TYPE Original Research PUBLISHED 02 February 2023 DOI 10.3389/fmed.2023.1079317 Lymphangioleiomyomatosis: OPEN ACCESS Searching for potential biomarkers EDITED BY Jin Woo Song, Asan Medical Center, Republic of Korea Eva Revilla-López1,2 , Victoria Ruiz de Miguel3 , REVIEWED BY Manuel López-Meseguer1 , Cristina Berastegui1 , Elena Lesma, University of Milan, Italy Meritxell Boada-Pérez3 , Alberto Mendoza-Valderrey3 , Elzbieta Radzikowska, Marta Arjona-Peris1 , Marta Zapata-Ortega1 , Victor Monforte1,4 , National Institute of Tuberculosis and Lung Diseases, Poland Carlos Bravo1,4 , Antonio Roman1,4 , Susana Gómez-Ollés2,3,4* and *CORRESPONDENCE Berta Sáez-Giménez1,2,5 Susana Gómez-Ollés susana.go@vhir.org 1 Lung Transplant Program, Department of Pulmonology, Hospital Universitari Vall d’Hebron, Barcelona, Spain, 2 Department of Medicine, Universitat Autònoma de Barcelona, Barcelona, Spain, 3 Department of SPECIALTY SECTION Pulmonology, Vall d’Hebron Institut de Recerca, Barcelona, Spain, 4 CIBER de Enfermedades Respiratorias, This article was submitted to Instituto de Salud Carlos III, Madrid, Spain, 5 Department of Cell Biology, Physiology and Immunology, Pulmonary Medicine, Universitat Autònoma de Barcelona, Barcelona, Spain a section of the journal Frontiers in Medicine RECEIVED 25October 2022 ACCEPTED 16 January 2023 PUBLISHED 02 February 2023 Background: Vascular endothelial growth factor-D (VEGF-D) is the most commonly CITATION used biomarker for diagnosing lymphangioleiomyomatosis (LAM). However, lung Revilla-López E, Ruiz de Miguel V, López-Meseguer M, Berastegui C, biopsy is often necessary as well; therefore, defining new biomarkers for LAM is Boada-Pérez M, Mendoza-Valderrey A, crucial. The aim of this study was to describe the diagnostic accuracy of a variety Arjona-Peris M, Zapata-Ortega M, Monforte V, of biomarkers. Bravo C, Roman A, Gómez-Ollés S and Sáez-Giménez B (2023) Methods: We assessed 13 analytes in serum related to extracellular matrix Lymphangioleiomyomatosis: Searching for potential biomarkers. remodeling, lymphatic involvement and angiogenesis in a cohort of patients with Front. Med. 10:1079317. LAM, comparing them with patients with other cystic lung diseases (OCLD) and doi: 10.3389/fmed.2023.1079317 healthy women. A scoring method based on the cut-point of each VEGF-D and COPYRIGHT © 2023 Revilla-López, Ruiz de Miguel, metalloproteinase-2 (MMP-2) was used to evaluate the diagnostic performance of López-Meseguer, Berastegui, Boada-Pérez, the marker combination. Mendoza-Valderrey, Arjona-Peris, Zapata-Ortega, Monforte, Bravo, Roman, Results: A total of 97 subjects were recruited: 59 (61%) LAM patients, 18 Gómez-Ollés and Sáez-Giménez. This is an open-access article distributed under the terms (19%) OCLD patients, and 20 (20%) healthy female controls. MMP-2 was the of the Creative Commons Attribution License only extracellular matrix remodeling biomarker able to differentiate LAM patients (CC BY). The use, distribution or reproduction in from OCLD and healthy patients. Serum MMP-2 was higher in LAM patients other forums is permitted, provided the original author(s) and the copyright owner(s) are [median 578 (465–832) ng/ml] than in patients with OCLD and healthy controls credited and that the original publication in this [medians 360 (314–546) and 427 (365–513) ng/ml, respectively (p < 0.0001)]. The journal is cited, in accordance with accepted academic practice. No use, distribution or area under ROC curve (AUC) of MMP-2 was 0.785 and that of VEGF-D 0.815 reproduction is permitted which does not (p = 0.6214). The sensitivity/specificity profiles of each biomarker (54/92% for comply with these terms. MMP-2, 59/95% for VEGF-D) yielded a composite score (−6.36 + 0.0059 × VEGF- D + 0.0069 × MMP-2) with higher accuracy than each component alone (AUC 0.88 and sensitivity/specificity 79/87%). Conclusion: Combining MMP-2 and VEGF-D may increase diagnostic accuracy for LAM. KEYWORDS lymphangioleiomyomatosis, biomarkers, serum, metalloproteinases, VEGF-D, diagnose Frontiers in Medicine 01 frontiersin.org Revilla-López et al. 10.3389/fmed.2023.1079317 1. Introduction LAM patients compared to patients with other cystic diseases and healthy volunteers. Lymphangioleiomyomatosis (LAM) is a rare multisystem disease that predominantly affects women and is characterized by the proliferation of abnormal smooth muscle-like cells 2. Materials and methods named LAM cells (1). LAM is secondary to mutations in the tuberous sclerosis complex (TSC) genes, mainly TSC2, which 2.1. Subjects and clinical data cause the activation of the mammalian target of rapamycin (mTOR) complex (2). LAM cells proliferate in the lungs A cross-sectional study was conducted from July 2019 to and the axial lymphatic system, but also invade the airways, June 2020. The study population comprised LAM patients with pulmonary artery, diaphragm, aorta, and retroperitoneal fat tissue or without lung transplantation (LT), patients with OCLD and (3, 4). healthy females. The cohort of OCLD included Langerhans cell Serum levels of vascular endothelial growth factor-D (VEGF- histiocytosis, emphysema, and Birt–Hogg–Dubé syndrome. LAM D), a ligand for the lymphatic growth-factor receptor VEGFR- patients were diagnosed by lung biopsy or by the presence of 3/Flt-4, are higher in most LAM patients than in healthy controls lung cysts plus: renal angiomyolipomas, tuberous sclerosis, or or in patients with other cystic lung diseases (OCLD) (5, 6). chylous effusions. It has been suggested that a VEGF-D level of > 800 pg/ml Clinical findings, diagnostic methods, sirolimus treatment, and the presence of pulmonary cysts is diagnostic for LAM (7). childbirth, menopausal status, lung function, LT, presence of tuberous However, a VEGF-D level of < 800 pg/ml does not rule out sclerosis, and extrapulmonary involvement were retrospectively the diagnosis (6). A wide range of VEGF-D levels has been recorded from electronic clinical records. described (8); they are higher in patients with lymphatic disease Cystic lung involvement was graded according to thoracic CT who develop chylous pleural effusion or ascites than in those scans in LAM patients and in OCLD patients. Lungs were divided without lymphatic involvement, which may explain these findings into three equal zones (upper, middle, and lower). Based on the lung (9), although the variability may also be due to the differences in volume replaced by cysts within each zone, cyst involvement was the assay kits used and in the access to standardized tests (10). graded individually and subjectively as follows: mild disease, less than 30%; moderate disease, cysts involve 30–60%; severe disease, cysts Even though the diagnostic value of VEGF-D is well established, involve greater than 60%, as previously reported (23). All CT scans its clinical utility to monitor follow-up and treatment response were retrospectively reviewed by the same operator. is not clear (11). Therefore, the search for new LAM biomarkers Analysis of sex hormones such as estradiol, luteinizing hormone, continues. follitropin, progesterone, and prolactin, as well as pulmonary In addition to lymphatic involvement, extracellular matrix function tests (PFTs) were performed at the time of inclusion in the remodeling, and angiogenesis in LAM (12), other biomarkers may study. PFT followed the American Thoracic Society (ATS)/European be related to these biological processes. Matrix metalloproteinases Respiratory Society protocols (24). (MMP) are a family of endopeptidases involved not only in Sirolimus was indicated in LAM patients in the case of extracellular matrix degradation but also in cell proliferation, FEV1 < 80%, an angiomyolipoma larger than 4 cm in diameter differentiation, cell signaling, and migration (13). Associations or multiple angiomyolipomas. In LAM patients who underwent between MMP and destructive lesions caused by LAM have been LT, sirolimus was prescribed in the case of chronic lung allograft demonstrated (14) and elevated MMP-2, MMP-7, and MMP-9 dysfunction (CLAD) or angiomyolipomas. Sirolimus was given at levels have been observed in serum of LAM patients (15–17). a dose between 1 and 4 mg once daily adjusted to obtain target However, these studies only compared LAM patients with healthy trough blood levels between 5 and 15 ng/ml measured with liquid controls. chromatography-mass spectrometry. Angiopoietin-1, endostatin, fibroblast growth factor-acidic The Ethics Committee of the Vall d’Hebron University Hospital (FGF-acidic), phosphatidylinositol-glycan biosynthesis class F (Barcelona, Spain ARB-SIR-2018-01) approved the study and written protein (PIGF), thrombospondin-2 and platelet derived growth informed consent was obtained from all subjects. factor-AA protein (PDGF-AA) are biomarkers associated with angiogenesis. Little information is available on these biomarkers in LAM; only one study has reported higher levels of endostatin 2.2. Blood collection and storage with LAM associated with TSC than in patients with sporadic LAM (S-LAM) and healthy volunteers (18). Von Willebrand Blood samples were collected in one BD Vacutainer Plastic R factor (vWF) is a protein associated with angiogenesis (19) SST II Advance Tube, one BD Vacutainer Citrate Tube, and one R and is reduced in LAM patients compared to healthy controls BD Vacutainer spray-coated K2EDTA tube. The serum tube was R (20). Cancer antigen-125 (CA-125) and Krebs von den Lungen allowed to clot at room temperature for 60 min prior to centrifugation 6 (KL-6) are also related to LAM: CA-125 levels have been at 1,300 g at 4◦ C for 10 min. Serum was obtained and stored at associated with pleural effusions and reduced pulmonary −80◦ C until used. The EDTA tube was immediately centrifuged function (21) and KL-6 levels are higher than in other lung after the blood extraction for 15 min at 1,000 g at 4◦ C. Plasma was diseases (22). aliquoted into 1 ml aliquots, and then an additional centrifugation Thus, the objective of the present study was to describe the was performed at 10,000 g at 4◦ C for 10 min to obtain platelet-free diagnostic accuracy of a range of biomarkers in a cohort of plasma. All plasma samples were stored at −70◦ C until used. Frontiers in Medicine 02 frontiersin.org Revilla-López et al. 10.3389/fmed.2023.1079317 2.3. MMP-2, MMP-9, endostatin, FGF disease, and 22 (50%) LAM patients had severe disease. As for acidic, PIGF, thrombospondin-2, VEGF-D, OCLD patients: 4 (22%) had mild disease, 6 (33%) had moderate disease, and 8 (44%) had severe disease. Dyspnea was the most VEGF- R3, angiopoietin-1, PDGF-AA, vWF, common initial symptom, recorded in 24 (41%) LAM patients, KL-6, and CA-125 assays followed by pneumothorax in 21 (36%), chylothorax in 5 (9%), hemoptysis in 1 (2%), and hematuria also in 1 (2%). A total of 7 Serum MMP-2 and MMP-9 were measured using Luminex (12%) diagnoses were incidental in asymptomatic patients. A total Human Magnetic Assay LXSAHM-02 (R&D Systems, Minneapolis, of 28 (47%) patients had renal angiomyolipomas, 12 (20%) of those MN, USA). vWF activity assay was performed on ACLTOP 700 LAS were bilateral and median size was 25 (14–42) mm × 24 (13– using reagents from the same supplier (IL Werfen, Barcelona, Spain). 25) mm. A total of 12 (20%) patients had lymphangioleiomyomas. Serum CA-125 was measured by chemiluminescence assay using During follow-up, 24 (41%) LAM patients presented recurrent Atellica automated analyzer with reagents from the same supplier R pneumothorax and 5 (9%) developed recurrent chylothorax. No (Siemens Healthcare Diagnostics, Erlangen, Germany). Endostatin, multifocal micronodular pneumocyte hyperplasia was observed on FGF acidic, PIGF, thrombospondin-2, angiopoietin-1, and PDGF-AA any thoracic CT. Concerning LT, 11 (73%) of them were bilateral were measured using Luminex Performance Human Angiogenesis and 4 (27%) were single. A total of 11 (73%) LT patients had Magnetic Panel A (6-Plex), R&D Systems. VEFD-R3 was measured extrapulmonary lesions: 8 (53%) had angiomyolipomas and 3 (20%) using Luminex Human Magnetic Assay (1-Plex) LXSAHM-01, had lymphangioleiomyomas. Only one patient had relapse of LAM R&D Systems. VEGF-D was measured by a human VEGF-D after LT and 7 (47%) of them developed CLAD. Quantikine ELISA kit, R&D Systems. KL-6 was measured using a R A total of 41 (69%) LAM patients were treated with sirolimus: chemiluminescence enzyme immunoassay (CLEIA) by Lumipulse G 33 (80%) non-LT patients and 8 (20%) LT patients, with a mean KL-6 IRC (Fujirebio, Tokyo, Japan). time under sirolimus treatment of 6.9 ± 2.7 and 7.5 ± 2.7 years, respectively. Mean sirolimus dose was 2 ± 0.7 mg/day and mean blood trough levels were 7.6 ± 4.3 ng/ml. Side effects were recorded in 2.4. Statistical analysis 11 (27%) patients and tended to occur early after starting treatment. The most common side effect was the appearance of oral ulcers, Continuous variables are expressed as means ± standard recorded in 6 (15%) patients, severe increases in serum cholesterol deviation (SD) and as medians and interquartile range (IQR). in 2 (5%), liver enzyme abnormalities in 1 (2%), leg swelling in 1 Categorical variables are expressed as numbers of cases and (2%), and acne in 1 (2%). A total of 3 (5%) LAM patients started proportions. Normality was evaluated in continuous variables using sirolimus treatment after the collection of the blood sample for the the Shapiro–Wilk test and comparisons of groups were made using study. Only 3 (30%) TSC-LAM patients took antiepileptic drugs and analysis of variance (ANOVA). Categorical variables were compared 15 (34%) LAM patients were on bronchodilators, with no differences using the chi-square test or Fisher’s exact test, as appropriate. For between TSC-LAM and S-LAM (p = 0.7). The proportion of patients comparisons between patients the Kruskal–Wallis test was used, under other treatments were also comparable between the two groups as appropriate. A receiver operating characteristic (ROC) curve (Supplemental Table 1). was created to detect the cut-off point to establish the optimal Table 1 shows the main demographic and clinical characteristics sensitivity and specificity of each biomarker. The previously reported of the study population. There were no differences between groups VEGF-D cut-off value of 800 pg/ml was also analysed. A logistic in post-menopause status, pregnancy, or body mass index, but regression model including VEGF-D and MMP-2 was used to create differences were found in hypertension, hypercholesterolemia, and a composite score. The DeLong test was used to compare ROC renal failure between the group of transplanted patients and the rest. curves. A p-value < 0.05 was considered statistically significant. SPSS In the PFTs, there were only differences in force vital capacity (FVC) 27.0 software version (IBM SPSS Statistics, Chicago, IL, USA) and as percentage of predicted value. GrapPhad Prism 9.1.2 (Graph Pad Software, San Diego, CA, USA) were used for the analysis. 3.1. Distribution of serum biomarkers between groups 3. Results Serum MMP-2 was highest in LAM patients, with a median of 578 A total of 97 subjects were recruited: 59 (61%) LAM patients, 18 (465–832) ng/ml compared with medians of 360 (314–546) ng/ml for (19%) OCLD patients, and 20 (20%) healthy female controls. Median OCLD and 427 (365–513) ng/ml for healthy controls (p < 0.0001) time from LAM diagnosis to blood sample collection was 9 (2– (Figure 2). MMP-2 was able to differentiate between LAM and 17) years. A total of 15 (25%) of the LAM patients had undergone LT OCLD patients (p = 0.0004). Differences in MMP-2 levels between prior to recruitment; median time from LT to blood sample collection S-LAM and TSC-LAM were non-significant: median values were was 12 (8–13) years. A total of 10 (17%) patients had LAM associated 573 (289–1,151) and 502 (404–688) ng/ml, respectively (p = 0.181). with TSC (TSC-LAM): 2 of them (20%) underwent LT and 6 (75%) There were no differences in the LAM group regarding MMP-2 were under sirolimus treatment. A total of 9 (50%) of the OCLD levels and LT: LAM patients had a median of 551 (462–783) ng/ml patients had emphysema, 8 (44%) had Langerhans cell histiocytosis compared with one of 630 (491–901) ng/ml in LAM LT patients and 1 (6%) had Birt–Hogg–Dubé syndrome (Figure 1). The method (p > 0.999). There were no significant differences either in the of LAM diagnosis was lung biopsy in 46 (78%) patients, and clinical LAM group regarding MMP-2 levels and sirolimus treatment: LAM and radiological findings in 13 (22%). with sirolimus treatment had a median of 578 (466–847) ng/ml, Regarding cystic lung involvement showed in CT scans: 8 (18%) while LAM without sirolimus treatment had a median of 480 (379– LAM patients had mild disease, 14 (32%) LAM patients had moderate 656) ng/ml (p = 0.081). The difference in MPP-2 levels between Frontiers in Medicine 03 frontiersin.org Revilla-López et al. 10.3389/fmed.2023.1079317 FIGURE 1 Subjects enrolled in the study. TABLE 1 Demographic and clinical characteristics of the study population. LAM (including LAM (no LT) LT (n = 15) OCLD (n = 18) Healthy p-value LT) n = 59 n = 44 controls (n = 20) Age (years), median (IQR) 52 (46–62) 51 (44–59) 61 (52–66) 57 (46–60) 51 (47–56) 0.037 Post-menopause [n (%)] 41 (69) 29 (66) 12 (80) 13 (72) 10 (50) 0.381 Pregnancy [n (%)] 37 (63) 26 (59) 11 (73) 14 (78) 10 (50) 0.392 Ex-smokers [n (%)] 21 (36) 17 (39) 4 (27) 11 (61) 3 (15) <0.001 Smokers [n (%)] 3 (5) 3 (7) 0 (0) 5 (28) 4 (20) BMI (kg/m2 ), mean ± SD 24 ± 4.8 23.6 ± 4 25.2 ± 6.4 24 ± 4.1 23 ± 3.6 0.9214 Hypertension [n (%)] 19 (32) 11 (25) 8 (53) 2 (11) 1 (5) <0.001 Hypercholesterolemia [n (%)] 30 (51) 19 (43) 11 (73) 0 (0) 6 (30) <0.001 Diabetes mellitus [n (%)] 7 (12) 4 (9) 3 (20) 1 (6) 0 (0) 0.320 Renal failure (GFR < 60) [n (%)] 13 (22) 1 (2) 12 (80) 0 (0) 0 (0) <0.001 PFT (mean ± SD) FVC (ml) 2,733 ± 847 2,927 ± 825 2,165 ± 647 2,370 ± 978 – 0.012 FVC (% pred.) 86 ± 21 91 ± 20 72 ± 18 72 ± 26 0.009 FEV1 (ml) 1,850 ± 698 1,922 ± 728 1,638 ± 573 1,629 ± 1,141 0.257 FEV1 (% pred.) 71 ± 22 72 ± 22 69 ± 22 59 ± 35 0.298 FEV1/FVC (%) 68 ± 14 65 ± 15 75 ± 9 61 ± 17 0.059 DLCO (%) 55 ± 25 54 ± 26 68 ± 20 48 ± 30 0.337 Supplemental oxygen [n (%)] 6 (10) 5 (11) 1 (6) 7 (39) – 0.012 LT, lung transplantation; OCLD, other cystic lung diseases; IQR, interquartile range; SD, standard deviation; BMI, body mass index; GFR, glomerular filtration rate (ml/min/1.73 m2 ); PFT, pulmonary function test. healthy controls and LAM patients under sirolimus remained under sirolimus treatment had a median VEGF-D of 610 (486– significant (p = 0.003). Also, the levels of MMP-2 were comparable 886) pg/ml compared to 658 (452–1,638) pg/ml in those without among patients with and without bronchodilators: median value was treatment (p = 0.457). The difference in VEGF-D levels between 625 (404–843) and 543 (465–738) ng/ml, respectively (p = 0.98). healthy controls and LAM patients under sirolimus remained Serum VEGF-D level was higher in LAM with a median of 629 significant (p = 0.004). Patients without bronchodilators showed (482–924) pg/ml, than in OCLD [median 459 (325–530) pg/ml], and higher VEGF-D levels [median 709 (521–893) pg/ml] compared to healthy controls [median 404 (352–456) pg/ml] (Figure 2). Levels those under treatment [median 458 (400–548) pg/ml] (p = 0.004). of VEGF-D were higher in TSC-LAM patients [median 733 (523– Of note, VEGF-D was persistently elevated in LAM patients after 928) pg/ml] compared with S-LAM [median 554 (457–790) pg/ml] LT though the difference with regard to those without LT was not (p = 0.039). Considering all LAM patients irrespective of LT, those significant (Figure 3). Median VEGF-D was 577 (461–861) pg/ml Frontiers in Medicine 04 frontiersin.org Revilla-López et al. 10.3389/fmed.2023.1079317 AUC for this composite score was 0.88 (95% CI 0.82–0.95) (p < 0.05 compared to either VEGF-D or MMP-2 alone), and the best cut-off for the composite score was 0.61, resulting in a sensitivity/specificity of 79/86% (Figure 4). We did not find any correlation between either MMP- 2 or VEGF-D values and age, estradiol, luteinizing hormone, follitropin, progesterone, or prolactin levels. Nor was there any correlation between the two biomarkers. MMP-2 and VEGF- D were not associated with chylothorax or angiomyolipomas or with recurrencies; VEGF-D correlated with lymphatic involvement, FIGURE 2 but MMP-2 did not. Comparison of serum metalloproteinase-2 (MMP-2) and vascular There was a nearly significant positive correlation between endothelial growth factor-D (VEGF-D) levels between the levels of serum MMP-2 and the severity of the cystic lung lymphangioleiomyomatosis (LAM) patients, other cystic lung diseases (OCLD), and healthy controls. Solid lines indicate median and involvement by CT: median MMP-2 was 458 (305–536) ng/ml interquartile range (IQR). in the mild disease group, 500 (348–783) ng/ml in the moderate disease group, and 551 (456–782) ng/ml in the severe disease group (p = 0.087). On the other hand, levels of serum VEGF-D were not in LT-free LAM vs. 893 (589–1,482) pg/ml in the LAM LT group correlated: median values were 574 (423–919), 503 (405–646), and (p = 0.384). 544 (444–816) pg/ml, respectively, for the three groups (p = 0.638). Endostatin, PIGF, thrombospondin-2, VEGF-R3, angiopoietin- We observed a tendency toward an association between higher 1, PDGF-AA, MMP-9, CA-125, vWF, FGF-acidic, and KL-6 levels of serum MMP-2 and worse lung function parameters in LAM biomarkers were not able to discriminate between LAM patients and patients who had not undergone LT. In patients with FEV1% pred. OCLD patients or healthy volunteers (Table 2 and Supplementary below the median, mean MMP-2 was 674 ± 203 ng/ml, while in Figures 1, 2). those above median FEV1% pred. it was 562 ± 226 ng/ml (p = 0.092), The area under ROC curve (AUC) of VEGF-D to predict LAM (Figure 5). VEGF-D did not correlate with lung function parameters. diagnosis, compared with OCLD and healthy volunteers, was 0.815 [95% confidence interval (CI): 0.732–0.899, p < 0.0001]. A VEGF- D cut-off level of 578 pg/ml showed a sensitivity and specificity 4. Discussion of 59 and 95%, respectively, whereas a cut-off value of 800 pg/ml had a sensitivity of 36 and a specificity of 97%, respectively. The Finding new diagnostic biomarkers for LAM is a pressing issue. AUC of MMP-2 was 0.785 (95% CI: 0.695–0.876, p < 0.0001). Our results show that serum levels of MMP-2 are significantly higher A cut-off level of 560 pg/ml showed a sensitivity of 54% for LAM in patients with LAM compared to patients with OCLD and healthy diagnosis and a specificity of 92%. Specifically, a cut-off value of volunteers. A composite score combining VEGF-D and MMP-2 can 824 pg/ml had a specificity of 100% to differentiate LAM from OCLD be used as a diagnostic biomarker for LAM and can increase the patients: 15 (25%) LAM patients fulfilled this criterion, 5 of them accuracy of the diagnosis. received a LT and 11 were under sirolimus treatment. A composite Associated with lymphangiogenesis and lung remodeling, MMPs marker was therefore constructed combining MMP-2 and VEGF- are components of the extracellular matrix that promote collagen and D (score = −6.36 + 0.0059 × VEGF-D + 0.0069 × MMP-2). The elastin degradation (13). MMPs are overexpressed in a number of FIGURE 3 Comparison of serum metalloproteinase-2 (MMP-2) and vascular endothelial growth factor-D (VEGF-D) levels between lymphangioleiomyomatosis (LAM) patients, LAM lung transplantation (LT) patients, other cystic lung diseases (OCLD), and healthy controls. Solid lines indicate median and interquartile range (IQR). Frontiers in Medicine 05 frontiersin.org Revilla-López et al. 10.3389/fmed.2023.1079317 TABLE 2 Biomarker levels from LAM patients, QQLT LAM patients, OCLD and healthy controls. LAM patients LT LAM (n = 15) OCLD (n = 18) Healthy controls p-value (n = 44) (n = 20) Angiopoietin-1 (pg/ml), median (IQR) 7,178 (5,033–13,450) 9,405 (5,911–11,662) 7,603 (4,726–12,581) 3,766 (2,998–6,604) 0.0035 CA-125 (U/ml), median (IQR) 15 (1.7–25.2) 18.5 (12.6–45.3) 8.2 (6.3–14.5) 10.4 (7.5–12.9) 0.0118 Endostatin (ng/ml), median (IQR) 37 (28–51) 72 (55–80) 35 (26–45) 31 (23–39) <0.0001 FGF acidic (pg/ml), median (IQR) 16 (12.4–19.2) 16 (11.4–19.2) 16.3 (11.34–18.9) 11.6 (10.3–20.8) 0.6219 KL-6 (U/ml), median (IQR) 342 (256–517) 424 (341–535) 415 (290–636) 336.5 (237–393) 0.1526 MMP-9 (ng/ml), median (IQR) 150 (93–209) 208 (90–283) 242 (163–374) 194 (102–229) 0.0246 PDGF-AA (pg/ml), median (IQR) 1,020 (768–1,697) 889 (647–1,236) 985.5 (702–1,497) 433.5 (282.8–665.8) <0.0001 PIGF (pg/ml), median (IQR) 3.8 (3.2–4.3) 4.5 (4.2–4.9) 3.9 (3–4.6) 2.4 (1.8–3.3) <0.0001 Thrombospondin-2 (ng/ml), median (IQR) 27 (21–36) 31 (24–44) 24.5 (16–38) 18.5 (15–25) 0.0027 VEGF-R3 (pg/ml), median (IQR) 182 (116.3–227) 165.1 (61.8–193.1) 168.9 (109.2–204.5) 86.94 (75.70–137.3) 0.0037 vWF (%), median (IQR) 143.9 (121.6–175.6) 157.3 (133.5–180.8) 97.15 (78–139.1) 124.2 (96.75–149.8) 0.0030 CA-125, cancer antigen-125; FGF-acidic, fibroblast growth factor-acidic; KL-6, Krebs von den Lungen 6; LAM, lymphangioleiomyomatosis; LT, lung transplantation; MMP, matrix metalloproteinases; OCLD, other cystic lung diseases; PDGF-AA, platelet derived growth factor-AA protein, PIGF, phosphatidylinositol-glycan biosynthesis class F protein; TSC, tuberous sclerosis complex; VEGF-D, vascular endothelial growth factor-D; vWF, von Willebrand factor. FIGURE 4 (A) Receiver operating characteristic (ROC) curves for metalloproteinase-2 (MMP-2) (green line) and vascular endothelial growth factor-D (VEGF-D) (blue line). (B) ROC curves of MMP-2 (green line) and the combination of MMP-2 and VEGF-D (red line). (C) ROC curves of VEGF-D (blue line) and the combination of MMP-2 and VEGF-D (red line). chronic pulmonary diseases, including idiopathic pulmonary fibrosis larger LAM cells stained more strongly for MMP-2, MMP-9, (25, 26), asthma (27), and emphysema (28), as well as tumor and Human Melanoma Black-45 monoclonal antibody. According proliferation (29). Hayashi et al. (30) demonstrated increased activity to other authors, the disproportion between MMPs and their of MMPs, particularly MMP-2, which might be responsible for the tissue inhibitors may also lead to the cystic destruction of lung cystic destruction of the lung parenchyma in LAM. Furthermore, parenchyma (31). these authors observed that the immunohistochemical heterogeneity Few studies have been published on MMPs and LAM. Chang of LAM cells was associated with different patterns; for example, et al. (17) found significantly higher MMP-2 serum levels among Frontiers in Medicine 06 frontiersin.org Revilla-López et al. 10.3389/fmed.2023.1079317 receiving sirolimus might have milder disease. In addition, the levels of pre-treatment VEGF-D are not available due to the design of the study. According to Taveira-DaSilva et al. (11) pre-sirolimus VEGF- D decreased especially in patients with lymphatic involvement, but they were unable to demonstrate an association between VEGF-D decline and lung function changes following sirolimus treatment. As a biomarker for monitoring the progression of LAM disease, VEGF-D has several limitations, a circumstance that highlights the importance of further research in this area. Another relevant finding from our study is that VEGF-D remained high in LAM patients after LT. This result supports the FIGURE 5 hypothesis that LAM cells do not originate exclusively in the lungs, Classification of lymphangioleiomyomatosis (LAM) patients based on median FEV1% pred. higher or lower than 73.5%. and that they may also proliferate in other tissues (36). Recently, Guo et al. (37) identified the presence of LAM cells in the uterus through single-cell RNA sequencing. To our knowledge, ours is the first study to measure VEGF-D after LT in LAM patients. LAM patients, but they only compared them with healthy volunteers The main limitation of this study is its cross-sectional design. and excluded patients taking sirolimus and LT. Those authors Also, the sample size was relatively small but this was due to the rarity hypothesized that MMP-2 may decrease in advanced disease, in view of the disease and is a drawback for most LAM studies. The strengths of an association they found between MMP-2 levels and better lung of the study include the homogeneity of the clinical care and criteria, function. In contrast, we found that MMP-2 was associated with as all patients were treated by the same team. poorer lung function. Considering that a higher number of cysts, and In conclusion, a composite two-biomarker assay combining therefore more lung destruction, impairs lung function, an increase MMP-2 and VEGF-D enhances diagnostic accuracy. However, in MMP-2 seems plausible in patients with poorer outcomes. In fact, further longitudinal studies with larger patient populations are we found a trend for an association between serum MMP-2 levels and needed to confirm the usefulness of MMP-2 in clinical practice. The the degree of lung involvement by CT. study also showed that serum levels of VEGF-D remained elevated in A recent study explored the role of MMP-2 and MMP-7 as LAM patients even after LT. biomarkers for S-LAM and TSC-LAM in comparison to healthy controls (16). In both S-LAM and TSC-LAM patients, serum MMP- 2 levels were higher than in controls, while MMP-seven levels were higher in TSC-LAM patients than in controls and S-LAM Data availability statement patients. However, higher levels of those biomarkers did not appear The raw data supporting the conclusions of this article will be to correlate with worse lung function parameters. In the same made available by the authors, without undue reservation. study (16). MMP-2 did not have the same ability to diagnose LAM disease as VEGF-D, in agreement with our data; however, in that study the authors did not attempt to combine biomarkers in order to increase diagnostic accuracy. A Japanese study used semi- Ethics statement quantitative analyses to assess serum activity of MMP-9 and MMP-2 in LAM patients and compared them with healthy females (15). The studies involving human participants were reviewed and MMP-9 serum activity was significantly higher than in controls, but approved by the Ethics Committee of the Vall d’Hebron University MMP-2 activity did not differ between groups. Due to the semi- Hospital. The patients/participants provided their written informed quantitative nature of the method used for the analysis, the results consent to participate in this study. of that study may differ from ours. In addition, some authors have investigated the role of doxycycline, an MMP inhibitor, on LAM progression. Pimenta Author contributions et al. (32) observed a decrease in MMP-2 levels after doxycycline treatment but this was not associated with functional improvement. ER-L was responsible for collecting blood samples, data On the other hand, Chang et al. (33) did not find differences in analysis, statistics, and writing of the manuscript. VR and AM-V MMP-two levels after 2 years of doxycycline. Altogether, further were responsible for sample analysis and critically reviewed the studies should clarify the prognostic role of MMP-2 in LAM. We manuscript. ML-M, CBe, MB-P, MA-P, MZ-O, VR, CBr, and AR found VEGF-D levels to be significantly higher in LAM patients, critically reviewed the manuscript. SG-O and BS-G were responsible with similar sensitivity and specificity values to those recorded in for study design and critical review. All authors contributed to the previous studies (6, 34). Young et al. (35) discovered significant article and approved the submitted version. correlations between measurements of VEGF-D at LAM diagnosis and lung function parameters. In contrast, we did not detect any association between lung function and serum VEGF-D levels, but Funding our measurements were not made at baseline. Along the same lines, Young et al. (6) did not find any correlation in their cross- This project was supported by the Spanish Society of sectional study. Even though we observed no differences in VEGF-D Pulmonology and Thoracic Surgery (SEPAR), grant number: PI levels between patients who received sirolimus and those who did 638/2018. The funders have no role in study design, data and analysis not, this finding could be explained by the fact that patients not collection, decision to publish, or preparation of the manuscript. Frontiers in Medicine 07 frontiersin.org Revilla-López et al. 10.3389/fmed.2023.1079317 Acknowledgments Publisher’s note We thank the individuals who provided their blood samples, and All claims expressed in this article are solely those of the the AELAM foundation for its continued support for LAM research. authors and do not necessarily represent those of their affiliated The logistic regression model has been carried out in the Statistics organizations, or those of the publisher, the editors and the reviewers. and Bioinformatics Unit (UEB) Vall d’Hebron Hospital Research Any product that may be evaluated in this article, or claim that may Institute (VHIR). be made by its manufacturer, is not guaranteed or endorsed by the publisher. Conflict of interest Supplementary material The authors declare that the research was conducted in the The Supplementary Material for this article can be found online absence of any commercial or financial relationships that could be at: https://www.frontiersin.org/articles/10.3389/fmed.2023.1079317/ construed as a potential conflict of interest. full#supplementary-material References 1. Johnson S, Taveira-Dasilva A, Moss J. Lymphangioleiomyomatosis. Clin Chest Med. 18. Lamattina A, Poli S, Kidambi P, Bagwe S, Courtwright A, Louis P, et al. Serum (2016) 37:389–403. endostatin levels are associated with diffusion capacity and with tuberous sclerosis- associated lymphangioleiomyomatosis. Orphanet J Rare Dis. (2019) 14:72. doi: 10.1186/ 2. Carsillo T, Astrinidis A, Henske E. Mutations in the tuberous sclerosis complex gene s13023-019-1050-4 TSC2 are a cause of sporadic pulmonary lymphangioleiomyomatosis. Proc Natl Acad Sci USA. (2000) 97:6085–90. 19. Lenting P, Casari C, Christophe O, Denis CV. Von willebrand factor: the old, the new and the unknown. J Thromb Haemost. (2012) 10:2428–37. 3. Kumasaka T, Seyama K, Mitani K, Sato T, Souma S, Kondo T, et al. Lymphangiogenesis in Lymphangioleiomyomatosis. Am J Surg Pathol. (2004) 28:1007–16. 20. Banville N, Burgess J, Jaffar J, Tjin G, Richeldi L, Cerri S, et al. A quantitative 4. Seyama K, Kumasaka T, Kurihara M, Mitani K, Sato T. Lymphangioleiomyomatosis: proteomic approach to identify significantly altered protein networks in the serum of a disease involving the lymphatic system. Lymphat Res Biol. (2010) 8:21–31. patients with lymphangioleiomyomatosis (LAM). PLoS One. (2014) 9:e105365. doi: 10. 1371/journal.pone.0105365 5. Young L, Inoue Y, McCormack F. Diagnostic potential of serum VEGF-D for lymphangioleiomyomatosis. N Engl J Med. (2008) 358:199–200. 21. Glasgow C, Pacheco-Rodriguez G, Steagall W, Haughey M, Julien- Williams P, Stylianou M, et al. CA-125 in disease progression and treatment of 6. Young L, VanDyke R, Gulleman P, Inoue Y, Brown K, Schmidt L, et al. Serum vascular lymphangioleiomyomatosis. Chest. (2018) 153:339–48. doi: 10.1016/j.chest.2017.05.018 endothelial growth factor-D prospectively distinguishes lymphangioleiomyomatosis from other diseases. Chest. (2010) 138:674–81. doi: 10.1378/chest.10-0573 22. D’Alessandro, M, Bergantini L, Cameli P, Perillo F, Remediani L, Refini R, et al. Prognostic role of KL-6 in lymphangioleiomyomatosis patients. Minerva Med. (2021) 7. Gupta N, Finlay G, Kotloff R, Strange C, Wilson K, Young L, et al. 113:727–29. doi: 10.23736/S0026-4806.21.07322-5 Lymphangioleiomyomatosis diagnosis and management: high-resolution chest computed tomography, transbronchial lung biopsy, and pleural disease management an Official 23. Avila N, Dwyer A, Rabel A, Moss J. Sporadic lymphangioleiomyomatosis and American thoracic society/Japanese respiratory society clinical practice guideline. Am J tuberous sclerosis complex with lymphangioleiomyomatosis: comparison of CT features. Respir Crit Care Med. (2017) 196:1337–48. doi: 10.1164/rccm.201709-1965ST Radiology. (2007) 242:277–85. doi: 10.1148/radiol.2421051767 8. Amaral A, de Oliveira M, Dias O, Arimura F, Freitas C, Acencio M, et al. 24. Graham B, Steenbruggen I, Barjaktarevic I, Cooper B, Hall G, Hallstrand T, et al. Concentration of serum vascular endothelial growth factor (VEGF-D) and its correlation Standardization of spirometry 2019 update an official American thoracic society and with functional and clinical parameters in patients with lymphangioleiomyomatosis from European respiratory society technical statement. Am J Respir Crit Care Med. (2019) a brazilian reference center. Lung. (2019) 197:139–46. doi: 10.1007/s00408-018-00191-3 200:E70–88. doi: 10.1164/rccm.201908-1590ST 9. Glasgow C, Avila N, Lin J, Stylianou M, Moss J. Serum vascular endothelial growth 25. Maher T, Oballa E, Simpson J, Porte J, Habgood A, Fahy W, et al. An epithelial factor-D levels in patients with lymphangioleiomyomatosis reflect lymphatic involvement. biomarker signature for idiopathic pulmonary fibrosis: an analysis from the multicentre Chest. (2009) 135:1293–300. doi: 10.1378/chest.08-1160 PROFILE cohort study. Lancet Respir Med. (2017) 5:946–55. doi: 10.1016/S2213-2600(17) 30430-7 10. McCarthy C, Gupta N, Johnson S, Yu J, McCormack F. Lymphangioleiomyomatosis: pathogenesis, clinical features, diagnosis, and management. Lancet Respir Med. (2021) 26. Todd J, Vinisko R, Liu Y, Neely M, Overton R, Flaherty K, et al. Circulating matrix 9:1313–27. metalloproteinases and tissue metalloproteinase inhibitors in patients with idiopathic pulmonary fibrosis in the multicenter IPF-PRO registry cohort. BMC Pulm Med. (2020) 11. Taveira-DaSilva A, Jones A, Julien-Williams P, Stylianou M, Moss J. Long-term 20:64. doi: 10.1186/s12890-020-1103-4 effect of sirolimus on serum vascular endothelial growth factor D levels in patients with lymphangioleiomyomatosis. Chest. (2018) 153:124–32. 27. Ko F, Diba C, Roth M, McKay K, Johnson P, Salome C, et al. A comparison of airway and serum matrix metalloproteinase-9 activity among normal subjects, asthmatic 12. Ji R. Lymphatic endothelial cells, lymphangiogenesis, and extracellular matrix. patients, and patients with asthmatic mucus hypersecretion. Chest. (2005) 127:1919–27. Lymphat Res Biol. (2006) 4:83–100. doi: 10.1378/chest.127.6.1919 13. Bonnans C, Chou J, Werb Z. Remodelling the extracellular matrix in development 28. Mao J, Tashkin D, Belloni P, Baileyhealy I, Baratelli F, Roth M. All-trans retinoic and disease. Nat Rev Mol Cell Biol. (2014) 15:786–801. acid modulates the balance of matrix metalloproteinase-9 and tissue inhibitor of 14. Matsui K, Takeda K, Yu Z, Travis W, Moss J, Ferrans V. Role for activation of matrix metalloproteinase-1 in patients with emphysema. Chest. (2003) 124:1724–32. doi: 10.1378/ metalloproteinases in the pathogenesis of pulmonary lymphangioleiomyomatosis. Arch chest.124.5.1724 Pathol Lab Med. (2000) 124:267–75. 29. Abdel-Hamid N, Abass S. Matrix metalloproteinase contribution in management of 15. Odajima N, Betsuyaku T, Nasuhara Y, Inoue H, Seyama K, Nishimura M. Matrix cancer proliferation, metastasis and drug targeting. Mol Biol Rep. (2021) 48:6525–38. metalloproteinases in blood from patients with LAM. Respir Med. (2009) 103:124–9. 30. Hayashi T, Fleming MV, Stetler-Stevenson W, Liotta L, Moss J, Ferrans V, et al. 16. Terraneo S, Lesma E, Ancona S, Imeri G, Palumbo G, Torre O, et al. Exploring the Immunohistochemical study of matrix metalloproteinases (MMPs) and their tissue role of matrix metalloproteinases as biomarkers in sporadic lymphangioleiomyomatosis inhibitors (TIMPs) in pulmonary lymphangioleiomyomatosis (LAM). Hum Pathol. (1997) and tuberous sclerosis complex. A pilot study. Front Med. (2021) 8:605909. doi: 10.3389/ 28:1071–8. doi: 10.1016/s0046-8177(97)90061-7 fmed.2021.605909 31. Nascimento E, Baldi B, Mariani A, Annoni R, Kairalla R, Pimenta 17. Chang W, Cane J, Blakey J, Kumaran M, Pointon K, Johnson S. Clinical utility of S, et al. Immunohistological features related to functional impairment in diagnostic guidelines and putative biomarkers in lymphangioleiomyomatosis. Respir Res. lymphangioleiomyomatosis. Respir Res. (2018) 19:83. doi: 10.1186/s12931-018- (2012) 13:34. doi: 10.1186/1465-9921-13-34 0797-9 Frontiers in Medicine 08 frontiersin.org Revilla-López et al. 10.3389/fmed.2023.1079317 32. Pimenta S, Baldi B, Kairalla R, Carvalho C. Doxycycline use in patients with and treatment response: a prospective analysis of the multicenter international lymphangioleiomyomatosis: biomarkers and pulmonary function response. J Bras lymphangioleiomyomatosis efficacy of sirolimus (MILES) trial. Lancet Respir Med. Pneumol. (2013) 39:5–15. doi: 10.1590/s1806-37132013000100002 (2013) 1:445–52. doi: 10.1016/S2213-2600(13)70090-0 33. Chang W, Cane J, Kumaran M, Lewis S, Tattersfield A, Johnson S, et al. A 2-year 36. Kumasaka T, Seyama K, Mitani K, Souma S, Kashiwagi S, Hebisawa A, et al. randomised placebo-controlled trial of doxycycline for lymphangioleiomyomatosis. Eur Lymphangiogenesis-mediated shedding of LAM cell clusters as a mechanism for Respir J. (2014) 43:1114–23. dissemination in lymphangioleiomyomatosis. Am J Surg Pathol. (2005) 29:1356–66. doi: 34. Seyama K, Kumasaka T, Souma S, Sato T, Kurihara M, Mitani K, et al. 10.1097/01.pas.0000172192.25295.45 Vascular endothelial growth factor-D is increased in serum of patients with 37. Guo M, Yu J, Perl A, Wikenheiser-Brokamp K, Riccetti M, Zhang E, et al. Single- lymphangioleiomyomatosis. Lymphat Res Biol. (2006) 4:143–52. cell transcriptomic analysis identifies a unique pulmonary lymphangioleiomyomatosis 35. Young L, Lee H, Inoue Y, Moss J, Singer L, Strange C, et al. Serum cell. Am J Respir Crit Care Med. (2020) 202:1373–87. doi: 10.1164/rccm.201912-2 VEGF-D concentration as a biomarker of lymphangioleiomyomatosis severity 445OC Frontiers in Medicine 09 frontiersin.org