Limits...
Multi-modal proteomic analysis of retinal protein expression alterations in a rat model of diabetic retinopathy.

VanGuilder HD, Bixler GV, Kutzler L, Brucklacher RM, Bronson SK, Kimball SR, Freeman WM - PLoS ONE (2011)

Bottom Line: The aim of this study was to identify retinal proteomic alterations associated with functional dysregulation of the diabetic retina to better understand diabetic retinopathy pathogenesis and that could be used as surrogate endpoints in preclinical drug testing studies.Alterations in pro-inflammatory, signaling and crystallin family proteins were confirmed by orthogonal methods in multiple independent animal cohorts.These proteins, especially those not normalized by insulin therapy, may also be useful in preclinical drug development studies.

View Article: PubMed Central - PubMed

Affiliation: Department of Pharmacology, Penn State College of Medicine, Hershey, Pennsylvania, United States of America.

ABSTRACT

Background: As a leading cause of adult blindness, diabetic retinopathy is a prevalent and profound complication of diabetes. We have previously reported duration-dependent changes in retinal vascular permeability, apoptosis, and mRNA expression with diabetes in a rat model system. The aim of this study was to identify retinal proteomic alterations associated with functional dysregulation of the diabetic retina to better understand diabetic retinopathy pathogenesis and that could be used as surrogate endpoints in preclinical drug testing studies.

Methodology/principal findings: A multi-modal proteomic approach of antibody (Luminex)-, electrophoresis (DIGE)-, and LC-MS (iTRAQ)-based quantitation methods was used to maximize coverage of the retinal proteome. Transcriptomic profiling through microarray analysis was included to identify additional targets and assess potential regulation of protein expression changes at the mRNA level. The proteomic approaches proved complementary, with limited overlap in proteomic coverage. Alterations in pro-inflammatory, signaling and crystallin family proteins were confirmed by orthogonal methods in multiple independent animal cohorts. In an independent experiment, insulin replacement therapy normalized the expression of some proteins (Dbi, Anxa5) while other proteins (Cp, Cryba3, Lgals3, Stat3) were only partially normalized and Fgf2 and Crybb2 expression remained elevated.

Conclusions/significance: These results expand the understanding of the changes in retinal protein expression occurring with diabetes and their responsiveness to normalization of blood glucose through insulin therapy. These proteins, especially those not normalized by insulin therapy, may also be useful in preclinical drug development studies.

Show MeSH

Related in: MedlinePlus

Coverage and differential expression across analysis methods.Bioinformatic comparison of datasets generated by the three proteomic and one transcriptomic approaches used was performed to compare commonalities and differences in coverage and differential expression. (A) Comparison of proteomic coverage of unique species identified across methods is illustrated by Venn diagram. iTRAQ, DIGE and Luminex provided complementary coverage with limited overlap between methods. Of the 527 unique proteins identified in this study, 110 were identified by both iTRAQ and DIGE, and only one protein included in the directed Luminex analysis was also identified by an open-profiling method. For the majority of identified proteins, corresponding transcripts were present at detectable level in the transcriptomic analysis. The greater coverage of the transcriptomic data compared to the proteomic coverage is primarily due to the greater sensitivity of the whole-genome microarray approach. (B) Comparison of differentially expressed species detected by each method is depicted. As with proteomic coverage, quantitative data proved complementary across methods, with a different subset of differentially expressed proteins detected by each of the three proteomic approaches. In the two cases where a protein was differentially-expressed in both iTRAQ and DIGE quantitation, the direction and magnitude of change were comparable. Numerous differentially-expressed transcripts were identified by microarray analysis, although only a portion of the proteomic changes were detected at the transcript level.
© Copyright Policy
Related In: Results  -  Collection


getmorefigures.php?uid=PMC3020973&req=5

pone-0016271-g004: Coverage and differential expression across analysis methods.Bioinformatic comparison of datasets generated by the three proteomic and one transcriptomic approaches used was performed to compare commonalities and differences in coverage and differential expression. (A) Comparison of proteomic coverage of unique species identified across methods is illustrated by Venn diagram. iTRAQ, DIGE and Luminex provided complementary coverage with limited overlap between methods. Of the 527 unique proteins identified in this study, 110 were identified by both iTRAQ and DIGE, and only one protein included in the directed Luminex analysis was also identified by an open-profiling method. For the majority of identified proteins, corresponding transcripts were present at detectable level in the transcriptomic analysis. The greater coverage of the transcriptomic data compared to the proteomic coverage is primarily due to the greater sensitivity of the whole-genome microarray approach. (B) Comparison of differentially expressed species detected by each method is depicted. As with proteomic coverage, quantitative data proved complementary across methods, with a different subset of differentially expressed proteins detected by each of the three proteomic approaches. In the two cases where a protein was differentially-expressed in both iTRAQ and DIGE quantitation, the direction and magnitude of change were comparable. Numerous differentially-expressed transcripts were identified by microarray analysis, although only a portion of the proteomic changes were detected at the transcript level.

Mentions: Complementarity of the multi-modal proteomic approach used here was assessed by comparing identities and biophysical properties of unique species detected/identified by technology. Overlap in proteomic coverage was limited (Figure 4A), with no commonality between DIGE and Luminex, and only one species (Got1) detected by both iTRAQ and Luminex. Non-directed approaches demonstrated more overlap, with 110 unique protein products identified by both DIGE (154 proteins total) and iTRAQ (439 proteins total). As expected, the majority (>80%) of species detected/identified by proteomic approaches were also detected at the transcriptomic level by high-density microarray analysis (Figure 4A). Commonality between methods at the level of differential expression was much more limited, with little overlap between proteomic methods. Only Hspa1b and Lap3, which were determined to be up-regulated with diabetes by both DIGE and iTRAQ, were observed by multiple methods (Figure 4B). Importantly, this limited overlap was not due to a protein being observed as differentially expressed by one proteomic method and unchanged by another. Rather, the differences in proteins identified as differentially expressed were largely observed by only one method. Comparison of differentially-expressed species identified by proteomic and transcriptomic analyses demonstrated that few (<10%) changes in protein expression were also observed in the mRNA analysis. While this comparative analysis of the different methods is illuminating it should be noted that the same number of samples were not used in each analysis due to differences in the techniques resulting in slightly differing levels of statistical power.


Multi-modal proteomic analysis of retinal protein expression alterations in a rat model of diabetic retinopathy.

VanGuilder HD, Bixler GV, Kutzler L, Brucklacher RM, Bronson SK, Kimball SR, Freeman WM - PLoS ONE (2011)

Coverage and differential expression across analysis methods.Bioinformatic comparison of datasets generated by the three proteomic and one transcriptomic approaches used was performed to compare commonalities and differences in coverage and differential expression. (A) Comparison of proteomic coverage of unique species identified across methods is illustrated by Venn diagram. iTRAQ, DIGE and Luminex provided complementary coverage with limited overlap between methods. Of the 527 unique proteins identified in this study, 110 were identified by both iTRAQ and DIGE, and only one protein included in the directed Luminex analysis was also identified by an open-profiling method. For the majority of identified proteins, corresponding transcripts were present at detectable level in the transcriptomic analysis. The greater coverage of the transcriptomic data compared to the proteomic coverage is primarily due to the greater sensitivity of the whole-genome microarray approach. (B) Comparison of differentially expressed species detected by each method is depicted. As with proteomic coverage, quantitative data proved complementary across methods, with a different subset of differentially expressed proteins detected by each of the three proteomic approaches. In the two cases where a protein was differentially-expressed in both iTRAQ and DIGE quantitation, the direction and magnitude of change were comparable. Numerous differentially-expressed transcripts were identified by microarray analysis, although only a portion of the proteomic changes were detected at the transcript level.
© Copyright Policy
Related In: Results  -  Collection

Show All Figures
getmorefigures.php?uid=PMC3020973&req=5

pone-0016271-g004: Coverage and differential expression across analysis methods.Bioinformatic comparison of datasets generated by the three proteomic and one transcriptomic approaches used was performed to compare commonalities and differences in coverage and differential expression. (A) Comparison of proteomic coverage of unique species identified across methods is illustrated by Venn diagram. iTRAQ, DIGE and Luminex provided complementary coverage with limited overlap between methods. Of the 527 unique proteins identified in this study, 110 were identified by both iTRAQ and DIGE, and only one protein included in the directed Luminex analysis was also identified by an open-profiling method. For the majority of identified proteins, corresponding transcripts were present at detectable level in the transcriptomic analysis. The greater coverage of the transcriptomic data compared to the proteomic coverage is primarily due to the greater sensitivity of the whole-genome microarray approach. (B) Comparison of differentially expressed species detected by each method is depicted. As with proteomic coverage, quantitative data proved complementary across methods, with a different subset of differentially expressed proteins detected by each of the three proteomic approaches. In the two cases where a protein was differentially-expressed in both iTRAQ and DIGE quantitation, the direction and magnitude of change were comparable. Numerous differentially-expressed transcripts were identified by microarray analysis, although only a portion of the proteomic changes were detected at the transcript level.
Mentions: Complementarity of the multi-modal proteomic approach used here was assessed by comparing identities and biophysical properties of unique species detected/identified by technology. Overlap in proteomic coverage was limited (Figure 4A), with no commonality between DIGE and Luminex, and only one species (Got1) detected by both iTRAQ and Luminex. Non-directed approaches demonstrated more overlap, with 110 unique protein products identified by both DIGE (154 proteins total) and iTRAQ (439 proteins total). As expected, the majority (>80%) of species detected/identified by proteomic approaches were also detected at the transcriptomic level by high-density microarray analysis (Figure 4A). Commonality between methods at the level of differential expression was much more limited, with little overlap between proteomic methods. Only Hspa1b and Lap3, which were determined to be up-regulated with diabetes by both DIGE and iTRAQ, were observed by multiple methods (Figure 4B). Importantly, this limited overlap was not due to a protein being observed as differentially expressed by one proteomic method and unchanged by another. Rather, the differences in proteins identified as differentially expressed were largely observed by only one method. Comparison of differentially-expressed species identified by proteomic and transcriptomic analyses demonstrated that few (<10%) changes in protein expression were also observed in the mRNA analysis. While this comparative analysis of the different methods is illuminating it should be noted that the same number of samples were not used in each analysis due to differences in the techniques resulting in slightly differing levels of statistical power.

Bottom Line: The aim of this study was to identify retinal proteomic alterations associated with functional dysregulation of the diabetic retina to better understand diabetic retinopathy pathogenesis and that could be used as surrogate endpoints in preclinical drug testing studies.Alterations in pro-inflammatory, signaling and crystallin family proteins were confirmed by orthogonal methods in multiple independent animal cohorts.These proteins, especially those not normalized by insulin therapy, may also be useful in preclinical drug development studies.

View Article: PubMed Central - PubMed

Affiliation: Department of Pharmacology, Penn State College of Medicine, Hershey, Pennsylvania, United States of America.

ABSTRACT

Background: As a leading cause of adult blindness, diabetic retinopathy is a prevalent and profound complication of diabetes. We have previously reported duration-dependent changes in retinal vascular permeability, apoptosis, and mRNA expression with diabetes in a rat model system. The aim of this study was to identify retinal proteomic alterations associated with functional dysregulation of the diabetic retina to better understand diabetic retinopathy pathogenesis and that could be used as surrogate endpoints in preclinical drug testing studies.

Methodology/principal findings: A multi-modal proteomic approach of antibody (Luminex)-, electrophoresis (DIGE)-, and LC-MS (iTRAQ)-based quantitation methods was used to maximize coverage of the retinal proteome. Transcriptomic profiling through microarray analysis was included to identify additional targets and assess potential regulation of protein expression changes at the mRNA level. The proteomic approaches proved complementary, with limited overlap in proteomic coverage. Alterations in pro-inflammatory, signaling and crystallin family proteins were confirmed by orthogonal methods in multiple independent animal cohorts. In an independent experiment, insulin replacement therapy normalized the expression of some proteins (Dbi, Anxa5) while other proteins (Cp, Cryba3, Lgals3, Stat3) were only partially normalized and Fgf2 and Crybb2 expression remained elevated.

Conclusions/significance: These results expand the understanding of the changes in retinal protein expression occurring with diabetes and their responsiveness to normalization of blood glucose through insulin therapy. These proteins, especially those not normalized by insulin therapy, may also be useful in preclinical drug development studies.

Show MeSH
Related in: MedlinePlus