Pruebas de la aceleración de la edad biológica y del acortamiento de los telómeros en los supervivientes de COVID-19
Abstract
La infección por el SARS-CoV-2 determina el síndrome COVID-19, caracterizado, en los peores casos, por una grave dificultad respiratoria, fibrosis pulmonar y cardíaca, liberación de citoquinas inflamatorias e inmunosupresión. Esta afección ha provocado hasta ahora la muerte de alrededor del 2,15% del total de la población mundial infectada. Entre los supervivientes, la presencia del llamado síndrome persistente post-COVID-19 (PPCS) es un hallazgo común. En los supervivientes del COVID-19, el PPCS presenta uno o más síntomas: fatiga, disnea, pérdida de memoria, trastornos del sueño y dificultad de concentración. En este estudio, se analizó una cohorte de 117 supervivientes de COVID-19 (post-COVID-19) y 144 voluntarios no infectados (sin COVID-19) utilizando la pirosecuenciación de islas CpG definidas e identificadas previamente como adecuadas para la determinación de la edad biológica. Los resultados muestran un aumento consistente de la edad biológica en la población post-COVID-19, determinando una aceleración DeltaAge de 10,45 ± 7,29 años (+5,25 años por encima del rango de normalidad) en comparación con 3,68 ± 8,17 años para la población libre de COVID-19 (p < 0,0001). Un acortamiento significativo de los telómeros es paralelo a este hallazgo en la cohorte post-COVID-19 en comparación con los sujetos libres de COVID-19 (p < 0,0001). Además, la expresión de la ECA2 se redujo en los pacientes post-COVID-19, en comparación con la población libre de COVID-19, mientras que la DPP-4 no cambió. A la luz de estas observaciones, planteamos la hipótesis de que algunas alteraciones epigenéticas están asociadas con la condición post-COVID-19, particularmente en pacientes más jóvenes (< 60 años).
1. Introduction
Tras la infección por SARS-CoV-2, el nivel de expresión de la enzima convertidora de angiotensina 2 (ACE2) en el sistema vascular tiende a disminuir [7]. Esta enzima está implicada en la regulación del sistema renina-angiotensina (SRA); la ECA2 contrasta con la actividad de la enzima convertidora de angiotensina (ECA) relacionada, convirtiendo la angiotensina II en angiotensina [1,2,3,4,5,6,7,8,9]. Una baja expresión de la ECA2 provoca una acumulación de angiotensina II, lo que puede agravar las condiciones que conducen a la dificultad respiratoria, la hipertensión, la arritmia, la hipertrofia cardíaca, el fallo de la función ventricular izquierda, la aterosclerosis y los aneurismas aórticos [9,10]. Además, la ECA2 se correlaciona negativamente con el envejecimiento; es relativamente abundante en personas jóvenes y sanas con un riesgo significativamente menor de ECV, mientras que se observa una menor cantidad en los ancianos [11].
La dipeptidil-peptidasa IV (DPP-4) es el receptor del coronavirus MERS (MERS-CoV) y se ha informado de que, en algunos casos, funciona como coreceptor del SARS-CoV-2 [12]. La expresión de la DPP-4 aumenta en la superficie de las células senescentes [13], y su forma transmembrana puede escindir muchas moléculas como quimiocinas, neuropéptidos y hormonas incretinas. Los inhibidores de la DPP-4 se han utilizado para tratar la DMT2, la isquemia cardíaca y la disfunción sistólica [14,15]. Algunas pruebas indican que los inhibidores de la DPP-4 podrían inhibir la entrada del coronavirus en las vías respiratorias, lo que sugiere un enfoque terapéutico adicional para el tratamiento de la COVID-19 [16]. No está claro si el nivel de ECA2 y DPP-4 en la sangre periférica puede representar biomarcadores valiosos para supervisar la recuperación de la COVID-19 o la aparición del PPCS.
En los últimos años, varios estudios han tratado de identificar marcadores biológicos o moleculares de envejecimiento que se correlacionan con la edad cronológica y que, por tanto, podrían ser útiles para estimar la edad biológica frente a la cronológica [22]. Algunos de estos parámetros se han definido a partir de modificaciones del metiloma del ADN que se correlacionan con la edad cronológica y podrían utilizarse en modelos de predicción de la edad para definir molecularmente la edad biológica: el llamado DNAmAge [23]. Muchos de estos estudios se centraron en individuos sanos o enfermos y en problemas forenses o de salud pública [24,25]. Se han desarrollado varios métodos para estimar la variación de los niveles de metilación en CpGs de ADN seleccionados. Estos enfoques se aplican a la determinación de la DNAmAge y se han utilizado para destacar la diferencia con la edad cronológica: la llamada DeltaAge. Algunos métodos se basan en la evaluación de muchos CpGs, explorados utilizando un array de todo el genoma o tecnologías de secuenciación de nueva generación [26]. Sin embargo, otros métodos se han desarrollado teniendo en cuenta el reducido número de CpGs analizados por pirosecuenciación [23,27,28]. Todos los sistemas se basan en los valores de metilación del ADN obtenidos a partir de muestras de sangre completa debido a su practicidad. Estos métodos simplificados tienen la ventaja adicional de ser rápidos y adecuados a la mayoría de los entornos de laboratorio sin requerir bioinformática [20,29,30]. Entre algunos de estos métodos "reduccionistas", el algoritmo propuesto por Bekaert B. et al. obtuvo buenos resultados para la predicción de la edad biológica en sujetos jóvenes y mayores [20]. Este algoritmo considera que un resultado de predicción es correcto para individuos de 60 años o más cuando la edad predicha coincide con la edad cronológica dentro de un rango de ± 5,2 años [20,31]. Teniendo en cuenta que la mayoría de los sujetos postCOVID-19 se encuentran en el grupo de edad de 50 a 60 años o más, este método se consideró adecuado para el presente estudio [20].
Un DeltaAge positivo se considera una aceleración del reloj biológico sanguíneo, mientras que un DeltaAge negativo indica una bioedad sanguínea más joven que la cronológica. Este parámetro ha demostrado ser útil para evaluar el riesgo de aparición de enfermedades cardiovasculares y neurodegenerativas, el cáncer y la aparición de la muerte por todas las causas [32].
En las enfermedades infecciosas, la aplicación de estos métodos es todavía limitada. Sin embargo, se ha observado una aceleración del DeltaAge en personas infectadas por el virus de la inmunodeficiencia humana (VIH), el citomegalovirus o bacterias como el Helicobacter pylori [32]. En el tejido cerebral post-mortem, el DNAmAge de los individuos crónicamente seropositivos era mayor que el de los controles negativos. Curiosamente, recientemente se ha observado una reversión parcial de la aceleración del DNAmAge tras la terapia antirretroviral [33,34]. La infección por VIH aumenta el riesgo de desarrollar enfermedades relacionadas con la edad, como los trastornos neurocognitivos [35]. Del mismo modo, en las personas infectadas por citomegalovirus, los análisis de metilación del ADN realizados en los leucocitos circulantes revelaron un aumento del DeltaAge [36]. Las consecuencias a largo plazo de estas alteraciones epigenómicas aún están por determinar.
El presente estudio investiga si, en los supervivientes del COVID-19, existe una alteración del DNAmAge y una aceleración del DeltaAge que, en asociación con otros parámetros moleculares como la longitud de los telómeros y la expresión de ACE2 en sangre periférica, podrían tipificar un conjunto de biomarcadores valiosos en otros y futuros estudios que exploren el riesgo de manifestaciones fisiopatológicas asociadas al PPCS.
2. Results
2.1. Evaluation of DNAmAge and DeltaAge in COVID-19 Survivors
A cohort of 117 COVID-19 survivors came to the attention of our physicians (the clinical features of volunteers are reported in Table 1). Results indicate that the y-axis intercept differs significantly between the COVID-19-free (Figure 1A) and the post-COVID-19 (Figure 1B) populations. The post-COVID-19 group intercepted the y-axis at value 35.22, while the COVID-19-free group intercepted the y-axis at 17.76. This difference determined an increment of DNAmAge of approximately nine years in the post-COVID-19 group, compared with the same group’s chronological age (p-value < 0.0001). No difference was appreciable in the controls (Table 2). Accordingly, the vast majority (76.6%) of the post-COVID-19 group had an average DeltaAge acceleration of 10.45 years (Figure 2, red dots). Considering that this method has a tolerance of about ± 5.2 years [20,31], the corrected average accelerated DeltaAge for this group was 5.25. On the other hand, the COVID-19-free volunteers together had a DeltaAge of 3.68, falling well within the range of normality [20] (Figure 2, blue squares). The post-COVID-19/COVID-19-free DeltaAge ratio was 2.84 (Table 2). Interestingly, the DeltaAge distribution within the two groups showed that the COVID-19-free samples were evenly distributed between the normal (39.9%) and the accelerated ranges (48.9%), while the remaining 12.8% had a decelerated biological clock. By contrast, 76.6% of the post-COVID-19 cohort had an accelerated DeltaAge, with only 23.4% falling within the normal or decelerated ranges (Figure 3A). Interestingly, while the COVID-19-free DeltaAge was distributed evenly among the different age groups, the increase of DeltaAge in the post-COVID-19 population was well represented among the younger people (age 56 ± 12.8 years; p-value < 0.0001, Figure 3A,B). The older individuals, in both COVID-19-free and COVID-19-survivors groups, did not show signs of DeltaAge acceleration. Interestingly, no differences were noticed between females and males in each age group. This result indicates that the younger survivors might be more sensitive to the SARS-CoV-2-dependent remodeling of the epigenome landscape (Figure 3B).
Biological age determination in COVID-19-free (blue squares) and post-COVID-19 (red dots) groups. (A) Linear regression of COVID-19-free volunteers’ DNAmAge. (B) Linear regression of DNAmAge in the post-COVID-19 subjects. In both graphs, the black dashed line is the bisector and represents the perfect correlation between chronological and biological age. The post-COVID-19 group (right panel) showed a statistically significant DNAmAge acceleration; p < 0.0001 (two-sided T-test).
DeltaAge distribution between COVID-free volunteers (left; blue squares) and post-COVID-19 survivors (right; red dots). The black dashed lines indicate the ± 5 years limit of the normal range according to the method. **** p-value of < 0.0001 (two-sided T-test).
(A) DeltaAge range of distribution within each age group. Specifically, in the COVID-19-free cohort (blue bars), 12.8% of the participants were decelerated (mean: −8.7 ± 5); 39.0% fell within the normal range, while 48.2% presented an accelerated DeltaAge (mean: + 5.15 ± 4.34). Interestingly, only a negligible portion (0.9%) of post-COVID-19 patients (red bars) were in the decelerated range. While 22.5% were within the normal range, the vast majority (76.6%) bore an accelerated bioclock (mean: + 8.7 ± 5.79). The average chronological age is reported above each bin. DeltaAge mean values were considered after subtraction of the ± 5.2 normality range distribution. (B) The graph shows DeltaAge distribution according to the different chronological age groups. A significance of **** p < 0.0001 between the younger COVID-19-free group (< 60) and the corresponding post-COVID-19 patients is shown (two-sided T-test).
2.2. Telomere Length Quantification
TL shortening has been reported as a risk factor for developing more severe COVID-19 syndrome [21]. We investigated this parameter, which is also associated with the progression of the aging process [19]. Comparing COVID-19-free and post-COVID-19 individuals revealed the presence of a significant shortness of chromosome ends in the COVID-19 survivors’ group (p-value < 0.0001; Figure 4A). Specifically, in the COVID-19-free (Figure 4A; blue squares) volunteers, TL was 3.5-fold longer than in the post-COVID-19 group (red dots). The correlation between DeltaAge distribution and TL indicates that post-COVID-19 survivors (Figure 4C; red dots), compared with the COVID-free group (Figure 4B; blue squares), have shorter telomeres (p-value < 0.0001) independent of an accelerated DeltaAge, suggesting that these two parameters might be regulated independently. Again, no significant differences emerged between females and males.
2.3. Peripheral Blood Expression of ACE2 and DPP-4
In a cell infected by SARS-Cov-2, ACE2 expression decreases, but little is known about the intensity of this biomarker in post-COVID-19 survivors. We evaluated the mRNA level of ACE2 (SARS-CoV and SARS-CoV-2 receptor) and DPP-4 (MERS-CoV receptor). The results are shown in Figure 5A,B. In the post-COVID-19 population, at the time point in which the blood samples were taken, which was more than four weeks from the end of the viral infection (see Table 1), ACE2 expression was significantly reduced (Figure 5A) (p-value < 0.0001). The expression level of DPP-4 was unchanged (Figure 5B).
3. Discussion
The global vaccination program against SARS-CoV-2 is actively ongoing, and the incidence of COVID-19 will soon decrease reasonably worldwide. Nevertheless, among the millions of COVID-19 survivors, many will require long-term assistance due to increased post-COVID-19 clinical sequelae defined as PPCS [37,38]. Despite the several manifestations associated with PPCS, there is a lack of potentially valuable molecular biomarkers for the monitoring of PPCS onset and evolution. In this study, we took advantage of the prior indication that biological age, defined as DNAmAge, could be altered in the presence of viral or bacterial infections [33,36,39], and the fact that shorter telomeres are associated with the risk of developing worse COVID-19 symptoms [21]. In this light, we found that a consistently accelerated DeltaAge (5.22 years above the normal range) characterized the post-COVID-19 population, and particularly those chronologically under 60 years (Figure 3A,B). This observation was paralleled by a significant telomere shortening (Figure 4). Although the two parameters seem independent (Figure 4B,C), both alterations coexisted in the post-COVID-19 population. All analyses were performed on blood with a minimally invasive procedure to obtain a source of genetic material exposed to critical environmental changes and associated with the “bona fide” health state of an individual [26,31,40].
However, much remains unknown about the effect of biological age on pulmonary and epithelial health following SARS-CoV-2 infection due to the lack of an appropriate algorithm and the invasive procedure that patients must undergo. The pathophysiology at the basis of these findings remains unclear; however, they may reflect a modified epigenetic environment, particularly evident among the younger COVID-19 survivors (Figure 3). The progression of aging is associated with critical metabolic changes. Some of these changes occur at the level of metabolites regulating the function of essential epigenetic enzymes, such as the decrease in NAD+ levels, the cofactor of sirtuins [41], and the reduction in alpha-ketoglutaric acid [42], the cofactor for all dioxygenases [43]. Although very speculative, it may be that older adults are relatively less sensitive to SARS-CoV-2-dependent epigenetic changes due to changes in their metabolic landscape. Additional experiments are necessary to elucidate this relevant aspect. In light of this consideration, a further question could be whether epigenetic changes might exist antecedent to the first viral contact, persisting or perhaps worsening progressively up to the post-COVID-19 period.
Several epigenetic phenomena have been associated with the SARS-CoV-2 infection [44], including the epigenetic regulation of ACE2 and IL-6. The latter has been associated with the development of worse COVID-19 symptoms due to excessive inflammation [45]. In addition, SARS-CoV-2 has been found to induce changes in DNA methylation, which affect the expression of immune response inhibitory genes that could, in part, contribute to the unfavorable progression of COVID-19 [46]. Finally, it is noteworthy that a recently identified signature made of 44 variably methylated CpGs has been predictive of subjects at risk of developing worse symptoms after SARS-CoV-2 infection [47]. Interestingly, none of these newly identified CpGs overlap with those involved in the DNAmAge prediction used in this [20] or other studies [26]. Hypothetically, it might be possible that distinct signals are regulating the structure of the epigenome regions determining a higher risk of developing a worse COVID-19 syndrome and those associated with DNAmAge prediction.
Even though epigenetics might provide clinically relevant information about COVID-19 [33] progression, no data is currently available regarding the involvement of epigenetic processes in the onset of the post-COVID-19 syndrome or PPCS. Although the post-COVID-19 cohort included in our study was heterogeneous, the range of symptoms observed during the infection varied from mild fever and smelling disturbance to a more severe condition that required assisted ventilation. Our evidence indicates changes in the methylation level of some CpGs associated with biological age calculation. This observation might reflect a more extensive phenomenon underlining unprecedented changes in the epigenome associated with the SARS-CoV-2 infection. A long-term follow-up of patients with an accelerated DeltaAge might help to clarify this critical point.
Telomere length is a marker of aging: progressive telomere shortening is a well-characterized phenomenon observed in older adults and attributed to the so-called telomere attrition. This condition is worsened by the absence of telomerase activity which is physiologically silenced in the early post-natal stage and throughout adulthood [19]. An accelerated TL shortening is a parameter associated with an increased risk of developing cardiovascular diseases and other disorders [48]. In COVID-19, patients bearing shorter telomeres in their peripheral leukocytes have been proposed to be at risk of worse prognoses [49]. In the post-COVID-19 group analyzed here, the average TL was 3.03 ± 2.39 kb, compared with 10.67 ± 11.69 kb in the control group (p < 0.0001). As shown in Table 2, the chronological ages of the two cohorts were approximately comparable. Hence, it is unlikely that the aging process was a determinant eliciting the difference. Accordingly, our results suggest that the observed TL shortening could be independent of DeltaAge (Figure 4B,C), indicating that the SARS-CoV-2 infection might directly contribute to telomere erosion in the blood cellular component.
ACE2 is a crucial component of the SARS-CoV-2 infection process. SARS-CoV-2 uses the ACE2 receptor to invade human alveolar epithelial cells and other cells, including cardiac fibroblasts [50]. In infected individuals, ACE2 is often down-regulated due to the infection [7,45]. The enzyme is expressed in several tissues, including alveolar lung cells, gastrointestinal tissue, vascular cells, and the brain; however, it is relatively under-represented in circulating blood cells. In all cases studied, the total relative ACE2 mRNA level in the peripheral blood of non-COVID-19 or post-COVID-19 subjects was significantly lower than that of the MERS-CoV receptor DDP4. However, in the post-COVID-19 group, ACE2 mRNA expression was reduced significantly compared with controls, while DPP-4 demonstrated similar expression levels in both groups. Interestingly, the accelerated DeltaAge, predominant in the younger Post-COVID-19 survivors, significantly correlated with a lower ACE2 mRNA level, suggesting an adverse effect of DNAmAge on ACE2 density in peripheral blood (Figure 5B,C).
The two groups considered in this study were not significantly different in terms of age, sex, and known clinical conditions before SARS-CoV-2 infection, except for a relatively higher incidence of BMI > 30 (15.3% vs. 9%) in the post-COVID-19 population compared with controls, as well as a record of more frequent lung diseases (20.2% vs. 1.6%; see Table 1). The origin of the persistent reduction in ACE2 expression in the post-COVID-19 group remains unsolved, and a longitudinal study should be performed monitoring this parameter.
4. Materials and Methods
Upon approval by the Ethical Committee and informed consent signing, peripheral blood was collected in EDTA vacutainers. A group of 144 age- and sex-matched COVID-19-free volunteers with some risk factors partially overlapping with the post-COVID-19 patients were recruited among the hospital workers and non-COVID-19 patients (see Table 1). Genomic DNA was extracted from the whole blood by a robotized station, as described below. After bisulfite conversion and PCR amplification, pyrosequencing was performed. DNAmAge calculations were completed according to Bekaert et al. [20].
The samples were classified into two groups: COVID-19-free (n = 144), a heterogeneous group that included healthy, cardiovascular disease-affected, and obstructive sleep apnea-affected patients, and the post-COVID-19 group, which included all of the previous types of patients who had also been infected with SARS-CoV-2 (n = 117). The clinical features of both populations are summarized in Table 1.
4.1. DNA Extraction from Whole Blood
Blood samples collected in EDTA (200 μL) were used to perform the extraction using the QIAmp DNA Blood Mini Kit (QIAGEN, cat. 55106, Hilden, Germany) associated with the automated system QIACube (QIAGEN, cat. 9002160), according to the manufacturer’s instructions. Subsequently, 2 μL of DNA was quantified with QIAxpert (QIAGEN, cat. 9002340, Hilden, Germany).
4.2. Bisulfite Conversion
One microgram of DNA was converted using the EpiTect Fast DNA Bisulfite Conversion Kit (QIAGEN, cat. 59824) associated with the RotorGene 2plex HRM Platform (QIAGEN, cat. 9001560) and the QIACube automated system, following the manufacturer’s instructions. Subsequently, 2 μL of converted DNA was quantified with QIAxpert.
4.3. Polymerase Chain Reactions for Pyrosequencing
PCR reaction mixes were performed using the PyroMark PCR Kit (QIAGEN, cat. 978103), following the manufacturer’s instructions. The sequences of primer used are available in the Supplementary Materials.
4.4. Pyrosequencing
The amplicons were sequenced in order to check the level of methylation in each CpG site. PyroMark Q24 Advanced Reagents (QIAGEN, cat. 970902) were loaded in the PyroMark Q24 Cartridge (QIAGEN, cat. 979202), following the manufacturer’s instructions, and 5 μL of PCR product was added to the reaction mix containing: Pyromark Binding Buffer (supplied in PyroMark Q24 Advanced Reagents kit), Streptavidin Sepharose High Performance (GE Healthcare, cat. GE17-5113-01), and DNase/RNase-free distilled water. Samples were shaken at room temperature for 15 min at 1400 rpm. Subsequently, the samples underwent the PyroMark Q24 Vacuum Station (QIAGEN, cat. 9001515) procedure, in which the target sequences were purified and put into an annealing buffer containing the sequencing primer (0.375 μM). The sequences of oligos are available in the Supplementary Data. The plate containing the sequence to analyze and the primer was heated at 80 °C for 5 min. Finally, the PyroMark Q24 Advanced System (QIAGEN, cat. 9001514) was set to analyze the target sequences (available in the Supplementary Methods).
4.5. DNAmAge Estimation
Bekaert’s algorithm was applied to estimate the biological age of the population [20] as reported in Daunay et al. [31]:
4.6. Telomere Length Quantification
The length of chromosome ends was quantified using a PCR Real-Time of Absolute Human Telomere Length Quantification qPCR Assay Kit (ScienCell, cat. 8918, Carlsbad, CA), following the manufacturer’s instructions.
4.7. RNA Extraction
The total RNA was isolated from whole blood using a QIAmp RNA Blood Mini Kit (QIAGEN, cat. 52304) and an automatized extractor QIACube, according to the manufacturer’s instructions. The RNA was quantified with QIAxpert.
4.8. cDNA Synthesis and qPCR Real-Time
An Omniscript RT Kit (QIAGEN, cat. 205113) was used to convert total RNA into cDNA according to the manufacturer’s instructions.
Real-time qPCR was performed on the RotorGene 2plex HRM Platform using RT2 SYBR Green ROX FAST Mastermix (QIAGEN, cat. 330620). The sequences of primers are available in the Supplementary Data. To perform the amplification, the machine settings were:
Initial denaturation: 95 °C, 5 min;
Denaturation: 95 °C, 15 s;
Annealing: 60 °C, 30 s;
Elongation: 72 °C, 30 s;
Final elongation: 72 °C, 1 min.
Denaturation, annealing, and elongation were repeated 45 times.
4.9. Data Analysis
All data were analyzed with GraphPad Prism 8.4.3 and p-values were calculated using two-sided T-tests.
5. Conclusions
This study has many significant limitations, including the limited number of subjects investigated and the low number of CpGs considered. Although we used a valid forensic method to establish the biological age in the examined groups [20,31], adopting other methods which evaluate a large set of CpGs might be preferable [26,40]. However, the application of such procedures is undermined by the elevated cost and relative complexity and therefore may not be feasible at the laboratory level in many hospitals.
Nevertheless, it was shown here that individuals belonging to a group of COVID-19 survivors exhibited a significant acceleration of their biological age, occurring mainly in the younger individuals. This information was correlated with TL shortening and the expression of ACE2 mRNA. It is too early to extrapolate whether relevant clinical indications may arise from this and other studies assessing the role of epigenetic changes in the COVID-19 syndrome [46,47]. However, a warning might be raised that sequelae of SARS-CoV-2 infection might rely on persistent epigenomic modifications, possibly underlying the presence of a COVID-19 epigenetic memory. The epigenomic landscape of actual post-COVID-19 survivors and prospective COVID-19 survivors from SARS-CoV-2 variants should be considered to gain predictive prognostic insights and monitor more accurately a patient’s response to treatment.
Acknowledgments
The authors would like to thank Marta Lovagnini, Elena Robbi, Riccardo Sideri, and Lorena Grano De Oro for their valuable work and assistance in providing patients’ information, tracking, and blood sample collection.
Supplementary Materials
The following are available online at https://www.mdpi.com/article/10.3390/ijms22116151/s1.
Author Contributions
Conceptualization, A.F., C.G., T.B. and M.T.L.R.; methodology, A.M., V.B., M.G.Z. and S.A.; validation, C.G., A.F., A.P., F.M., M.M. and M.N.; formal analysis, A.M., C.G., M.T.L.R., T.B. and S.N.; investigation, A.M., V.B., M.G.Z., S.A. and L.F.; resources, M.N., V.B. and S.N.; data curation, O.C., M.B., L.A.D.V.; patient recruitment and clinical assessment, A.M.; writing—original draft preparation, A.M. and C.G.; writing—review and editing, A.F., T.B. and M.T.L.R.; supervision, C.G.; project administration, C.G.; funding acquisition, C.G. and M.T.L.R. All authors have read and agreed to the published version of the manuscript.
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8201243/
No hay comentarios:
Publicar un comentario