Reader Comments
Post a new comment on this article
Post Your Discussion Comment
Please follow our guidelines for comments and review our competing interests policy. Comments that do not conform to our guidelines will be promptly removed and the user account disabled. The following must be avoided:
- Remarks that could be interpreted as allegations of misconduct
- Unsupported assertions or statements
- Inflammatory or insulting language
Thank You!
Thank you for taking the time to flag this posting; we review flagged postings on a regular basis.
closeComment on: Endogenous morphine levels are increased in sepsis: a partial implication of neutrophilsr your comment title...
Posted by zhuwei on 09 Mar 2010 at 01:49 GMT
We have read the publication by Glattard, Welters et al. with great interest and acknowledge the significant efforts that were expended by all investigators. The results and conclusions forwarded by the authors indicate a significant physiological response mediated by the endogenous morphine signaling system in subpopulations of patients afflicted with sepsis. Although the overall message of the work appears to be valid, the study suffers from numerous methodological flaws that clearly distort accurate descriptions of key mechanistic aspects of endogenous morphine signaling in human disease. Our major concerns are listed below.
1. Molecular biological characterization of the mu opiate receptor expressed by PMNs
We have demonstrated with considerable cross-validation that PMNs express the novel 6 transmembrane helical (TMH) type of mu3/mu4 opiate receptor. Mu3 and mu4 receptors are cognate GPCRs selectively activated by morphine, related morphinan alkaloids, and are unresponsive to all known families of endogenous opioid peptides(Cadet et al., 2003,Cadet et al., 2007,Kream et al., 2007). The PCR primer set utilized in the present study is designed to amplify an 85 bp sequence within the protein coding region that is common to the traditional mu1 opioid receptor as well as mu3 and mu4 receptors. Importantly, the PCR primer set does not span an Exon/Exon splice site boundary, thereby introducing the possibility for amplification of genomic DNA.
Previous publications from our laboratory have demonstrated via the use of highly selective primer sets exclusive expression of mu3 and mu4 opiate receptors by PMNs in the absence of traditional mu1 opioid receptor expression (above). These data have been fully validated and replicated by state-of-the art real time PCR analyses indicating selective expression of 6TMH domain mu3/mu4 opiate receptors by PMNs. Presently, the authors further confound this critical issue by showing PCR amplification of an 85 bp product from extracted RNA obtained from a SH-SY5Y neuroblastoma cell line. Many laboratories including ours have previously demonstrated SH-SY5Y cells predominantly express traditional mu1 opioid receptors as well as mu3 and mu4 opiate receptors as minor species.
It is our firm contention that PMNs express the “short” mu3/mu4 opiate receptor and not the traditional mu1 7TMH domain opioid receptor. This has been confirmed in a recent publication, whereby the investigators utilize a mu1 opioid selective primer set and fail to amplify a PCR product containing an Exon 1/Exon2 specific nucleotide sequence from human monocytes (Williams et al., 2007 Anesth Analg 105:998-1005).
In the present publication, the lack of selectivity of the employed PCR primer set as well as the reliance on only one PCR primer set underscores a significant lack of resolution in the characterization of the type of mu opiate receptor expressed by PMNs. It is likely, however, that the authors have effectively amplified mu3/mu4 opiate receptor-encoding mRNA obtained from PMNs, not MOR1 as stated in the publication. Finally, the observed alteration of mu opiate receptor expression subsequent to LPS treatment is consistent with previously published data from our laboratory indicating enhancement of mu3 and mu4 opiate receptor-encoding mRNA following immune challenge.
2. Biochemical and cellular characterization of the mu opiate receptor expressed by PMNs
The veracity of the comparative Western Blot analysis of gel fractionated membrane protein obtained from SH-SY5Y neuroblastoma cells and PMNs is markedly diminished by a lack of critical understanding of the relative abundance of mu opioid/opiate receptors in SH-SY5Y cells vs. human leukocytes. It is well established that cellular distributions of mu opiate receptors are found in extremely low abundance in non-neuronal cells such as blood monocytes, i.e., approximately 2 fmol/mg membrane protein (Stefano et al. 1993, Proc Natl Acad Sci USA 90: 11099-11103). SH-SY5Y cells represent a neuroblastoma cell line and express significantly higher levels of traditional 7TMH domain mu1opioid receptors on the order of 50-100fmol/mg membrane protein (Borner et al. 2004, Mol Pharmacol 66: 1719-1726). Unfortunately, the investigators chose to perform the Western blot analysis utilizing a 50 ug protein aliquot from both cell sources. Based on the 1-2 orders of magnitude lower abundance of mu opioid/opiate receptors in leukocytes, an immunoprecipitation trace enrichment step was required to produce a credible Western Blot analysis. Despite the authors’ contentions, the clearly visible 55 KDa band from SH-SY5Y cells is barely detectable in the PMN lane, and its authenticity is questionable. Finally, the omission of a crucial control experiment demonstrating selective inhibition of the 55 KDa Western Blot signal by a cognate blocking peptide sequence further diminishes enthusiasm for these experiments.
The flow cytometry experiments designed to quantitatively detect mu1 opioid receptors on the extracellular surface of intact leukocytes in organotypic culture are equivalently flawed. As stated above, the authors have not critically estimated the extremely low abundance of mu opioid/opiate expressed by leukocytes and instilled confidence that the intrinsically low quantum yield of the FITC-labeled N-terminally directed mu1 opioid receptor antibody molecules will provide acceptable S/N ratios in the analyses. Additionally, the lack of a convincing Western Blot analysis greatly diminishes the credulity of the flow cytometry studies. Accordingly, the authors do not provide secondary validation that the FITC-labeled N-terminally directed mu1 opioid receptor antibody binds to the membrane protein in its native conformation. The omission of a requisite control experiment demonstrating selective inhibition of the fluorescent signal by a cognate blocking peptide sequence raises major concerns that the labeling pattern achieved by the antibody accurately reflects cellular distributions of mu1 opioid receptors. In strong support of our contentions, Williams and coworkers, in the same study cited above, were unable to achieve a fluorescent flow cytometry signals from human blood monocytes utilizing 2 different N-terminally directed mu1 opioid receptor antibody molecules linked to significantly more sensitive reporter systems (Williams et al., 2007 Anesth Analg 105:998-1005).
3. Biochemical quantification of endogenous morphine in PMNs and human plasma
In similar fashion to the molecular biological and biochemical analyses of mu opioid/opiate receptors expressed by leukocytes, the authors have not critically evaluated the requisite type of immunoassay capable of accurate and reliable quantification of trace levels of endogenous morphine in leukocytes and human plasma. Instead, they have employed an intrinsically insensitive, optically driven, commercially available morphine ELISA for quantification of endogenous morphine in the cellular release experiments and in the plasma of sepsis patients and healthy volunteers. Based on data presented in Fig. 2, the true limit of detection, defined as a statistically significant 10% inhibition of maximal signal, is approximately 20 pg/ml. The sensitivity of the morphine ELISA, defined as the 50% displacement point, is approximately 200pg/ml. Furthermore, the investigators’ general lack of expertise in RIA/ELISA procedures is reflected by their basic misunderstanding and incorrect reporting of Coefficient of Variation (CV) values for the ELISA.
It has been well established in the biochemical literature that reliance on data points clustered at the limit of detection introduces significant levels of inaccuracy into calculated data sets due to a very high level of variance. Accurate immunoassay quantification of a low abundance analyte such as endogenous morphine requires a serial dilution of each biological sample with displacement values falling within the sensitive linear range of the assay, i.e., the 50% displacement point. It is therefore quite evident from inspection of the ELISA scatter plot depicted in Fig.2 and the use of 40ul duplicate samples of unextracted cell culture medium, that morphine concentrations in the cellular release experiments are markedly overestimated. Back calculations from the reported values of immunoreactive morphine released into cell culture medium yield absurdly high PMN cellular concentrations of greater than 200 pg morphine/million cells. Furthermore, the estimation of PMN morphine concentrations, as determined by ELISA, differs markedly by at least 2 orders of magnitude from values of <1pg morphine/million PMN cells, as determined by Q-TOF MS/MS, an accepted reference standard for authentication of all chemical compounds. Although the presently reported Q-TOF MS/MS value is approximately 10 fold lower than the 10pg morphine/million PMN cells previously established by our group using similar Q-TOF MS/MS methodology (Zhu et al., 2005 J Immunol 175:7357-7362), differences in PMN cell pellet preparation as well as the inclusion of erythrocyte lysis buffer may have resulted in significant losses of cellular morphine during extraction. It is our firm contention, however, that accumulated Q-TOF MS/MS data from both studies support a true PMN morphine concentration varying between 1 and 10pg morphine/million cells.
Major concerns also apply to the chemical characterization and quantification of morphine in extracted human plasma. Comparative analysis of the paired Q-TOF MS/MS spectra depicted in Fig. 5 indicates a true value of approximately 0.5 pmol morphine/ml of patient plasma, equivalent to a concentration of 0.5 nM. In effect, The Q-TOF MS/MS analysis of extracted patient plasma clearly demonstrates a marked 10 fold overestimation of plasma morphine concentrations in sepsis patients at 8nM, as determined by ELISA. Interestingly, the Q-TOF fragmentation spectrum of morphine-like material from extracted patient plasma differs markedly in its relative peak ratios from authentic morphine standard, thereby indicating significant contamination by one or more morphine metabolites in the preparation.
In sum, major deficiencies in the employed ELISA procedures, as well as the lack of secondary validation measures such as a standard combined HPLC/immunoassay analysis to demonstrate coelution of morphine immunoreactivity with authentic morphine standard, markedly diminish enthusiasm for the PMN morphine release experiments and for accurate assessment of enhanced levels of circulating endogenous morphine in sepsis patients. It is also unfortunate that internal inconsistencies in the presentation of markedly overestimated and unreliable ELISA morphine concentrations with true values established by Q-TOF MS/MS are decidedly underplayed. This lack of critical evaluation of complementary data sets can only obscure hard scientific observations, interpretations, and conclusions emanating from the present study. The engenderment of artificial controversy by citing a letter from the Zenk group that denies expression of endogenous morphine by PMNs (Boettcher et al., 2006 J Immunol 176:5703-5704) appears disingenuous in light of past (Zhu et al., 2005 J Immunol 175:7357-7362) and present Q-TOF MS/MS data.
In conclusion, circulating PMNs contain intrinsically low but biologically meaningful levels of chemically authentic morphine. PMNs utilize endogenously expressed morphine to selectively activate cognate mu3 and mu4 receptors coupled to constitutive nitric oxide (NO) production to effectively regulate local circuit autocrine/paracrine cellular responses to proinflammatory mediators. Pathophysiological changes in endogenous morphine/mu3 and mu4 receptor/NO-coupled regulatory events reflect first line defense mechanisms by which living organisms fight to survive.
Critical Reference List
Cadet P, Mantione K J, Stefano G B. Molecular identification and functional expression of mu3, a novel alternatively spliced variant of the human mu opiate receptor gene. Journal of Immunology 2003; 170(10): 5118-5123.
Cadet P, Mantione K J, Zhu W, Kream R M, Sheehan M, Stefano G B. A functionally coupled mu3-like opiate receptor/nitric oxide regulatory pathway in human multi-lineage progenitor cells. Journal of Immunology 2007; 179(9): 5839-5844.
Kream R M, Sheehan M, Cadet P, Mantione K J, Zhu W, Casares F M, Stefano G B. Persistence of evolutionary memory: Primordial six-transmembrane helical domain mu opiate receptors selectively linked to endogenous morphine signaling. Medical Science Monitor 2007; 13(12): SC5-SC6.
Richard Kream Ph.D.
Kirk J Mantione Ph.D.
George B. Stefano Ph.D. Dr. h.c.
Wei Zhu Ph.D.
Neuroscience Research Institute
SUNY Old Westbury
Old Westbury, NY 11568
RE: Comment on: Endogenous morphine levels are increased in sepsis: a partial implication of neutrophilsr your comment title...
ingewelt replied to zhuwei on 29 Mar 2010 at 20:44 GMT
We would like to thank Kream and colleagues for their interest in our work and we appreciate the very detailed nature of their comments. However, we feel that several aspects of our work may have been misinterpreted and we wish to counter most of the criticism expressed.
Firstly, and most importantly, we would like to emphasise that the aim of our investigation was to investigate plasma levels of endogenous morphine as a marker of infection, not to specify the role of different µ opioid receptor subtypes in neutrophil activation. Secondly, we have identified circulating neutrophils as a source of endogenous morphine production in severe life-threatening infection. As discussed in our paper, these results contradict previous data published by Boettcher et al. (Boettcher C, J Immunol 2006;176:5703), but support previous reports on the presence of endogenous morphine in white blood cells (Zhu W, J Immunol. 2005;175:7357). By investigating granular localisation, LPS and IL-8 induced secretion of morphine as well as involvement of µ receptors we have now further elucidated the release of morphine in response to pro-inflammatory stimulation. We provide evidence that the average endogenous morphine concentration found in septic patients is able to inhibit IL-8 secretion by neutrophils, suggesting an immunomodulatory role for endogenous morphine in systemic infection. Overall, our results strongly support the concept of endogenous morphine production by neutrophils as an important factor in immune homeostasis during life-threatening infection.
1. PCR and antibody detection of µ opioid receptors
The OPRM1 gene, which was amplified in our experiments, encodes the µ opioid receptor which is the primary site of action of morphine. We appreciate that different splice variants of the µ opioid receptor are expressed on various types of cells. However, we have to reiterate that the specification of µ opioid receptor splice variants and their specific role in immune regulation was not the intention of our study; our primary aim was to establish a general involvement of the morphinergic system in the regulation of neutrophil function during sepsis and systemic inflammation. As correctly outlined by Kream and colleagues, neutrophils express the µ3/µ4 opioid receptor subtype, which is selectively activated by morphine. However, the use of specific primers for these opioid receptor subtypes would only serve to confirm our results and corroborate previous reports about the effects of µ3/4 activation on neutrophils. Further investigation of the specific µ opioid receptor subtype would not provide any additional information on morphine as a biomarker of severe infection or how it regulates neutrophil function.
Although undoubtedly a variety of splicing variants with their own pharmacological profile are expressed on different cell types, numerous articles using primers directed against the conserved region of the MOR-1 have been published (www.ncbi.nlm.nih.gov/site...). Thus, identification of the core region of µ opioid receptors is an established technique to demonstrate presence of µ opioid receptor expression in general and this formed the overall objective of our study.
To date, no specific antibody against the µ3 opioid receptor subtype exists. Thus, the use of a validated antibody directed against the conserved region of the µ opioid receptor is the only way to assess and quantify expression on the surface of blood cells. However, we extended the specification of µ opioid receptors by using antibodies against both the C-terminal and the N-terminal region of MOR-1. The suitability of this µ opioid receptor antibody against the external N-terminal region for flow cytometry has been described in a previous study (Beck M et al. Pain 2002;98:187), and our results confirm this report. The N-20 antibody is directed against an epitope mapping an extracellular domain of human MOR-1. Transcript variants of the µ opioid family may lead to N-terminal changes, but do not necessarily affect the binding capacity of the antibodies we used. Thus, the µ3 hypothesis, although not relevant for our investigation, is not contradictory to our results. The fact that an antibody against the highly variable C-terminal part of the µ opioid receptor failed to produce a band in Western blots and did not generate a relevant fluorescence in our flow cytometry experiments further supports the presence of MOR-transcripts and translational products in neutrophils.
2. Characterisation of the µ opioid receptor:
Our Western blot experiments add further information on the expression of µ opioid receptors. Firstly, they confirm their presence on blood neutrophils. Secondly, they demonstrate the suitability of this antibody for the detection of µ opioid receptors, as the antibodies used in our experiments produced results, which were in line with previous reports (Beck M et al. Pain 2002;98:187). The low expression of opioid receptors on neutrophils is reflected in relatively lower fluorescence signals compared to those obtained after antibody staining of more abundant receptors, e. g. CD11b. The cross-validation of the antibodies by Flow cytometry and Western blot, by blocking experiments and by testing for non-specific binding is in keeping with current scientific standards. The study cited in Kream’s comments (Williams JP, Anesth Analg 2007;105:997) was exclusively performed on peripheral blood mononuclear cells, a cell preparation which excludes neutrophils. This study therefore neither supports nor contradicts our results on neutrophil µ opioid receptor expression.
The SH-SY5Y extract was used as a positive control of the molecular weight of the respective band, not for quantification. The fact that we obtained a visible band in both neutrophil and SH-SY5Y extracts despite a significantly lower expression of opioid receptors in immune cells confirms previous reports regarding the detectability of µ opioid receptors in various cell types (Makman MH et al. Adv Neuroimmunol 1994;4:69). Blocking with a synthetic peptide was performed and showed the absence of the respective band after immunoadsorption of the antibody. Since, as mentioned above, the characterization of the µ opioid receptor subtype was not the aim of our study, immunoprecipitation experiments were not considered.
3. Quantification of endogenous morphine in PMNs and human plasma:
Most of the concerns expressed in Kream’s comments are outlined in our publication in sufficient detail. Our choice of an optically driven ELISA represents the appropriate method for screening blood samples. The ELISA kit used in our study is one of the two non-radioactive methods currently available and carries the benefit of exclusively recognising authentic morphine and not its derivatives. In preliminary experiments we validated the kit by testing all samples with ELISA as well as MS-MS (positive samples, spiked standard), confirming the kit is accurate and reliable for serial samples from septic patients. Furthermore, the detection limit of the ELISA is equivalent to the detection limit of a RIA used in the study cited by Kream (0.5ng/ml) (Zhu et al. J Immunol 2005;175:7357). In addition to direct analysis, standards and samples were deproteinised and purified either by a Sep Pack cartridge (elution with 15% acetonitrile) or by HPLC. After purification, the peak corresponding to morphine (samples and standard) was analysed by both ELISA and MS-MS. In summary, our ELISA results clearly demonstrate that morphine is present in the concentration range reported.
In the present study the 50% displacement point was not defined as the sensitivity of the ELISA. Instead, the methodological Detection Limit (DL) was calculated according to well described procedures (Immuno-Stat methodology book, Nucleon; ISBN: 2-84332-001-1).
This methodological DL reflects the EC50. As outlined in our original publication, the DL value was 0.01ng for the batch of the kit used in this study. In view of the concerns of Kream and colleagues, we feel that a lack of familiarity with ELISA techniques does not warrant general criticism of results, which have been validated by several techniques.
Morphine secretion experiments were performed on 20 x 106 neutrophils to ensure that a sufficient amount of morphine was secreted. The same experiments were repeated on 70 x 106 neutrophils. A parallel decrease of morphine concentrations was observed after dilution of the supernatant confirming the accuracy of our ELISA analysis for cell media.
Figure 2 clearly shows the linearity of the ELISA to detect morphine in plasma and tissue culture media samples. As stated in our publication, 40 µl samples of unextracted tissue culture medium were assayed in duplicate and yielded coefficients of variations between 0 and 8%, a variation which is well within the range for biological samples and fulfils all quality standards (Immuno-Stat methodology book, Nucleon; ISBN:2-84332-001-1). It is surprising that the CV range was omitted in the publication cited in Kream’s comments (Zhu et al. J Immunol 2005;175:7357).
Kream and colleagues state that the back calculation estimates the morphine amount to be greater than 200 pg morphine/million cells. As outlined in our publication, 200 µl of incubation medium containing 20x106 cells was used for the stimulation experiments. Thus, the concentration of morphine present within a million of cells was extrapolated to be 1.6 pmole/million cells and 0.6 pmole/million cells, respectively. As obtained by back calculation, the concentrations in the culture medium of neutrophils stimulated with IL-8 for 7 min (1.6 pg/million cells and 0.6 pg/million cells, respectively) are comparable to the actual concentrations obtained from neutrophil extracts (0.32+/−0.2 pg/million of neutrophils). Interestingly, the culture medium of neutrophils exposed to LPS for 6 hours yielded higher morphine concentrations, as biosynthesis and release of endogenous morphine were stimulated. These results are in line with previously described rises in morphine concentrations in serum, brain spinal chord and adrenal gland after surgery and LPS administration (Brix-Christensen V et al. Int J Cardiol 1997;62:191; Brix-Christensen V et al. Acta Anaesthesiol Scand 2000;44:1204; Yoshida S et al. Surg Endosc 2000;14:137; Goumon Y et al. Neurosci Lett 2000;293:135; Meijerink WJ et al. Shock 1999;12:165).
Differences in quantification of morphine released by neutrophils are potentially due to different methodological approaches. In our experiments the neutrophils were purified from buffy coat, whereas the study cited by Kream et al (Zhu W et al. J Immunol 2005;175:7357) used heparinised whole blood. Disparate blood preparations as well as different cell isolation techniques are known to alter neutrophil activity and may thereby modify morphine production. Variations of the morphine concentrations between different studies are therefore easily explained by the methodology used.
Figure 5 represents the morphine detected in the plasma of a septic patient. MS-MS spectra were used to calculate the endogenous morphine concentration in the sample. Using peak intensities, a back calculation revealed the presence of morphine in the range of 2.4 nM in the plasma of this particular patient. This is consistent with the morphine concentrations (3.9 and 5.4 nM) in two further patient samples analysed with MS-MS.
Combined SPE and HPLC/immunoassay analysis were performed to demonstrate coelution of morphine immunoreactivity with an authentic morphine standard before analysis of patient samples. This confirmed the presence of authentic morphine. Confirmation of authentic morphine by the methods used in study was an essential prerequisite for the analysis of any clinical samples and is therefore not covered in the present publication. The deproteinised extracts of patients with sepsis were purified by HPLC and the peaks corresponding to morphine were analysed by both ELISA and MS-MS. Thus, the coelution of morphine immunoreactivity with authentic morphine standard was unequivocally demonstrated.
Regarding the Q-TOF fragmentation spectrum, Kream and colleagues attempt to attribute differences between the peak ratios of the morphine-like substance extracted from patient plasma and the authentic morphine standard to contamination by one or more morphine metabolites in the preparation. However, the additional peaks result from the fragmentation of molecules not corresponding to morphine but displaying the same molecular weight (MW) before fragmentation. These results strongly support the use of MS-MS approach instead of classical MS analysis. In biological samples such as tissue extracts or body fluids such as plasma, molecules with the same MW are isolated in the first stage of MS and new fragments different from those expected for morphine fragmentation are produced. These nonspecific peaks are not present in the external standard but can be found in the plasma of healthy donors in the absence of morphine fragments. Although we have not identified the biochemical nature of these unspecific peaks, our ELISA technique, which has a high specificity for morphine and lacks cross-reactivity for any morphine derivatives, confirmed the presence of endogenous morphine only in patient plasma, not in healthy individuals.
In conclusion, we have provided new information about the mechanism by which endogenous morphine regulates neutrophil function. As a matter of good scientific practice, contradictory reports were included in our discussion in an effort to provide a comprehensive overview of morphine’s immunoregulatory capacity. We fully agree with Kream and colleagues that neutrophils represent an important source of endogenous morphine which regulates immune homeostasis by autocrine feedback mechanisms involving µ opioid receptors.
Ingeborg Welters, MD/PhD
Yannick Goumon, PhD
University of Liverpool
Clinical Sciences
Duncan Building, Daulby Street
L69 3GA
Liverpool, UK