Main

Sudden infant death syndrome (SIDS) is defined as “the sudden unexpected death of an infant <1 y of age with onset of the fatal episode apparently occurring during sleep, that remains unexplained after a thorough investigation, including performance of a complete autopsy and review of the circumstances of death and the clinical history” and subcategorized by revised definition (1). Today, SIDS is still the major cause of death among infants between 1 mo and 1 y of age in developed countries (2,3). SIDS is generally agreed to be a multifactorial disease and according to the frequently encountered “triple-risk hypothesis,” SIDS may occur when an infant left vulnerable by a combination of predisposing and developmental risk factors encounters an environmental trigger event (e.g., sleeping in prone position, overheating) (4,5). Predisposing risk factors include alterations in a plethora of genes, e.g., involved in thermoregulation and catecholamine production, as well as conditions such as abnormal brainstem processes in the regulation of autonomic respiration and/or arousal (6,7).

Thermal stress, e.g., caused by slight infections, exaggerated room heating, or excessive insulation of the infant, is known to be a risk factor for SIDS (8,9,10,11,12,13,14,15,16,17), especially affecting infants with impaired thermal regulation. This is because there is a strong association between thermal regulation and ventilatory control, specifically for prolonged apnea (18,19), which can be the proximate mechanism of death in SIDS. Thermal regulation is in part mediated by the thermal stress response, so it has been hypothesized that a deficient thermal stress response due to genetic alterations in the gene for heat shock protein 60 (Hsp60), encoded by HSPD1 (20), may play a role in the pathogenesis of a subgroup of SIDS cases (21).

However, to date only two studies investigated HSPD1 genetic alterations in SIDS and these brought forth contradictory results as to their potential impact: Rahim et al. analyzed an MspI restriction fragment length polymorphism in HSPD1 and reported one specific fragment to be significantly underrepresented in the SIDS group (22). The meaning of this study’s findings is limited, however, by its small collective (n = 12) and its lack of correction for multiple testing in statistical analysis (23). Bross et al. investigated the distribution of several single-nucleotide polymorphisms in HSPD1 in a series of SIDS (n = 61) and control cases but in no instance could they establish a significant difference between SIDS and control cases (24). Moreover, both the studies focused on variants that had not previously been shown or suspected to be pathogenic by functionally impairing Hsp60.

This study is the first to examine in a substantially larger number of cases whether the nonsynonymous and known to be pathogenic (25) rare variant rs72466451 in the gene for Hsp60 is associated with the occurrence of SIDS.

Results

Distribution of HSPD1 SNV Allelic Variants

Table 1 shows the results of the distribution analysis of the allelic variants of the HSPD1 SNV in controls and cases of SIDS. A significant association of the pathogenic “C” allele with SIDS was observed (P = 0.05).

Table 1 Distribution of allelic variants of HSPD1 SNV
Full size table

The “C” allele was unevenly, although not significantly differently, distributed between genders, with both control cases and five (of six) SIDS cases bearing a “C” allele being male ( Table 2 ). Moreover, although there was an uneven distribution of “C” alleles between the age-grouped SIDS cases, with four (of six) “C” alleles found in SIDS cases in the risk age between 46 and 153 d, there was no significant association (P = 0.532).

Table 2 Properties of SIDS and control cases bearing a pathogenic allele
Full size table

Distribution of HSPD1 SNV Genotype Variants

Table 3 shows the results of the distribution analysis of the genotype variants of the HSPD1 SNV in controls and cases of SIDS. A significant association of the C/T genotype with SIDS was shown (P = 0.05).

Table 3 Distribution of genotype variants of the HSPD1 SNV
Full size table

However, the homozygous and pathogenic genotype C/C was not detected at all, neither in control nor in SIDS cases. Therefore, the distribution of the heterozygous C/T genotype equals the respective allelic distribution, with the C/T genotype being more frequent in males and in SIDS cases in the risk age between 46 and 153 d, although it was not significantly different in any of these instances (data not shown). Moreover, genotype distribution was calculated not to depart from Hardy–Weinberg equilibrium for SIDS and control cases (data not shown).

Discussion

SIDS is most likely a complex and multifactorial disease. According to the “triple-risk hypothesis,” contributing or predisposing factors are classified into genetic, developmental, and environmental factors (“triggers”). Therefore, each of a vast number of possible combinations of these factors could precipitate SIDS. However, with SIDS being a diagnosis of exclusion, it is important to keep in mind that no single factor, predisposing or else, can be determined that “causes” SIDS, because, per definition, SIDS is to be diagnosed only if no (other) cause can be determined (1). Therefore, many of the predisposing, i.e., genetic factors that have already been reported to be associated with the occurrence of SIDS (26,27,28) do not cause manifest disease that would thwart compatibility with a diagnosis of SIDS. Instead, they are thought to elicit or to be associated with only subtle changes in, e.g., biochemical, physiological, or immunological processes. These small alterations could then, if coalescing with a complementary instable developmental state and appropriate environmental conditions, provoke fatal crisis.

Many of these studies fail, however, to provide evidence that the genetic factors in question that had been associated with SIDS do have some, however small, effect. Hence, there is no shortage of reports acclaiming associations but providing only contrived hypotheses as to what these associations might actually mean.

The present study, therefore focuses on an SNV in the coding region of the gene for Hsp60, HSPD1, that creates a bona fide amino acid change in the protein and that has already been shown to cause manifest disease if present in a homozygous genotype (25). In addition, evidence for functional impairment of Hsp60 by this SNV has been produced by demonstrating reduced survival in transgenic Escherichia coli, especially when confronted with heat stress (25).

We hypothesize that homozygosity for this rare variant’s pathogenic allelic variant, which notably was not observed in any single case in this study, probably causes a manifest, i.e., “stand-alone” disease that under particular circumstances might even be lethal in early phases of embryonic or fetal development and thereby precludes diagnosability of SIDS. If this hypothesis is true, it should also be investigated whether some deaths initially reported as SIDS were in fact due to a disease caused by a pathogenic homozygous genotype of the rs72466451 SNV.

Heterozygosity for this rare variant, however, is non-lethal, but the amino acid change it causes might create a somewhat impaired heat shock reaction by restricting the chaperonin’s ability to bind or refold thermally deformed protein structures or to target them for degradation in the proteasome. This may cause an increased vulnerability to overheating that is not sufficient for symptomatic disease but is applicable to classify as a predisposing factor for SIDS that affects both genders alike. This hypothesis is corroborated by the fact that the pathogenic variant was also found in two controls who had not died of SIDS.

It follows from the above explicated interpretation of the “triple-risk hypothesis”—that is to say “many combinations of factors may lead to SIDS”—that certain combinations of complementary factors may have a higher plausibility of working together in causing SIDS. For instance, an infant with a genetic predisposition that impairs heat stress response might be more at risk when this predisposition is met with the “overheating” trigger. We propose that further studies of association take into account as far as possible potential reciprocity between the genetic factor in question and the actual trigger (e.g., prone position, overheating, slight infection) that was reported for each SIDS patient.

Conclusions

We observed a significant relationship between the genotype of a functional HSPD1 rare variant (rs72466451) and SIDS. This is the first study investigating the distribution of this pathogenic SNV in a SIDS case–control collective. Our results suggest considering a possible role of a genetically impaired heat shock reaction in the pathogenesis of a small subgroup of SIDS cases and should direct further work.

Methods

Patients and Diagnosis

Our study collective consisted of 133 formalin-fixed, paraffin-embedded samples of German Caucasian SIDS cases (82 males, 51 females, age of death 26–341 d, median 98 d) and 192 gender-matched control cases (120 males, 72 females). In all the cases, diagnosis of SIDS was according to the current definition (1) and based on a detailed and comprehensive postmortem investigation at the Institute of Legal Medicine (University of Bonn, Bonn, Germany). A total of 57.9% of SIDS patients were in the risk interval between 46 and 153 d postnatal upon death.

The German Caucasian controls consisted of biopsy material taken from the psoas major muscle during autopsy (n = 154; 94 males, 60 females, age of death 3–96 y, median 59 y) and voluntarily donated saliva samples from healthy subjects (n = 38; 26 males, 12 females, aged 2–38 y, median 12 y). For deceased controls, diagnosis of SIDS was excluded. The study design provided for informed consent and was approved by the ethics committee of the medical faculty of the University of Bonn.

DNA Extraction

For SIDS cases, DNA was extracted from formalin-fixed paraffin-embedded tissue using the QIAamp DNA FFPE Tissue Kit (Qiagen, Hilden, Germany). For control cases, the QIAamp DNA Investigator Kit (Qiagen) was used to extract DNA from saliva swabs and the Qiagen DNA Micro Kit (Qiagen) was used with tissue specimens. All kits were utilized according to the manufacturer’s respective instructions. DNA concentration was determined fluorometrically using the Qubit fluorometer (Invitrogen/Life Technologies, Darmstadt, Germany) according to the supplied recommendations.

Primer Design and PCR Amplification of HSPD1

Amplification and typing primers were designed using the free online resource “Primer 3” (Whitehead Institute for Biomedical Research, Cambridge, MA), checked for secondary structures and inter- and intracomplementarity using the “Netprimer” software (Premier Biosoft, Palo Alto, CA), and tested for specificity by performing a query using the Basic Local Alignment Search Tool (“BLAST”, hosted online by the National Library of Medicine, Bethesda, MD). The primer details are listed in Table 4 . The SNV in question, rs72466451, lies on chromosome 2 at position 198363487 (contig NT_005403.17) in exon 2 of HSPD1, and its G allele has been classified as “pathogenic.” There are as yet no data concerning population diversity in regard to rs72466451 (29).

Table 4 Primers for HSPD1 locus amplification and SNV genotyping
Full size table

A 139 bp stretch of the SNV-containing region was then amplified by PCR using the following cycling conditions: initial denaturation at 94 °C for 10 min, denaturation at 94 °C for 30 s, annealing at 55 °C for 30 s, elongation at 72 °C for 30 s, and final elongation at 72 °C for 5 min, with 30 cycles. PCR was set up using standard conditions, with 1 ng of DNA template and 1 U of Taq-Polymerase (New England Biolabs, Ipswich, MA) used per reaction.

Genotyping the HSPD1 Variant

PCR products were cleaned up before sequencing using Exo-SAP-It (Affymetrix, Santa Clara, CA). Genotyping of the amplified SNV was then performed by minisequencing using the typing primer listed in Table 4 and the SNaPshot Multiplex System (Applied Biosystems/Life Technologies, Darmstadt, Germany) according to the manufacturer’s instructions. Afterward, the products were treated with calf intestine phosphatase (Affymetrix) to eliminate interference with subsequent fluorescence detection. Sequence analysis could then be performed using a 310 Genetic Analyzer and GeneMapper software v.3.2 (Applied Biosystems/Life Technologies) according to the manual.

Statistical Analysis

The frequencies of the pathogenic allelic variants in the control group examined in this study were too small for a Pearson’s χ2 test to be reliably performed. Therefore, Fisher’s exact test was used instead; all P values reported herein are one-sided and derived from it. Calculation of Fisher’s exact tests, Bonferroni correction, and possible departure from Hardy–Weinberg equilibrium were performed using SPSS software v. 19 (SPSS, Chicago, IL).

Statement of Financial Support

No financial support was received for this study.

Disclosure

The authors report no financial ties or conflicts of interest.