3.6.1. HHIP Serum Level Comparison

This analysis was performed by comparing HHIP serum protein levels between both groups, regardless of polymorphism; however, this comparison was not significant (*p* = 0.62) (Figure 2).

The comparison between cases and controls depending on the polymorphism showed significant differences with the SNPs rs13147758 (AA; *p* = 0.021) and rs1828591 (AA; *p* = 0.023); however, after the logistic regression analysis, these associations were not significant. A similar result was observed in the intra-case analysis, as rs13147758 showed a difference between cases and controls with the AA genotype (*p* = 0.021), and after adjustment for covariates, it changed to *p* = 0.28. The effect same occurred with rs1828591 (*p* = 0.046, after adjustment, *p* = 0.053). In this analysis, rs13118928 did not present any significant difference before adjustment for covariates.

#### 3.6.2. Analysis of Serum Protein Levels by Applying Genetic Models

The analysis of genetic association models was performed using the dominant and recessive models, comparing protein levels in both groups.

Among individuals exposed to biomass-burning smoke without the disease (BBES), the dominant model (AA vs. AG + GG) of rs13118928 showed that individuals who carry one or two copies of the minor allele (G) have higher protein levels compared with individuals who are homozygous for the common allele (A) (*p* = 0.005, and after adjustment, *p* = 0.04). However, in the COPD-BS group, significant differences were not observed (Figure 3).

**Figure 3.** Levels of HHIP protein in serum by the dominant model of rs13118928. \* The statistical difference between BBES with AA vs. AG+GG (*p* < 0.05).

3.6.3. Correlations of Serum Protein Levels in COPD-BS and BBES with Lung Function

The rs1828591 showed a positive correlation among heterozygosity (AG) and the FEV1/FVC ratio in the COPD-BS group (*p* = 0.01, r2 = 0.56), and after logistic regression analysis, this value remained significant (*p* = 0.02, r<sup>2</sup> = 0.52) (Figure 4A).

Two positive correlations were observed in rs13118928: the first in the BBES group with the AG genotype and FEV1 (*p* = 0.04, r2 = 0.50; after the analysis by covariates, *p* = 0.03, r2 = 0.53). In addition, in the COPD-BS group, the heterozygous (AG) genotype showed a positive correlation with FEV1/FVC (*p* = 0.006, r<sup>2</sup> = 0.56 and adjusted values *p* = 0.012 and r<sup>2</sup> = 0.54) (Figure 4B,C).

**Figure 4.** Correlations of protein levels and lung function. (**A**) correlation of FEV1/FVC and protein levels in heterozygous (AG) BBES group of the rs1828591; (**B**) correlation of FEV1 and protein levels in heterozygous (AG) BBES group; and (**C**) correlation of FEV1/FVC and protein levels in heterozygous (AG) COPD-BS group, both in the rs13118928.

#### *3.7. HHIP Levels in Supernatant Sputum of Smoke Biomass Burning*

Analysis of HHIP protein levels in sputum supernatant (SS) from subjects exposed to smoke by biomass burning was performed in a subgroup of 40 randomly selected participants (20 COPD-BS and 20 BBES).

The age of the COPD-BS group (median = 72) was higher than that of the BBES group (median = 61.5) ~10 years, which was determined to be significant. For BBEI, the COPD group was more heavily exposed with a median = 470 h/year, and the BBES group had a median of 360 h/year; however, this difference was not significant. On the other hand, pulmonary function test results are lower in the COPD-BS group than in the BBES group, and this difference is expected in our study, since lung function tests are the diagnostic criteria (Supplementary Table S3).

Protein Level Comparison in Sputum Supernatant

In general, when comparing the protein levels between both groups, regardless of the polymorphism, the BBES group had higher protein levels than did the COPD-BS group (*p* = 0.09); however, this difference was not significant (Figure 5).

**Figure 5.** Comparison of HHIP protein in supernatant sputum levels among COPD-BS 4. 3.

In the analysis, according to genotypes among case-controls, no differences were found in the general association, genetic models, or correlations.

### *3.8. HHIP Levels and Metabolism of Drugs for COPD*

Nine drugs were identified as the first-line treatment (tiotropium bromide, umeclidinium, vilanterol, fluticasone furoate, salbutamol, ipratropium bromide, budesonide, fluticasone, and/or salmeterol). Based on the above, the in silico analysis showed a strong interaction between the different drugs and the enzymes, carriers, targets, and transporters that participate in their metabolism. However, as shown in supplementary Figure S2, no interaction among the drugs used in treating COPD, and the HHIP protein was identified.

#### **4. Discussions**

COPD is caused by exposure to noxious particles or gases, such as cigarette smoke or smoke caused by the combustion of solid fuels, known as biomass, which may include coal or crop residue, grass, dry branches, animal dung, charcoal, and wood [24,25].

In the current study, three polymorphisms (rs13118928, rs1828591, and rs13147758) in the *HHIP* gene were analyzed in Mexican women exposed to biomass-burning smoke; this kind of environmental risk factor is relatively common in low- and middle-income countries, such as Mexico, in addition to women being more likely than men to be affected by this exposure [26,27]. Additionally, some authors have suggested that biomass-burning smoke has a different pathophysiological mechanism from that caused by cigarette smoke [28].

The clinical and demographic characteristics of our study group are similar to those of previous reports; Moran-Mendoza et al., in 2008, described a group of women affected by exposure to smoke from biomass burning, aged approximately 67 years, mostly residing in rural or suburban areas, who had 5 h/day of direct exposure to smoke for up to 45 years [29]. Additionally, the BSEI may reach 275 h/year [25]. In our current study, the COPD group exceeded 300 h per year for harmful particles.

Previously, two GWAS seeking to identify genetic susceptibility to COPD in the Caucasian population found two SNPs in the *HHIP* gene (rs1828591 and rs13118928) with consistent replications in three cohorts; however, the combined *p*-values did not reach levels of significance at the GWAS level (1.74 <sup>×</sup> 10−<sup>7</sup> and 1.67 <sup>×</sup> 10−7) [7]. In addition, in the Boston Early-Onset COPD (BEOCOPD) family cohort, rs1828591 and rs13118928 were associated with FEV1 (*p* = 0.0025 and *p* = 0.0014, respectively), but neither SNP was significantly associated with COPD. Additionally, these same polymorphisms were associated with FEV1 in the British Birth Cohort (*p* = 0.039 and *p* = 0.038, respectively) [7]. In another GWAS, rs13147758 had a genome-wide *p*-value associated with FEV1/FVC and FEV1, as well as airflow obstruction, in smokers [8].

Some studies in Asian populations have evaluated these particular polymorphisms in the HHIP gene but focused on their relationship with tobacco smoking; in the KOLD cohort, 15 SNPs were analyzed, and none was associated with COPD; however, four were significantly associated with FEV1 value [30]. Similarly, in a Chinese Han population, none of the SNPs in HHIP had an association with COPD, but the rs12509311, rs13118928, and 1,828,591 were associated with the FEV1/FVC ratio in COPD smokers [31]. In contrast, three SNPs (rs13147758, rs1828591, and rs13118928) were associated with a decreased COPD risk (OR = 0.57, 0.54 and 0.56, respectively) with the GG genotype in a comparative study between Chinese Han and Mongolian populations [32]. We selected these SNPs because a critical role in the disease has been observed in other populations; however, in biomass-burning smoke-exposed individuals, their participation has not been evaluated to date. In our current study in biomass-burning smoke-exposed women, we found that the rs13118928 GG genotype was associated with a decreased risk (OR = 0.51) of the disease; interestingly, these populations have a higher Amerindian component [33]. Regarding the analysis by genetic association models (recessive model), it was shown that carrying two copies of the minor allele (GG) also provides a decreased risk of disease. In a recent study in Mexican-mestizo smokers, we found an association with COPD susceptibility with rs13147758 [34] but not rs13118928. [33,34].

Interestingly, the GG haplotype (rs13118928–rs1828591) shows that both SNPs minor alleles confer a decreased risk of COPD. Previously, we reported that rs13147758 (not associated in the current exposure comparison) and the haplotype formed with rs1828591 are associated with smokers' COPD susceptibility. In the current study, we did not observe any genetic association with susceptibility at the allele or haplotype level. This lack of association is probably due to the minimal differences in the genotype frequencies among cases and controls (approximately 3% between groups), which is in contrast to the tobacco-smoking study, where the difference in genotype frequencies reached 10%. [33,34].

The *HHIP* gene encodes a protein with the same symbol; this protein is a regulatory component of the Hedgehog pathway for embryonal development in vertebrates. HHIP is a transmembrane protein that attenuates the signals of this pathway and, from an evolutionary perspective, it is highly conserved. HHIP plays an essential role in embryogenesis processes, such as lung development and the development of other organs [11]. In a haploinsufficient murine model exposed to cigarette smoke, the importance of the HHIP protein in lung development was demonstrated; homozygous (Hhip-/-) mice died in the short term after birth due to defects in lung morphogenesis. In contrast, heterozygous mice (Hhip+/−) were viable with normal lung development but with an approximately 33% decrease in protein expression and an increase in functional and histological emphysema [35].

The diagnosis of COPD is performed through the spirometry test, and no other procedures are indicated (such as bronchoalveolar lavage or lung tissue biopsy) [1]. We believe that the biological sample closest to the pulmonary microenvironment is sputum; therefore, in this study, we decided to determine protein levels in sputum supernatant and serum. In addition, these measurements have not been previously reported for this pathology.

Interestingly, most of the results associated with the disease were obtained in serum protein levels; however, comparing cases and controls, regardless of polymorphisms, no significant differences were found.

In the case-control comparison, depending on the genotype/polymorphism, with rs13118928 being associated at the genetic level, no differences at the serum protein level were observed. Interestingly, among non-COPD exposed subjects, according to the dominant model (AA vs. AG + GG), the HHIP serum levels are increased among subjects carrying one or two copies for the minor allele (G). Notably, the GG genotype was associated with decreased risk. In this work, we observe that all subjects with

one or two copies of the minor allele have higher levels of the protein, which acquires relevance at the biological level, since previous investigations in murine models observed that the protein may play a protective role against harmful stimuli, such as exposure to cigarette smoke [36]. However, this protective role has not been previously described in people exposed to smoke from biomass burning.

In the lung function correlations with serum protein, the HHIP levels correlated positively with the FEV1 and FEV1/FVC parameters. Additionally, the rs13118928 AG genotype showed that both in controls with FEV1 and in cases with FEV1/FVC, lung function was better at higher serum protein levels, suggesting that the protein is stimulated by the oxidative stress caused by noxious particles or gases, such as those contained in the smoke from biomass combustion, similar to that described in murine models of tobacco smoke exposure [36]. Our findings with the AG genotype but not with the GG genotype could be due to the reduced number of subjects carrying the GG genotype.

On the other hand, supernatant sputum protein levels did not exhibit significant differences, either in the first case-control comparison or in subsequent analyses, where comparisons depend on genotypes/SNPs. Notably, the HHIP protein levels in the sputum supernatant were consistently lower than those in serum samples.

According to the databases and software consulted [22,23], no direct or indirect interactions of the different pharmacological drugs with the HHIP protein were identified; this allows us to presume so far HHIP levels are not they are altered by the pharmacological treatment used.

Our study has a number of limitations, among which we consider the small number of sputum supernatant samples obtained for protein level determination; although the statistical power for serum and supernatant protein levels is 68% and 60%, respectively, in a study such as the one that we conducted and in a poorly explored population, this approach presents good statistical power between protein levels and lung function.

Finally, our study analyzed three polymorphisms that had been previously associated with COPD in other populations. However, to the best of our knowledge, no previous association studies have investigated populations exposed to smoke due to biomass burning. Additionally, our study is the first to describe the protein levels in serum and sputum in subjects exposed to smoke by biomass burning with and without COPD. Our results may pave the way for further studies investigating COPD pathophysiology, where the HHIP participates at the gene-variation and protein levels.

#### **5. Conclusions**

The rs13118928 GG genotype and the rs13118928–rs1828591 (GG) haplotype are associated with decreased COPD risk in Mexican women exposed to smoke from biomass burning.

In addition, the HHIP serum protein levels in subjects harboring the rs13118928 AG genotype exposed to smoke by biomass burning, both with and without COPD, are associated with better lung function.

**Supplementary Materials:** The following are available online at http://www.mdpi.com/2075-4418/10/10/734/s1, Figure S1: Haplotypes of the rs13147758 and rs18285918 in the HHIP gene, Figure S2: In silico analysis from the metabolism of drugs in the treatment of COPD, and the interaction of HHIP with pharmacological treatment, Table S1: Genotype frequencies comparison among COPD patients, according to G2 vs. G1 groups, Table S2: Characteristics of selected women for serum protein levels analysis, Table S3: Demographical, clinical, exposition and lung function of exposed to biomass-burning smoke of the subgroup for protein levels in supernatant sputum; Table S4: General characteristics of SNPs included in the analysis.

**Author Contributions:** Conceptualization, A.O.-M., E.A.-R. and R.F.-V.; data curation, A.O.-M.; formal analysis, A.O.-M.; funding acquisition, A.R.-V. and R.F.-V.; investigation, A.O.-M., M.E.R.-D., F.C.-V., M.d.L.M.-G., E.R.-M. and E.A.-R.; methodology, A.O.-M., M.E.R.-D., F.C.-V., M.d.L.M.-G., E.R.-M. and E.A.-R.; project administration, G.P.-R., A.R.-V. and R.F.-V.; resources, G.P.-R., A.R.-V. and R.F.-V.; software, A.O.-M.; supervision, G.P.-R., A.R.-V. and R.F.-V.; validation, R.F.-V.; visualization, M.E.R.-D., F.C.-V., M.d.L.M.-G., E.R.-M., E.A.-R. and R.F.-V.; writing—original draft, A.O.-M. and R.F.-V.; writing—review and editing, A.O.-M., G.P.-R., A.R.-V., E.R.-M. and R.F.-V. All authors have read and agreed to the published version of the manuscript.

**Funding:** This work was supported by the budget allocated to research (RFV-HLA Laboratory) from the Instituto Nacional de Enfermedades Respiratorias Ismael Cosío Villegas (INER). This paper constitutes a partial fulfillment of the Graduate Program in Maestría en Ciencias de la Salud of the Instituto Politécnico Nacional (IPN) for A.O.-M. A. Ortega-Martínez acknowledges the scholarship (920256) and financial support provided by the National Council of Science and Technology (CONACyT). The authors acknowledge the support received from physicians and technicians from the COPD clinic at INER for confirmation of diagnosis, acquisition of data on lung function, and clinical care of the study participants.

**Conflicts of Interest:** The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.

**Data Availability Statement:** The datasets generated and analyzed for this study can be found in ClinVar SCV001423136, SCV001423137, and SCV001423138.
