Edited by: Maria Elisabetta Baldassarre, University of Bari Aldo Moro, Italy
Reviewed by: Uday Kishore, Brunel University London, United Kingdom; Ga Won Jeon, Inje University Busan Paik Hospital, South Korea; Anthony George Tsolaki, Brunel University London, United Kingdom
This article was submitted to Genetics of Common and Rare Diseases, a section of the journal Frontiers in Pediatrics
This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
Neonatal respiratory distress syndrome (RDS) is the most common cause of respiratory failure in premature infants due to surfactant deficiency (
Major risk factors, such as prematurity and low birth weight (BW) along with sex and race (
Furthermore, it has been noted that infants dying with RDS have low levels of surfactant proteins (SP) (
Multiple genetic variants and single nucleotide polymorphisms (SNP) of the surfactant protein gene (
However, the impact of SNP-SNP interactions on RDS susceptibility has not been addressed before. The synergistic (epistatic) interactions among genetic variants of the surfactant proteins may alter disease susceptibility (
In the present study, we studied intergenic and intragenic SNP-SNP interactions of the
This study used available genotype data and clinical information in the Floros biobank at Penn State University, College of Medicine. These were collected and processed under an approved protocol by the institutional review board from the human subject protection office of the Pennsylvania State University (PSU) College of Medicine as well as the institutional review board of the respective centers where samples were collected in other Institutions other than PSU, as described previously (
Clinical Characteristics of the cohort with and without RDS.
Gestational age (weeks): median (IQR) | 33 (31, 35) | 30 (26, 34) | <0.001 |
Sex: |
|||
Female | 236 (51) | 198 (41) | 0.02 |
Male | 220 (48) | 277 (58) | |
Race: |
|||
Non-Hispanic white | 328 (71) | 343 (72) | |
Non-Hispanic black | 64 (14) | 82 (17) | 0.09 |
Hispanic | 20 (4) | 25 (5) | |
Asian-pacific islander | 23 (5) | 13 (2) | |
Other/mixed parents | 22 (4) | 13 (2) | |
Infant birth weight (g) ± SD | 1,818 ± 515 | 1,474 ± 606 | <0.001 |
Preterm labor: |
|||
Absent | 64 (14) | 74 (15) | 0.36 |
Present | 203 (44) | 196 (41) | |
Maternal diabetes mellitus: |
|||
No | 419 (92) | 412 (94) | 0.27 |
Yes | 33 (7) | 21 (5) | |
Chorioamnionitis: |
|||
No | 161 (35) | 204 (43) | 0.26 |
Yes | 35 (8) | 33 (7) | |
Antenatal steroid: |
|||
No | 1 (0.6%) | 16 (3%) | 0.0003 |
Yes | 280 (61%) | 273 (57%) | |
Surfactant use: |
|||
No | 448 (97) | 167 (35) | <0.001 |
Yes | 8 (2) | 305 (64) | |
Chronic lung disease: |
|||
No | 297 (65) | 238 (50) | <0.001 |
Yes | 16 (4) | 92 (20) |
**
***
A total of 17 SNPs of the SP genes
Wang et al. (
For each type of data analysis, case-control genotype observations were sorted into a 2 × 2 contingency table to test each of the genetic effects described above. For example, consider a SNP with three genotypes AA, Aa, and aa. To estimate its dominant effect, the effect size was compared to that of the heterozygote Aa against the average size of each of the two homozygotes AA and aa in cases and controls, respectively. Based on the resulting 2 × 2 contingency table, the logistic regression model was implemented to estimate the dominant effect of this SNP, and the effects were adjusted for age and sex. The odds ratio (OR) was estimated to assess the magnitude of the dominant/additive effect.
To estimate the additive effect, the size was compared as below,
Odds of genotype for cases = number of cases with AA/number of cases with aa
Odds of genotype for controls = number of controls with AA/number of controls with aa
OR = odds for cases/odds for controls
=
For example-
OR = 1: Genotype difference is not associated with the disease;
OR > 1.0: Genotype AA is “more risky” (i.e., associated with higher risk for the disease than genotype aa)
OR < 1.0: Genotype aa is “more risky” for the disease than genotype AA
A similar procedure was applied to analyze all other genetic effects.
The significance of each effect was adjusted for multiple comparisons using the false discovery rate (FDR) controlled at 1%. Wang et al.'s simulation data indicate that a 100 × 100 sample size combination in an epistatic case-control model has a power of > 0.80 to detect significant associations in a 2 × 2 contingency table analysis (
The associations of single SNP and intergenic/intragenic two and three SNP interactions with RDS are shown in
Single SNP associated with RDS.
rs17886395 | Additive | 0.16 | 0.06–0.43 | 0.0006 | 0.03 |
The two SNP interactions associated with RDS susceptibility.
rs17886395 | rs721917 | d1d2 | 0.56 | 0.45–0.69 | 9.33E-08 | 9.77E-05 | ||
rs4253527 | d1d2 | 1.69 | 1.32–2.07 | 8.88E-06 | 0.003097 | |||
rs1136450 | rs4253527 | d1d2 | 1.77 | 1.42–2.19 | 3.08E-07 | 0.000161 | ||
d1a2 | 0.54 | 0.41–0.72 | 2.91E-05 | 0.004226 | ||||
rs1965708 | rs1059047 | d2 | 0.43 | 0.29–0.62 | 1.61E-05 | 0.004507 | ||
d1d2 | 1.69 | 1.32–2.17 | 2.85E-05 | 0.004507 | ||||
rs2077079 | rs4715 | a1d2 | 0.19 | 0.09–0.38 | 3.04E-05 | 0.004507 | ||
rs3024798 | a1d2 | 5.7 | 2.56–12.65 | 3.44E-05 | 0.004507 |
Three SNP-SNP-SNP interactions of surfactant protein genes associated with RDS.
rs1059046 | rs1965707 | rs1965708 | d1d2d3 | 0.55 | 0.46–0.65 | 7.74E-08 | |||
rs1965707 | rs1965708 | rs1136450 | d1d2d3 | 0.55 | 0.46–0.65 | 1.30E-07 | |||
d1d3 | 1.92 | 1.47–2.51 | 0.001018 | ||||||
rs1059046 | rs17886395 | rs1059047 | d1d2d3 | 0.57 | 0.47–0.69 | 3.54E-05 | |||
d1d2d3 | 0.59 | 0.49–0.72 | 0.000159 | ||||||
rs17886395 | rs1965707 | rs1136451 | d1d2d3 | 1.57 | 1.3–1.89 | 0.001033 | |||
rs1059046 | rs1136451 | rs1059057 | d1d2d3 | 0.54 | 0.44–0.65 | 8.12E-07 | |||
rs17886395 | rs1059047 | rs1059057 | a1d2d3 | 4.76 | 2.67–8.47 | 0.001024 | |||
rs17886395 | rs1136450 | rs1059057 | d1d2d3 | 0.57 | 0.47–0.69 | 0.000401 | |||
rs17886395 | rs1059047 | rs1136450 | d1d2d3 | 0.53 | 0.44–0.65 | 8.12E-07 | |||
rs1059046 | rs1136450 | rs4253527 | d1d2d3 | 1.53 | 1.28–1.81 | 0.001018 | |||
rs17886395 | rs1059047 | rs1136451 | d1d2d3 | 0.62 | 0.51–0.75 | 0.001235 | |||
rs1059047 | rs1136450 | rs1136451 | d1d2d3 | 0.53 | 0.43–0.64 | 2.82E-07 | |||
rs1136450 | rs1136451 | rs1059057 | d1d2d3 | 0.57 | 0.47–0.69 | 1.77E-05 | |||
rs1059047 | rs1136451 | rs1059057 | d1a2d3 | 4.09 | 2.39–7.00 | 0.0012 | |||
rs1059046 | rs721917 | rs7316 | d1d2 | 0.53 | 0.41–0.67 | 0.000197 | |||
d1d2a3 | 0.51 | 0.40–0.64 | 6.71E-05 | ||||||
rs1136450 | rs4253527 | rs7316 | d1d2 | 2.01 | 1.56–2.60 | 6.71E-05 | |||
d1d2a3 | 1.96 | 1.52–2.52 | 0.000196 | ||||||
rs1965708 | rs721917 | rs1130866 | d2d3 | 0.52 | 0.40–0.67 | 0.000362 | |||
rs1059046 | rs4253527 | rs721917 | d1d3 | 0.51 | 0.40–0.64 | 3.54E-05 | |||
d1a2d3 | 0.49 | 0.39–0.62 | 1.40E-05 | ||||||
rs1059046 | rs1136450 | rs721917 | d1d2d3 | 0.53 | 0.45–0.63 | 1.12E-08 | |||
rs17886395 | rs1136451 | rs721917 | d1d2d3 | 0.61 | 0.49–0.73 | 0.000467 | |||
rs1965708 | rs1136450 | rs2243639 | d1d2d3 | 1.62 | 1.34–1.95 | 0.000273 | |||
rs1965708 | rs1059057 | rs2077079 | d1d2d3 | 1.64 | 1.33–2.01 | 0.001275 | |||
rs1059046 | rs2077079 | rs1130866 | d1d2d3 | 0.67 | 0.57–0.79 | 0.001295 | |||
rs2077079 | rs3024798 | rs7316 | d1d2d3 | 0.63 | 0.52–0.76 | 0.001029 |
Out of the 17 SNPs of the five
Among the two SNP interactions, the only intragenic interaction included
There were five intragenic interactions associated with RDS. Three interactions were among SNPs of the
The two SNP interactions are shown in
The five
Intergenic three SNP interaction and RDS susceptibility. It shows the schematic presentation of
The
The two SNP interaction in RDS susceptibility. Associations between RDS and the two SNP-SNP interactions are shown. The star marks the
The three SNP interactions associated with RDS susceptibility. The figure depicts intergenic and intragenic interactions of
Although
Using the stringent criteria of FDR correction with 1% (
The rs17886395 (C/G) is located in the collagen-like domain of
We observed an association of the intragenic interaction between two SNPs (rs1136450 and rs4253527) of the
There were no significant interactions observed between SNPs of the hydrophilic and hydrophobic SPs. In contrast, previous observations have shown an association of
This study, to our knowledge, is the first to show that interactions among three SNPs of the SP genes and their epistatic effect associate with RDS susceptibility. The majority of prior studies have at most reported interactions between two SNPs of the SP genes. The three SNP models in the present study showed that the highest number of intergenic and intragenic interactions involved
The SNP rs1136450 with a dominant effect had the highest number of interactions (
The rs1059046 SNP of
There was one significant intragenic interaction (rs2077079, rs3024798, and rs7316). Each SNP exhibited a dominant effect and this interaction was associated with decreased risk of RDS. The rs2077079 (C/A) is located 10 nt downstream of the TATAA box, 5′ regulatory region and may affect gene transcription. The rs3024798 (A/C) is located at the splice sequence of intron 2-exon 3 and may affect splicing. The rs7316 (A/G) is located in the 3′UTR, at 4 nt upstream of the TAATAAA polyadenylation signal and may affect polyadenylation (
None of the SNPs were identified in the three SNP model, even though single
The
Although the present study has a relatively large sample size, one limitation is that the patient population differs from that of the controls in terms of age, birth weight, and sex. However, the analyses were adjusted for age and sex (birth weight was not corrected due to collinearity with gestational age). Another study limitation may be reduced generalizability as both study groups were predominantly whites. It is also possible that we have missed some significant interactions due to the use of stringent criteria such as those imposed by the FDR correction, set at 1% to avoid spurious associations. Nonetheless, the present findings need to be replicated. The SNP interactions and their association with the disease phenotype may be affected by the severity of RDS, which was not captured in this study. Around 40% of the data on important parameters such as antenatal use of steroids were missing and that may have introduced bias in the estimation of the difference between groups. The diagnosis of chronic lung disease included oxygen use at 28 days or oxygen at 36 weeks postmenstrual age. The definition for BPD has evolved over time and hence the study characteristic does not capture the current definition of BPD, consistently, as per NICHD 2019 (
Despite the above limitations, this study indicates a greater role of
Furthermore, surfactant lipids and SP-A exhibit anti- and pro-inflammatory effects, respectively, on immune cells under baseline conditions, and surfactant lipids have been shown to attenuate the SP-A effect (
The data analyzed in this study are subject to the following licenses/restrictions: the de-identified dataset is part of the FLOROS biobank at the Penn State University, College of Medicine. Requests to access these datasets should be directed to Joanna FLoros,
The studies involving human participants were reviewed and approved by Institutional Research Board (IRB) at Penn State University, College of Medicine. Written informed consent to participate in this study was provided by the participants' legal guardian/next of kin.
SA: data curation. MY, LY, and RW: formal analysis. JF: funding acquisition. BN and JF: resources. RW and JF: supervision and writing—review and editing. SA, CG, and JF: writing—original draft. All authors read and approved the final manuscript.
This study was supported by NIH grant R37 HL34788 to JF.
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.
The authors would like to acknowledge all the collaborators associated with the different institutions as they have been mentioned in previously published papers and Dr. R. Auten for contributing nine samples and Dr. T. Weaver and P. Ballard for contributing one specimen each.
Single nucleotide polymorphism
respiratory distress syndrome
surfactant protein gene
birth weight
surfactant protein.