-8, a proinflammatory neutrophil chemotactic cytokine; IL-12, a key cytokine in the stimulation of NK cells and T lymphocytes, and in the Th1 lymphocyte differentiation; and IL-2, a growth factor for antigen-stimulated T lymphocytes, also responsible for the proliferation of antigen-specific cells and NK cells [14-16].
Forty-two patients transplanted between 2002 and 2003 at the University of Toronto were included in this retrospective analysis of prospectively banked samples. Samples were collected specifically to test the hypothesis of gene and protein expression associated with posttransplant outcomes [3, 17]. Thirty–seven patients were selected based on completeness of sample set with regard to availability of allograft samples and posttransplant BALF. Five consecutive recipients with 30–day mortality also transplanted in the same period were included in this study.
The protocol was approved by the Research Ethics Board of the Toronto General Hospital, University Health Network, and by the Institutional Review Board of the Pennsylvania State University College of Medicine. Informed consent was obtained from each patient in adherence to the principles set forth in the Helsinki declaration.
Donor and recipient clinical information
Donor data were collected with regard to age, smoking history, gender, last pO2/FiO2 prior to procurement and duration of cold ischemic preservation. All donors received 2 g methylprednisolone.
Recipient data were collected with regard to the development of primary graft dysfunction (PGD) in the first 3 posttransplant days; development of CLAD was determined by the permanent drop of the FEV1 by more than 20% from their posttransplant baseline and survival. The time from transplant to development of CLAD and death was monitored.
Biological samples from lung allograft
Biopsies of the allografts were collected prior to implantation, immediately snap-frozen in liquid nitrogen and stored at −80°C. Total RNA and DNA were extracted from lung tissue with an RNA/DNA simultaneous extraction kit (Qiagen, Mississauga, Canada).
Lung allograft surfactant protein mRNA expression
SP mRNA expression was measured in a blinded fashion by quantitative real-time RT-PCR. The primers used to amplify and quantitate the SP-A, SP-D, SP-B and SP-C mRNA were designed by using Primer3 website (http://frodo.wi.mit.edu/cgi-bin/primer3/primer3_www.cgi) (primers and real–time PCR methods are described in the Online Supplemental Method Section). The expression level of the SP mRNA was normalized to the level of 18S ribosomal RNA.
Lung allograft SP-A gene variants
The SP-A1 and SP-A2 gene polymorphisms were assessed in a blinded fashion. The SP-A1 and SP-A2 single nucleotide polymorphism (SNPs) assessment was done using a pyrosequencing protocol. The PCR–based RFLP genotype method for SP-A provided the basis for the pyrosequencing protocol, which is a primer–based DNA sequencing method. Pyrograms are scored by pattern–recognition software that compares the predicted SNP pattern (histogram) to the observed pattern (pyrogram) (Pyrosequencing AB, Uppsala, Sweden). Scoring of SP-A1 (6A, 6A2–20) and SP-A2 (1A, 1A0–13) gene variants was done as previously described [18, 19].
Lung allograft posttransplant samples
Bronchoscopies with trans–bronchial biopsies (TBBx) and BALF collection are routinely performed at 2 and 6 weeks following transplantation, every 3 months for the first year, every 6 months for the second year and thereafter as clinically indicated. Aliquots of the BALF were collected and snap-frozen at −80°C. After thawing, protease inhibitors (Complete Mini Tabs, Boehringer-Mannheim, Germany) were added to the samples that were then clarified by centrifugation at 5000g for 10 min. The resulting supernatant was assayed for SP-A and SP-D using ELISA (Test/Kokusai-F, Sysmex Corporation, Kobe, Japan), IL-8 (BioSource, International, Inc., CytoSet kit, Camarillo, CA), IL-2 and IL-12 (Lincoplex, Linco Research, Inc., St. Louis, MO).
The TBBx were assessed for acute rejection and graded according to the ISHLT Lung Rejection Study Group revised criteria . The samples were evaluated for the presence or absence of any grade of rejection and also for signs of inflammation other than acute rejection. Results were expressed as the ratio between number of diagnoses over number of pathologic assessments.
We summarized data as median and the 25th to 75th percentile range and as frequency and percentage. We used Kruskall–Wallis tests to compare continuous variables between categories. When the Kruskall–Wallis test p-value was less than 0.05, we used exact Wilcoxon rank-sum tests to compare continuous variables between categories with a Bonferroni correction to preserve an overall alpha of 0.05. We analyzed categorical data using Chi-squared and Fisher's exact tests. A receiver operating curve (ROC) method was used to identify best cutoff value of mRNA expression for poor clinical outcomes (identified as CLAD development or death within 24 months after transplantation). For time-to-event analyses, we examined mRNA expression as a binary exposure variable (dichotomized at the cutoff value). Stratified Cox proportional hazards models were constructed to examine associations between mRNA expression and time to events of interest with stratification for recipient diagnosis and with adjustment for donor variables that could affect SP gene expression and that are known to affect outcomes after transplantation. Kaplan–Meier survival curves were compared using log–rank tests. We used linear mixed models to compare SP-A and SP-D protein and cytokine levels in BALF between groups. Since SP-A, SP-D and cytokines were measured in BALF collected multiple times after transplantation, we specified an autoregressive covariance structure and included a covariate for time since transplantation to account for clustering of observations within individual study participants. Differences were considered significant when the p-value was less than 0.05. Statistical analysis was performed using SAS 9.2 software (SAS Institute, Inc., Cary, NC).
Thirty of the 42 patients died within the follow-up period, 5 within 30 days posttransplantation and 25 with median survival of 27 months (range 13.5–63). The 12 patients alive have median follow-up of 124 months (124–127.5). CLAD was diagnosed in 26 patients; 18 died with survival of 27 months (14–58); 8 are alive with median follow-up of 124 months (121–128). Four patients are alive and free of CLAD with median follow-up of 122 months (121–125). Twenty–four patients who developed CLAD or died within 24 months, median 5.5 months (1.5–14.5), after transplantation were identified as the poor outcome group.
Lung allograft SP mRNA expression
Patients with 30–day mortality and those with a poor clinical outcome (CLAD/death within the first 24 months) had significantly lower levels of SP-A mRNA (Table 1). The median SP-A mRNA/18s rRNA for all patients was 0.55 (0.25–1.2). No difference was noted for the other SP-(B, C, D) mRNA expressions.
|30–Day mortality||Clinical outcome|
|SP-A||0.2 (0.14–0.29)||0.7 (0.38–1.35)||0.015||0.3 (0.2–0.9)||0.9 (0.5–1.5)||0.016|
|SP-B||0.4 (0.3–1.75)||0.7 (0.4–1.4)||0.8||0.7 (0.4–1.3)||0.8 (0.4–1.5)||0.7|
|SP-C||1.1 (0.7–2.5)||2.2 (1.2–3.2)||0.3||1.9 (1.1–2.9)||2.2 (1.3–3.6)||0.5|
|SP-D||1.1 (0.5–2.2)||2 (0.9–3.8)||0.16||1.9 (0.8–2.7)||2.1 (0.9–3.8)||0.3|
Table 2 shows the SP-A mRNA expression levels and clinical outcomes. This uses a SPA mRNA cutoff value of 0.3 (low SP-A mRNA ≤0.3, normal >0.3) for best accuracy for the diagnosis of poor clinical outcome (identified as CLAD development or death within the first 24 months after transplantation). This cutoff value was identified utilizing an ROC method. The matrix shows a 50% sensitivity, 94% specificity, 92% positive predictive value, 59% negative predictive value, Likelihood Ratio Positive test = 9 and Likelihood Ratio Negative test = 0.5.
|Low SP-A||12 (29%)||1 (2%)||13 (31%)|
|Normal SP-A||12 (29%)||17 (40%)||29 (69%)|
|Total||24 (58%)||18 (42%)||42 (100%)|
The donor and recipient characteristics grouped according to the allograft levels of SP-A mRNA expression are shown in Table 3. Table S1 shows the characteristics of all patients transplanted in the 2002–2003 period.
|Age||48 (36–55)||57 (38–66)||0.13|
|Male||19 (70%)||6 (46%)||0.1|
|Smokers||15 (58%)||4 (31%)||0.1|
|Last PaO2/FiO2||420 (393–445)||438 (409–550)||0.2|
|Cold ischemic time||201 (163–243)||229 (190–302)||0.3|
|Age||52 (42–62)||46 (38–55)||0.4|
|Male||18 (62%)||5 (38%)||0.16|
|Bilateral||23 (79%)||10 (77%)||0.4|
|COPD||14 (48%)||1 (7.8%)|
|ILD||7 (24%)||6 (46%)|
|CF||5 (18%)||3 (23%)|
|Other||3 (10%)||3 (23%)|
|CMV mismatch||17 (59%)||7 (54%)||0.8|
Ad hoc competing risk survival analysis for CLAD and death showed that low SP-A mRNA in donor lungs was associated with a rate of death with HR of 3.14 (CI 1.02–9.66) (p = 0.046) and a rate of CLAD with HR 0.86 (CI 0.35–2.12) (p = 0.7). Table 4 shows the survival models for time to death, by levels of SP-A mRNA stratified for recipient diagnosis unadjusted, and adjusted for donor age, donor smoking history and duration of cold ischemic preservation. An eightfold increased rate of death is demonstrated for low levels of SP-A mRNA in the donor lung. Ad hoc bootstrap analysis with 2000 replications testing the model without the stratification for diagnosis showed HR 5 for death (95% CI 1.6–15.6, p = 0.005). Table 5 shows the Cox survival model adjusted also for CLAD as a time-varying covariate. Table 6 shows the cause of death according to the donor lung SP-A mRNA levels. Table 7 shows the association between SP-A mRNA levels and PGD.
|Nonadjusted survival models|
|Low SP-A mRNA||5.8||3.5–9.4||0.0003|
|Survival model stratified for recipient diagnosis and adjusted for donor variables|
|Low SP-A mRNA||8.5||2.03–35.16||0.003|
|Cold ischemic time||1.0||0.99–1.0||0.5|
|Low SP-A mRNA||5.9||1.4–25.7||0.018|
|Cold ischemic time||1.0||0.99–1.01||0.9|
|CLAD||7 (44%)||4 (30%)||11|
|Lung cancer||0||1 (8%)||1|
|Aneurysm rupture||1 (6%)||0||1|
|Cardiogenic shock—MOF||0||1 (8%)||1|
|Total||16 (100%)||13 (100%)||29|
|PGD 0–1||PGD 2–3|
|Low SP-A mRNA||4 (31%)||9 (69%)||13 (100%)|
|Normal SP-A mRNA||17 (59%)||12 (41%)||29 (100%)|
|Total no. of patients||21 (50%)||21 (50%)||42 (100%)|
Recipients of allografts with low levels of SP-A mRNA expression prior to implantation showed, in the BALF collected posttransplantation (Tables 8 and 9), a trend toward lower levels of SP-A protein expression and significantly greater levels of IL-2 and IL-12 expression. Moreover, they had a significantly greater number of episodes of acute rejection per number of bronchoscopies with TBBx (Low SP-A mRNA 0.45, range 0.37–0.62, vs. Normal SP-A mRNA 0.2, range 0.03–0.33; p = 0.003). No difference was noted with respect to positive BALF cultures, or alveolar neutrophilia.
|Log difference||95% CI||p-value|
|SP-A (ng/mL)||880 (488–1935)||1610 (1044–2742)||0.1|
|SP-D (ng/mL||627.7 (162–994)||621.8 (347–1010)||0.6|
|IL-2 (pg/mL)||2.45 (2.1–3.2)||1.9 (1.25–2.3)||0.03|
|IL-12 (pg/mL)||3.6 (3.4–3.8)||3.2 (3–3.3)||<0.0001|
|IL-8 (pg/mL)||176.1 (59–261.2)||64.1 (29.7–118.7)||0.013|
Donor SP-A1 and SP-A2 genetic variants and allograft SP-A mRNA expression
Table 10 shows the frequency of SP-A1 and SP-A2 variants and the genotypes for SP-A1 and SP-A2 for all 42 patients. The most frequent variants for SP-A1 were 6A2 and 6A3, representing 87% of the total, with the remaining 13% of variants being 6A, 6A4 and 6A5. The most frequent SP-A1 genotypes are 6A6A2, 6A26A2 and 6A26A3, representing 79% of our population. For the SP-A2 gene, the most frequent variants were 1A0 and 1A1, representing 80% of the total and with the remaining 20% being distributed among the less frequently found variants 1A, and 1A2,3,5,8,9,10. The most common genotypes are 1A1A0, 1A01A0 and 1A11A0, representing 70% of our patients.
|Variant frequency||Genotype frequency||Variant frequency||Genotype frequency|
|6A2||63% (53)||6A–6A2||10% (4)||1A0||64% (54)||1A0–1A||10% (4)|
|6A3||24% (20)||6A–6A4||2% (1)||1A1||15% (13)||1A0–1A0||40% (17)|
|6A||6% (5)||6A2–6A2||40% (17)||1A||5% (4)||1A0–1A1||20% (8)|
|6A4||6% (5)||6A2–6A3||29% (12)||1A2||4% (3)||1A0–1A2||7% (3)|
|6A5||1% (1)||6A2–6A4||5% (2)||1A3||4% (3)||1A0–1A3||2% (1)|
|6A2–6A5||2% (1)||1A5||4% (3)||1A0–1A5||5% (2)|
|6A3–6A3||7% (3)||1A8||1% (1)||1A0–1A9||5% (2)|
|6A3–6A4||5% (2)||1A9||2% (2)||1A1–1A3||5% (2)|
|1A10||1% (1)||1A1–1A5||2% (1)|
|Total pts||100% (42)||Total pts||100% (42)|
Table 11 shows the levels of SP-A mRNA expression and SP-A2 genotype with a significant association (χ2 = 10.5; p = 0.015). The less frequent genotypes (<10%, or fewer than 4pts) were clustered together in one group named “other.” No significant association was noted for SP-A1 genotype.
|Low SP-A mRNA||4 (100%)||3 (18%)||2 (25%)||4 (31%)||13|
|Normal SP-A mRNA||14 (82%)||6 (75%)||9 (69%)||31|
|Total no. of patients||4||17||7||14||42|
The SP-A mRNA expression according to the SP-A1 genotype groups was: for genotype 6A6A2, 0.14 (0.1–1.3); 6A26A2, 0.76 (0.5–2.1); 6A26A3, 0.44 (0.3–0.9); other, 0.39 (0.2–1.2), (Kruskal–Wallis, p = 0.18). Figure 3 shows the SP-A mRNA expression according to SP-A2 genotype: for genotype 1AA0 0.1 (0.05–0.14); 1A01A0, 0.78 (0.5–2.4); 1A01A1, 0.44 (0.3–1.1); “other,” 0.5 (0.3–1.2) (Kruskal–Wallis, p = 0.0075; Wilcoxon and Bonferroni 1A1A0 vs. 1A01A0, p = 0.014; and vs. “other”, p = 0.04).
We report a pilot study examining the role of SP-A gene expression in clinical outcomes of lung transplantation. Allograft SP-A mRNA expression measured prior to implantation may serve as a predictor of early and late posttransplant survival. Low levels of SP-A mRNA measured in the lung allograft just prior to implantation appears to be a negative predictor with respect to posttransplant survival. Donor lungs with low levels of SP-A mRNA expression prior to implantation showed a trend to lower levels of SP-A and significantly greater levels of IL-2 and IL-12 measured in the BALF collected at routine posttransplant surveillance bronchoscopies. Moreover, allografts with low levels of SP-A mRNA had a greater incidence of subsequent biopsy–proven acute rejection episodes.
Our findings also suggest a role of the SP-A polymorphisms in the expression of SP-A in the context of lung transplantation, as we observed an association between donor lung SP-A mRNA expression levels and SP-A2 genotype. In particular SP-A2 genotype 1A1A0 had lower levels of SP-A mRNA expression in the allograft prior to implantation. Although explored, no signal was noted between the variants and clinical outcome, possibly due to lack of statistical power.
Liver, kidney and heart transplantation have a recipient and graft 5–year survival in excess of 70–80%, while lung and intestine transplantation 5–year survival rates linger around 50%. Interestingly, both of the latter organs are characterized by having continuous exposure to the environment. Both organs rely on a more active organ–specific innate immune process under normal situations for the first line of protection from environmental pathogens and noxious agents [21-24]. Interactions between innate and adaptive immune responses in these organs, in the context of transplantation, are likely major detrimental factors contributing to the increased graft dysfunction that is observed. Inflammatory activation of the innate immune system likely fuels the acquired immune system response and vice versa.
SPs are important components of the lung–specific innate immunity. In our study, the association with lung transplantation clinical outcomes was noted only for SP-A. SP-A is produced in the lung primarily by the type II pneumocytes, and belongs to the collectin super-family serving as an opsonin for bacteria, fungi and viruses. SP-A also participates in the phospholipid homeostasis and in the orchestration of lung immunity by regulating cytokine production from macrophages and neutrophils and by providing direct or indirect modulation of lymphocyte activity and proliferation [4, 5].
SP-A is a polymer of SP-A1 and SP-A2 molecules that have a significant genetic variability [19, 25]. A variation in response to glucocorticoids (one of the regulators of SP-A gene expression) has been observed between the SP-A1 and SP-A2 genes and between certain SP-A1 and SP-A2 variants [8, 26, 27]. This is of interest given that high dose of glucocorticoids are given to organ donors before procurement and to recipients after transplantation as part of standard immunosuppressive regimens. It is conceivable that SP-A genetic variability may lead to more complex interactions including a variable response to steroids and a variable pattern of regulation of the adaptive immune system [6-9]. A significant pharmacogenetic relationship between SP-A polymorphic genotype and methylprednisolone treatment on SP-A expression has been observed in a human donor lung micro–organ culture model .
The SP-A1 and the SP-A2 variant frequencies seen in our study patient population (Table 5) reflect the larger patient populations described in the literature [6, 11-13, 18]. The association of SP-A2 genotype 1A1A0 with lower levels of SP-A mRNA within the graft prior to implantation suggests a genotype–dependent regulation of the quantity of SP-A mRNA (and possibly protein) production by the donor lung in response to stress events (mechanical ventilation, brain death, organ flushing and recovery and cold ischemic preservation). There is likely a multifactorial mechanism of expression regulation related to the gene variants and possibly also related to cAMP, TTF and glucocorticoids as documented in the literature [26, 27, 29].
In our study, recipients that received allografts with low levels of SP-A mRNA prior to implantation had a significantly reduced survival. All the patients who died within 30 days of transplant had a significantly lower level of SP-A mRNA in the allograft. Moreover, recipients of lungs with low levels of SP-A mRNA expression had lower subsequent SP-A protein levels in routine surveillance BALF sampling and significantly greater levels of IL-2 and IL-12, along with a significantly greater incidence of acute rejection episodes. Of note SP-A-enriched surfactant treatment has been shown to improve early lung graft function in a rodent model .
In conclusion, this pilot study shows a relationship between lung allograft SP-A mRNA expression prior to implantation with postlung transplant survival. Moreover, a significant association is shown between the SP-A mRNA expression and the SP-A gene polymorphisms. The consistent information with regard to SP-A gene variants, mRNA expression and concordant subsequent SP-A protein findings over time support the association. However, we acknowledge that the limitations of this study are the small number of patients, the selection bias determined by the sample availability (see Table S1) and the possibility that mRNA levels could be confounded by other factors not considered in this study. Moreover, the mRNA cutoff value was identified by ROC analysis for poor clinical outcome. No single threshold value had both high sensitivity and specificity. The most specific and most sensitive threshold was chosen to perform the multivariable–adjusted survival analysis. Additional studies are needed to confirm our findings as well as to validate the mRNA cutoff value.
Lung transplantation may be considered the ultimate “test bench” to study the lung and its responses to stress. In fact, no other model provides the lung with a series of sequential acute and chronic stresses: brain death, resuscitation, mechanical ventilation, cold ischemic preservation, reperfusion injury, rejection, chronic immunosuppression, viral, bacterial and fungal infections, and chronic silent aspiration from gastro-esophageal reflux. The organ–specific innate immunity provided by surfactant and its associated proteins plays a significant role in the first-line defense mechanisms that the lung possesses in its protective strategy against various insults. Further study of the contribution of genetic variability in the lung allograft's innate immunity molecules may provide insights into the mechanism of injury occurring posttransplantation. In addition, studying specific deficiencies in the donor lung, as we have done here, opens the door to the development of targeted treatment or repair strategies to correct these deficiencies and to improve the short– and long–term performance of the lung graft.
The study is supported by the National Institute of Health, NHLBI, R21HL092478 and HL34788.
The authors of this manuscript have no conflicts of interest to disclose as described by the American Journal of Transplantation.