A degradation fragment of type X collagen is a real-time marker for bone growth velocity

Abstract

Despite its importance as a key parameter of child health and development, growth velocity is difficult to determine in real time because skeletal growth is slow and clinical tools to accurately detect very small increments of growth do not exist. We report discovery of a marker for skeletal growth in infants and children. The intact trimeric noncollagenous 1 (NC1) domain of type X collagen, the marker we designated as CXM for Collagen X Marker, is a degradation by-product of endochondral ossification that is released into the circulation in proportion to overall growth plate activity. This marker corresponds to the rate of linear bone growth at time of measurement. Serum concentrations of CXM plotted against age show a pattern similar to well-established height growth velocity curves and correlate with height growth velocity calculated from incremental height measurements in this study. The CXM marker is stable once collected and can be accurately assayed in serum, plasma, and dried blood spots. CXM testing may be useful for monitoring growth in the pediatric population, especially responses of infants and children with genetic and acquired growth disorders to interventions that target the underlying growth disturbances. The utility of CXM may potentially extend to managing other conditions such as fracture healing, scoliosis, arthritis, or cancer.

Copyright © 2017 The Authors, some rights reserved; exclusive licensee American Association for the Advancement of Science. No claim to original U.S. Government Works.

Figures

Fig. 1

Depiction ofmammalian type X collagen. (A) Noncollagenous N-terminal (NC2) and C-terminal (NC1) domains are connected by a collagenous triple helix. The NC1 domain is subdivided into a compact “C1q-like” region that resolves in the crystal structure and a linker region that does not. (B) Schematic of antibodybinding regions and collagenase sites. Solid lines indicate peptide sequences to which polyclonal antibodies (pAbs) were raised. Hatched lines indicate regions within which X53 and X34 monoclonal antibodies (mAbs) bind. Also shown are two sites susceptible to collagenase cleavage. aa, amino acids.

Fig. 2

Identification and subunit characterization of CXM marker. (A) Western blots of umbilical cord serum, adult serum, and recombinant full-length human type X collagen (rCOLX) (positive control). Equivalent blots of 4 to 12% gels were probed with antibodies to the noncollagenous NC2 domain (left panel), collagen helix (center panel), and noncollagenous NC1 domain (right panel). Fourth panel: Representative Coomassie stain of serum proteins present in cord and adult lanes. (B) Left panel: Western blot of immunoprecipitated collagen X marker (CXM) eluted at pH 7.0 versus pH 2.5, separated on a 12% gel, and probed with a pAb (USCNK) to the NC1 domain. Right panel: Recombinant trimeric NC1 (rNC1) separated by SDS–polyacrylamide gel before (left lane) or after (right lane) pH 2.5 treatment and stained for protein. std refers to molecular mass standards.

Fig. 3

Mass spectrometry analysis of CXM marker. (A) Boxed area: Region defined by high-confidence peptides identified in mass spectrometry analysis. Above box: Amino acids immediately upstream of identified region that include the proposed collagenase cut site (↓). Below box: The lack of tryptic cut site in C-terminal 50 amino acids makes this peptide too large to identify. (B) Semitryptic high-confidence peptide sequences identified by mass spectrometry are represented by stacked horizontal lines corresponding to their placement within the CXM marker. Proposed collagenase cut site (↓) corresponds to amino acid position 480. Functional domains are diagrammed above the graph with the linker region defined by a shaded box. See fig. S1 for the graph of peptides whose N and C termini are both tryptic.

Fig. 4

Marker decreases with age and is detected in human urine and mouse blood. Western blots of CXM aptoprecipitated with SOMA1and probed with X34 mAb from (A) serum of individuals of increasing ages (0 year, umbilical cord serum) or (B) matched urine and serum samples from a 2-month-old infant (Vol, volume of sample; Exp, exposure time for autoradiography). (C) Aptoprecipitated trimeric markers from human serum (CXM) or mouse serum (Cxm) probed with pAbs raised against their respective recombinant NC1 domains and compared to Coomassiestained gels of the same recombinant proteins (rNC1).

Fig. 5

Correlation of tail and long bone growth velocities with Cxm serum concentrations in mice. (A) Cxm serum concentration (blue) and the growth velocity (red) of mouse tails were plotted against age of mice (n = 29). (B and C) Cxm serum concentrations were plotted against matched femur (B) or tibia (C) growth velocities (n = 29), with linear fit lines in black and 95% confidence intervals in red. Respective Pearson’s correlations are as follows: femur, r = 0.82, P < 0.0001; tibia, r = 0.89, P < 0.0001.

Fig. 6

CXM correlates with age and growth velocity. (A) Serum CXM is plotted against age for normally growing infants and children (n = 129). Established height velocity curve averages for males (blue line) and females (red line) are superimposed for comparison. (B) CXM is plotted against age, grouped by sex, and shown as means ± SE. Sex-matched velocity norms for males and females are superimposed as before. (C) Infants and children 0.18 to 16 years of age were measured for length/ height and assayed for serum CXM at 0-, 6-, and 12-month periods (n = 44). Height velocities were calculated as change in length/height over time interval, converted to centimeter per year, and plotted against CXM [adjusted R2 (weighted) = 0.88, P < 0.001). (D) Log-transformed CXM serum concentrations for normally growing children and nongrowing adults are plotted against age (n = 139).

Fig. 7

CXM concentration increases during adult fracture healing. Plot showing CXM concentration measured at different time points after acute long bone fractures in a 29-year-old male and in 47- and 64-year-old females. Arrow indicates refracture in the 47-year-old patient.

Fig. 8

Diurnal variation of CXM. Morning and afternoon CXM concentrations from dried blood spots for different aged children. Subject A: a 4-year-old female tested morning and afternoon for three consecutive weeks (n = 27). Average CXM readings were plotted. Subject B: an 11-year-old female tested morning and afternoon for three consecutive weeks (n = 28). Average CXM readings and SD are plotted.