Growth failure during infancy is a major global problem that has adverse effects on long-term health and neurodevelopment. Preterm infants are disproportionately affected by growth failure and its effects. Herein we found that extremely preterm infants with postnatal growth failure have disrupted maturation of the intestinal microbiota, characterized by persistently low diversity, dominance of pathogenic bacteria within the Enterobacteriaceae family, and a paucity of strictly anaerobic taxa including Veillonella relative to infants with appropriate postnatal growth. Metabolomic profiling of infants with growth failure demonstrated elevated serum acylcarnitines, fatty acids, and other byproducts of lipolysis and fatty acid oxidation. Machine learning algorithms for normal maturation of the microbiota and metabolome among infants with appropriate growth revealed a pattern of delayed maturation of the microbiota and metabolome among infants with growth failure. Collectively, we identified novel microbial and metabolic features of growth failure in preterm infants and potentially modifiable targets for intervention.

Background: Human milk reduces morbidities in extremely low birth weight (ELBW) infants. However, clinical instability often precludes ELBW infants from receiving early enteral feeds. This study compared clinical outcomes before and after implementing an oropharyngeal colostrum (COL) protocol in a cohort of inborn (born at our facility) ELBW infants. Study Design: This is a retrospective cohort study of inborn ELBW infants admitted to the Duke Intensive Care Nursery from January 2007 to September 2011. In November 2010, we initiated a COL protocol for infants not enterally fed whose mothers were providing breastmilk. Infants received 0.1 mL of fresh COL to each cheek every 4 hours for 5 days beginning in the first 48 postnatal hours. We assessed demographics, diagnoses, feeding history, and mortality and for the presence of medical necrotizing enterocolitis (NEC), surgical NEC, and spontaneous perforation. Between-group comparisons were made using Fisher's exact test or Wilcoxon rank sum testing where appropriate. Results: Of the 369 infants included, 280 (76%) were born prior to the COL protocol (Pre-COL Cohort [PCC]), and 89 (24%) were born after (COL Cohort [CC]). Mortality and the percentage of infants with surgical NEC and spontaneous perforations were statistically similar between the groups. The CC weighed an average (interquartile range) of 1,666 (1,399, 1,940) g at 36 weeks versus 1,380 (1,190, 1,650) g for the PCC ( p < 0.001). In a multivariable analysis with birth weight as a covariable, weight at 36 weeks was significantly greater (37 g; p < 0.01). Conclusions: Initiating oropharyngeal COL in ELBW infants in the first 2 postnatal days appears feasible and safe and may be nutritionally beneficial. Further research is needed to determine if early COL administration reduces neonatal morbidity and mortality.

We have cloned via recombinant DNA technology the mRNA sequence for rat pancreatic preprokallikrein. Four cloned overlapping double-stranded cDNAs gave a continuous mRNA sequence of 867 nucleotides beginning within the 5'-noncoding region and extending to the poly(A) tail. The mRNA sequence reveals that pancreatic kallikrein is synthesized as a prezymogen of 265 amino acids, including a proposed secretory prepeptide of 17 amino acids and a proposed activation peptide of 11 amino acids. The activation peptide, although similar in length, is distinct from those of the other classes of pancreatic serine proteases. The amino acid sequence of the predicted active form of the enzyme is closely related to the partial sequences obtained for other kallikrein-like serine proteases including rat submaxillary gland kallikrein, pig pancreatic and submaxillary gland kallikreins, the y subunit of mouse nerve growth factor, and rat tonin. Key amino acid residues thought to be involved in the substrate-cleavage specificity of kallikreins are retained. Hybridization analysis showed relatively high levels of kallikrein mRNA in the rat pancreas, submaxillary and parotid glands, spleen, and kidney, indicating the active synthesis of kallikrein in these tissues.Glandular kallikreins (EC 3.4.21.8) are members of a closely related subfamily of serine proteases that process polypeptide hormone precursors. Other members include the y subunit of nerve growth factor (1), the epidermal growth factor-binding protein (2), and tonin (3). Characteristically, each has a much more limited substrate-cleavage specificity than other serine proteases such as trypsin, chymotrypsin, and elastase. These kallikrein-like proteases cleave at one or a very few peptide bonds in their natural substrates with a general, but not absolute, preference for residues that have positively charged side chains and a strong bias for arginine over lysine (4,5 4).Kallikreins are found in many exocrine tissues, although their site(s) of synthesis remains unverified, and the nature and processing of presumed kallikrein precursors are unknown. Inactive precursor forms of kallikrein have been purified from rat (6) and porcine (7) pancreas. The precursors are acidic glycoproteins of apparent Mr 37,000. Slow activation occurs spontaneously and is accelerated by catalytic amounts oftrypsin. The activated enzymes are single polypeptides with slightly reduced molecular weight, suggesting the release of an activation peptide, which remains uncharacterized. Active kallikrein isolated from autolysed porcine pancreas consists of two polypeptide chains held together by disulfide bridges (8), indicative of further proteolysis.We report the identification and sequence analysis of the cloned mRNA sequence for rat pancreatic kallikrein and describe the nature of the preproenzyme and the presence of kallikrein mRNA in a number of rat tissues.

Four distinct kallikrein-related mRNAs (PS, S1, S2, and S3), encoded by members of a multigene family, are selectively expressed in various combinations in several rat tissues. Although closely related along most of the mRNA sequence, the four mRNAs can be selectively detected with synthetic oligonucleotide probes complementary to highly variable mRNA subregions. PS mRNA, which encodes an enzyme with true kallikrein activity, is present at high levels in the submaxillary gland, pancreas, and kidney. S1 mRNA, which encodes an enzyme similar to the PS kallikrein, is detected only in the submaxillary gland and is present at one-fifth the PS mRNA level. S2 mRNA, which encodes the enzyme tonin, is present in the submaxillary gland at half the PS mRNA level and at a slightly higher level in the prostate. S3 mRNA, which encodes an enzyme very similar to tonin, is present in the submaxillary gland at one-tenth the PS mRNA level and in the prostate at about the same level as tonin mRNA.

We have determined the nucleotide sequence of four submaxillary gland mRNAs, designated PS, S1, S2, and S3, that encode kallikrein and kallikrein-like serine proteases. The four enzymes share between 74% and 86% amino acid sequence identity and are identical in length with the exception of single two amino acid deletions in the S2 and S3 enzymes. The PS enzyme appears to be a true tissue kallikrein. The S1 enzyme shares 86% amino acid sequence homology with the PS enzyme and retains key amino acid residues thought to be primary determinants of kallikrein cleavage specificity. The S2 enzyme is rat submaxillary tonin. The amino acid sequence of the S3 enzyme is identical with tonin at 84% of its amino acid positions and retains the same amino acid substitutions at positions likely to determine substrate cleavage preferences.