Hyperhomocysteinemia interaction with Protein C and Increased Thrombotic Risk
Reporter and Curator: Larry H Bernstein, MD, FCAP
This document explores the relationship between thromboembolic risk related to hyperhomocysteinemia related to the HHcy interaction with and blocking the protective effect of APC.
Previous Venous Thromboembolism Relationships With Plasma Homocysteine Levels
Marco Cattaneo, Franca Franchi, Maddalena L. Zighetti, Ida Martinelli, Daniela Asti, P. Mannuccio Mannucci
Arterioscler Thromb Vasc Biol. 1998;18:1371-1375.
Received January 28, 1998; revision accepted March 16, 1998. From the Angelo Bianchi Bonomi Hemophilia and Thrombosis Center, Institute of Internal Medicine, IRCCS Ospedale Maggiore, University of Milano, Italy.
Correspondence to Marco Cattaneo, MD, Hemophilia and Thrombosis Center, Via Pace 9, 20122 Milano, Italy. E-mail marco.cattaneo@unimi.it © 1998 American Heart Association, Inc. 1371 Original Contributions
Abstract—
The proteolytic enzyme activated protein C (APC) is a normal plasma component, indicating that protein C (PC) is continuously activated in vivo. High concentrations of homocysteine (Hcy) inhibit the activation of PC in vitro;
- this effect may account for the high risk for thrombosis in patients with hyperhomocysteinemia (HyperHcy).
We measured the plasma levels of APC in 128 patients with previous venous thromboembolism (VTE) and in 98 age- and sex-matched healthy controls and
- correlated them with the plasma levels of total Hcy (tHcy) measured before and after an oral methionine loading (PML).
Forty- eight patients had HyperHcy and 80 had normal levels of tHcy. No subject was known to have any of the congenital or acquired thrombophilic states at the time of the study. Because the plasma levels of APC and PC were correlated in healthy controls, the APC/PC ratios were also analyzed.
Plasma APC levels and APC/PC ratios were significantly higher in VTE patients than in controls (P < 0.03 and 0.0004, respectively).
- Most of the increase in APC levels and APC/PC ratios were attributable to patients with HyperHcy.
Patients with normal tHcy had intermediate values, which did not differ significantly from those of healthy controls.
- There was no correlation between the plasma levels of tHcy or its PML increments and APC or APC/PC ratios in controls.
- The fasting plasma levels of APC and APC/PC ratios of 10 controls did not increase 4 hours PML, despite a 2-fold increase in tHcy.
This study indicates that
- APC plasma levels are sensitive markers of activation of the hemostatic system in vivo and
- that Hcy does not interfere with the activation of PC in vivo.
Key Words: homocysteine, protein C, thromboembolism, activated protein C, hypercoagulability, T mechanism.
The zymogen protein C is converted to the active protease, activated protein C (APC),
- through proteolytic cleavage by thrombin bound to its endothelial membrane receptor thrombomodulin.1
The demonstration that APC is a normal plasma component,2,3whose enzymatic activity can be detected with specific and sensitive methods,4,5indicates that
- the protein C anticoagulant pathway is continuously activated in vivo.
Measurement of APC plasma levels might therefore be helpful in determining the in vivo integrity of the protein C anticoagulant pathway. More generally,
- APC levels might mirror the in vivo activation of the coagulation system and
- serve as a marker of thrombin activity in the circulation.4
The mechanism(s) by which a moderate elevation of plasma levels of homocysteine (Hcy) increases the risk for arterial and venous thrombotic disease is still unclear.6,7 In vitro studies showed that
- Hcy inhibits the thrombomodulin- dependent protein C activation to APC and
- interferes with the expression of thrombomodulin on human umbilical vein endothelial cells.8–10
These findings may be relevant to unravel the thrombogenic mechanism of Hcy, because
the protein C anticoagulant system is of major physiological importance in the regulation of the hemostatic congenital or acquired disorders
- characterized by impaired production or function of APC are associated with a high risk for venous thromboembolism (VTE).11
It must be noted, how ever, that these in vitro findings have been obtained by using very high concentrations of Hcy,
- at least 1 order of magnitude higher than the plasma concentrations found in patients with homozygous homocystinuria.12,13
Their clinical relevance is therefore uncertain and awaits confirmation from ex vivo and/or in vivo studies in humans. In this study, we compared the plasma levels of APC with those of the prothrombin fragment F1,2, a marker of thrombin generation,14in healthy subjects and patients with previous episodes of VTE and
- tested whether the levels are affected by plasma Hcy concentrations.
Methods
Materials
L-Methionine, tri-n-butylphosphine, 7-fluoro-2,1,3-benzoxadiazole- 4-sulfonamide (ABDF), L-cystine, Tween 20, Tween 80, benzamidine, and HEPES were from Sigma. (4-Amidinophenyl)-methanesulfonylfluoride (APMSF) was from Boehringer, BSA from Calbiochem, and the chromogenic substrate L-homocystine, ovalbumin, S-2366 from Chromogenics. The monoclonal antibody directed against the light chain of protein C (C3-Mab) was a kind gift of Dr H.P. Schwarz (Immuno, Vienna, Austria). All other chemicals were of reagent grade. Subjects We studied 128 patients with previous VTE and 98 healthy controls. All diagnoses of thrombotic episodes, excluding those of superficial veins, had been confirmed by objective methods: compression ultrasonography or venography for deep vein thrombosis; and ventilation/perfusion scintigraphy for pulmonary embolism. The contemporary presence of deep vein thrombosis in patients with superficial vein thrombosis had not been excluded by objective methods. Table 1 shows the characteristics of the patients studied.
They belonged to a cohort of 315 patients who had been screened for thrombophilic states at our Center between December 1993 and July 1995 and were selected on the basis of the following characteristics:
(1) absence of congenital or acquired thrombophilic states except hyperhomocysteinemia (HyperHcy) (see below);
(2) oral anticoagu- lant therapy discontinued at least 1 month before screening;
(3) at least 4 months elapsed since the last thrombotic episode; and
(4) willingness to participate in the study.
The screening for thrombophilia included the following tests:
- prothrombin time;
- activated partial thromboplastin time;
- thrombin time;
- plasma levels of fibrinogen,
- protein C,
- protein S, and
- antithrombin;
- APC resistance; and
- screening for antiphospholipid syndrome15 and
- plasma levels of total homocysteine (tHcy)
before and 4 hours after an oral methionine load. Patients with abnormal APC resistance were also screened for factor V Leiden.16
The study was designed and completed before the demonstration that the mutation G20210A of the prothrombin gene is a risk factor for deep vein thrombosis.17 This mutation therefore was looked for retrospectively only in those subjects whose DNA was still available for analysis (all controls and 50 patients): 5 patients (10%) and 2 controls (2.1%) were heterozygous for the mutation. Of the 128 patients enrolled in the study,
- 48 had hyperhomocysteinemia (VTE-HyperHcy) according to the diagnostic criteria outlined below, and
- 80 had normal Hcy levels (VTE-NormoHcy).
- The healthy controls, who were age and sex matched with the patients (male/female, 50/45; median age, 40 years [range, 20 to 73 years]), had been chosen from the same geographical area and with the same socioeconomic background as the patients.
- Previous episodes of thrombosis had been ruled out by a validated structured questionnaire.18
- No subject had abnormal liver or renal function, or overt autoimmune or neoplastic disease.
- Informed consent to participate in the study was obtained from all subjects.
- The study was approved by the ethics committee of the University of Milano.
Study Protocol
After an overnight fast, blood samples were drawn between 8:30 and 9:30 AM in K3-EDTA for measurement of total Hcy (tHcy), in 0.013 mol/L trisodium citrate for measurement of F1?2 and protein C, and in citrate plus 0.03 mol/L benzamidine (a reversible inhibitor of APC) for measurement of APC. L-Methionine (3.8 g/m2body surface area) was then administered orally in approximately 200 mL of orange juice. Four hours later, a second blood sample was collected in EDTA for tHcy measurement from all subjects and in citrate plus benzamidine for measurement of APC plasma levels from 10 controls. All subjects remained in the fasting state until the second blood sample had been taken. Plasma Hcy Assay Blood samples in K3-EDTA were immediately placed on ice and centrifuged at 2000xG, 4°C, for 15 minutes. The supernatant was stored in aliquots at < 70°C until assay.
The plasma levels of tHcy (free and protein bound) were determined by high-performance liquid chromatography (Waters Millipore 6000A pump, Millipore) and fluorescence detection (Waters 474) by the method of Ubbink et al,19with slight modifications.20 Briefly, 100 uL of plasma was incubated with 10 uL of 10% tri-n-butylphosphine in dimethylfor- mamide at 4°C for 30 minutes to reduce homocystine and mixed disulfide and deconjugate Hcy from plasma proteins. Then, 100 uL of 10% trichloroacetic acid was added, and the mixture was centrifuged in an Eppendorf microcentrifuge at 13 000 rpm for 10 minutes.
After centrifugation, the mixture was incubated with 1 mg/mL ABDF in borate buffer to derivatize the thiols. The mobile phase, pumped at 1 mL/min, consisted of 0.1 mol/L potassium dihydrogenophosphate, 0.06 mmol/L EDTA, and 12% acetonitrile (pH = 2.1).
- fasting plasma levels of tHcy or its postmethionine load absolute increments above fasting levels exceeded the 95th percentiles of distribution of values obtained in 388 healthy controls.
TABLE 1.
Patients With Previous VTE-NormoHcy
VTE-HyperHcyVTE-NormoHcy 4880
Median age, y (range) 36 (19–69)
Median age at the first thrombotic episode, y (range) 32 (17–62)
Time elapsed since last episode, mo (range) 14 (4–70)
Time elapsed since discontinuation of oral anticoagulant therapy, mo (range) 11 (1–64)Type of first thrombotic episode
Deep vein thrombosis 31/49
Pulmonary embolism 36 (14–62)
Superficial vein thrombosis 31 (13–60)
Venous thrombosis of other sites 14 (4–90)
2233 26 (54.2%)
With circumstantial risk factors* at first episode 44 (55%)
1372
Activated Protein C, Thrombosis, and Homocysteine
centrifuged within 60 minutes from collection at 1200xG, 4°C, for 30 minutes to obtain platelet-poor plasma, which was frozen in aliquots at < 70°C. A plasma pool from 30 healthy individuals (15 men, 15 women) was obtained in the same way and used to prepare the standards.
(removed)… The chromogenic substrate for APC S-2366 (0.46 mmol/L in Tris-buffered saline, pH 7.4) was then added to the wells. After incubation of the sealed plates at 4°C in wet chambers for 3 weeks, hydrolysis of the substrate was monitored at a dual wavelength setting of 405/655 nm.
The concentration of APC in the unknown samples was calculated from the absorbance of each sample with the standard curve as a reference. Results were expressed as percentage of pooled normal plasma. Measurement of Plasma F1?2 F1?2 was assayed by a commercial ELISA (Behringwerke), as previously described.21
Statistical Analysis
The two-tailed t test was used to compare VTE patients and healthy controls. ANOVA was used to compare VTE-HyperHcy, VTE controls, and healthy controls, followed by the Dunnett’s test for internal contrasts. The Pearson r value was calculated for correla- tions between the variables studied.
Results
The results obtained in all VTE patients and controls are presented, including those with the heterozygous G20210A mutation of the prothrombin gene. A subanalysis of the results obtained in the 40 patients and 98 controls, in whom the mutation was looked for, revealed that
- exclusion of the subjects heterozygous for the mutation did not significantly affect the results.
Plasma tHcy Levels
The mean (SD) fasting levels of plasma tHcy were significantly higher in VTE-HyperHcy (28.8?19.5 ?mol/L) than in VTE-NormoHcy (12.0+5.2, P<0.001) and healthy con- trols (11.0+5.3, P<0.001). The mean postmethionine load increments of tHcy above fasting levels were also higher in VTE-HyperHcy (32.9+13.5 umol/L) than in VTE- NormoHcy (19.8+7.5, P<0.001) and healthy controls (16.1+7.6, P<0.001). Differences between VTE-NormoHcy and healthy controls were not statistically significant. Six healthy controls (6.3%) had HyperHcy, according to the diagnostic criteria previously outlined. Plasma Levels of APC Healthy Controls The mean plasma level of APC in healthy controls was 116(20%). There was a statistically significant correlation between the plasma levels of APC and those of protein C (r?0.48, P?0.001) (Figure 1). Therefore, because APC levels are influenced by the concentration of their zymogen, both the absolute APC levels and the activated protein C/protein C (APC/PC) ratios were used for subsequent analysis. The mean value of the APC/PC ratio in healthy controls was 1.01?0.2.
There was no correlation between the plasma levels of APC (not shown) or the APC/PC ratios and the fasting plasma levels of tHcy (Figure 2) or its postmethionine load increments above fasting levels (not shown). The mean APC plasma levels and APC/PC ratios were similar in healthy controls whose tHcy plasma levels fell within the first (115 and 1.0), second (118 and 0.96), or third (115 and 1.01) tertiles of distribution. The mean fasting plasma levels of APC and the APC/PC ratios of 10 healthy controls
– did not significantly differ from those measured in the same subjects 4 hours after an oral methionine load,
– which increased the concentration of tHcy by more than 2-fold (Table 2).
VTE Patients
The mean plasma levels of APC and APC/PC ratios were higher in VTE patients than in healthy controls (124?32 versus 116?20, P?0.03 and 1.12?0.32 versus 0.99?0.19, P?0.0004). This difference was mostly due to VTE- HyperHcy patients whose plasma APC levels and APC/PC ratios were significantly higher than those of healthy controls (Table 3). In contrast, differences between VTE-NormoHcy and healthy controls and between VTE-HyperHcy and VTE- NormoHcy did not reach statistical significance (Table 3). Results did not change substantially when we excluded patients with thrombosis of the superficial veins (APC levels, 124+26 in VTE-HyperHcy and 121?31 in VTE-NormoHcy; APC/PC ratio, 1.17?0.25 in VTE-HyperHcy and 1.09?0.3 Figure 1. Correlation between the plasma levels of protein C and APC in 98 healthy volunteers. Values are expressed as per- centage of the concentrations measured in pooled normal plasma from 30 healthy blood donors. Figure 2. Correlation between the fasting plasma levels of tHcy and APC/PC ratios of 98 healthy volunteers. Cattaneo et al September 1998 1373 in VTE-NormoHcy) or women taking oral contraceptives (APC levels, 115?19 in controls, 130?29 in VTE- HyperHcy, and 121+33 in VTE-NormoHcy; APC/PC ratio, 0.98?0.23 in controls, 1.13?0.4 in VTE-HyperHcy, and 1.08?0.3 in VTE-NormoHcy). The prevalence of high APC/PC ratios was significantly higher in VTE patients than in controls, independent of the tHcy levels in their plasma (Table 4),
-whereas that of high plasma APC levels was significantly increased in VTE- HyperHcy patients only (Table 4).
Plasma Levels of F1?2
The mean plasma level of F1?2 in VTE patients (1.6?0.5 nmol/L) did not significantly differ from that measured in healthy controls (1.5?0.6 nmol/L). There was no statistically significant difference between plasma levels of F1?2 in VTE-HyperHcy (1.6?0.6 nmol/L), VTE-NormoHcy (1.6?0.6 nmol/L), and healthy controls. The mean F1?2 plasma levels were similar in healthy controls whose plasma levels of tHcy fell within the first, second, or third tertiles of distribution (not shown). F1?2 levels and APC/PC ratios were significantly correlated in controls (r?0.28, P?0.005) but not in VTE-HyperHcy (r? ?0.03, P?0.05) or VTE- NormoHcy (r?0.08, P?0.05).
Discussion
This study shows that
– patients with previous episodes of VTE have higher circulating plasma levels of APC than healthy controls, particularly if they have HyperHcy.
The patients studied had none of the known congenital or acquired thrombophilic states, in which
– the circulating levels of markers of activation of the coagulation system may be increased.21–24Even
– though the recently described G20210A mutation of the prothrombin gene17could be looked for retrospectively in only approximately one third of the pa- tients, also those patients in whom the prothrombin mutation was ruled out had high APC levels,
– excluding that they were mainly due to the presence of the mutation.
APC is generated from its plasma precursor, protein C, on activation by thrombin-thrombomodulin complex on the endothelial cell surface, probably acting in concert with the endothelial cell protein C receptor.1Subcoagulant amounts of thrombin in the circulation may increase the plasma levels of endogenous APC, which can therefore be considered markers of a hypercoagulable state.4Accordingly, the high APC plasma levels that we measured in patients with previous episodes of VTE may be interpreted as an index of ongoing thrombin formation,
despite the fact that at least 4 months (and a median of 14 months) elapsed since their last thrombotic episode. However,
– the plasma concentrations of F1?2, a marker of thrombin generation, were not increased signifi cantly in the same VTE patients and were not correlated with APC levels or APC/PC ratios.
In contrast to VTE patients, a statistically significant correlation between APC and F1?2 plasma levels was found in healthy controls. On the basis of these data, we hypothesize that
– the increased plasma levels of APC found in patients with previous episodes of VTE are not caused by heightened thrombin generation but by alternative mechanisms. Although we did not measure markers of activation of the fibrinolytic system,
– the possibility that high plasma levels of plasmin could be responsible for protein C activation25in these patients should be considered.
The greatest increase of APC plasma levels in VTE patients was observed in subjects with fasting and/or postmethionine-loading HyperHcy. VTE patients with nor mal plasma levels of tHcy had lower concentrations of APC than patients with HyperHcy, but this
– difference could be due to chance alone, because it was not statistically significant. These results contrast with the alleged inhibitory effect of Hcy on protein C activation that was shown in in vitro studies.8–10
Our data obtained in healthy individuals
– support the view that Hcy does not affect protein C activation in vivo, because the
– mean plasma levels of APC of subjects in the highest tertile of distribution of tHcy levels were not different from those of subjects in the lowest tertile. Moreover,
– the rapid increase in plasma tHcy brought about by an oral methionine load did not affect the concentration of circulating APC
TABLE 2. Healthy Controls Before and 4 Hours After Methionine Loading (PML) Plasma Levels tHcy, APC, and APC/PC Ratios in 10 tHcy, ?mol/LAPC, %APC/PC Ratio Baseline 4 h PML* P† 10.5?3.8 29.5?7.6 0.0001 118?43 113?32 0.57 0.98?0.2 0.95?0.1 0.7 Data are mean?SD. *Methionine was given orally at a dose of 3.8 g/m2body surface area. †t test for paired samples. TABLE 4. APC/PC Ratios in Healthy Controls, Patients With Previous VTE-HyperHcy, and Patients With Previous VTE-NormoHcy Prevalences of High Plasma Levels of APC and Subjectsn With High APC LevelsWith High APC/PC Ratio n (%)OR (95% CI) n (%)OR (95% CI) Healthy controls 98 10 (10.2) 1.0 (reference) 10 (10.2) 1.0 (reference) VTE-HyperHcy48 12 (25.0) 2.9 (1.1–8.3) VTE-NormoHcy80 16 (20.0) 2.2 (0.9–5.7) 16 (33.3) 4.4 (1.7–11.4) 22 (27.5) 3.3 (1.4–8.1) CI indicates confidence interval. The cutoff points, which corresponded to the 90th percentiles of distribution among healthy controls, were 143.1% for APC levels and 1.22 for APC/PC ratios.
TABLE 3. Controls, Patients With Previous VTE-HyperHcy, and Patients With Previous VTE-NormoHcy
Plasma Levels of APC and APC/PC Ratios in Healthy Subjects nAPC,* %APC/PC Ratio† Healthy controls VTE-HyperHcy VTE-NormoHcy P (ANOVA) 98 48 80 116?20 128?29 121?33 0.03 0.99?0.19 1.15?0.33 1.10?0.31 0.002 Data are mean?SD. *VTE-HyperHcy versus VTE-NormoHcy (Dunnett’s test), P?NS; VTE- HyperHcy versus healthy controls, P?0.01; VTE-NormoHcy versus healthy controls, P?NS. †VTE-HyperHcy versus VTE-NormoHcy (Dunnett’s test), P?NS; VTE- yperHcy versus healthy controls, P?0.001; VTE-NormoHcy versus healthy controls, P?0.01. 1374
Activated Protein C, Thrombosis, and Homocysteine
Therefore, the results of our study suggest that Hcy does not negatively influence the plasma APC levels and argue against the hypothesis that
– it inhibits the activation of protein C in vivo by interfering with the activity of thrombomodulin.
Recently, Lentz et al,26in an experimental study of mon- keys with diet-induced moderate HyperHcy, showed that
– the thrombin-stimulated endothelium of aortas from hyperhomocysteinemic animals activated protein C in vitro less effectively than that of control animals.
This study, which supports the hypothesis that Hcy interferes with protein C activation, is in apparent contradiction with our results. At least two possible explanations for their different results can be proposed.
First, Hcy would not affect protein C activation that is ongoing in vivo under physiological conditions, whereas it would interfere with its activation at sites at which athero- genic or thrombogenic stimuli injured the endothelium and increased the local concentration of thrombin.
Second, due to the different relative densities of endothelial cell protein C receptor and thrombomodulin on the endothelium of large vessels and capillaries,1the regulation of protein C activation may differ in the two vascular districts. Although Lentz et al26 measured protein C activation by the endothelium of the aorta, we measured circulating APC, which mostly reflects protein C activation occurring in the microcirculation.
On the basis of the considerations above, we speculate that
– Hcy does not interfere with protein C activation ongoing in the micro- circulation under physiological conditions, whereas
– it could inhibit protein C activation on large, injured vessels.
In conclusion, our study shows that APC plasma levels are high in patients with previous episodes of VTE in whom the plasma levels of F1?2 are normal. Therefore, APC plasma levels represent a sensitive marker of activation of the hemostatic system. In addition, the study showed that high Hcy levels are not associated with heightened thrombin generation and do not interfere with the activation of protein C under physiological conditions in vivo. Further studies are needed to unravel the mechanism(s) by which HyperHcy increases the risks for atherosclerosis and thrombosis.
References
1. Esmon CT, Ding W, Yasuhiro K, Gu J-M, Ferrel G, Regan LM, Stearns- Kurosawa DJ, Kurosawa S, Mather T, Laszik Z, Esmon NL. The protein C pathway: new insights. Thromb Haemost. 1997;78:70–74.
2. Bauer KA, Kass BL, Beeler DL, Rosenberg RD. Detection of protein C activation in humans. J Clin Invest. 1984;74:2033–2041.
3. Heeb MJ, Mosher D, Griffin JH. Inhibition and complexation of activated protein C by two major inhibitors in plasma. Blood. 1989;73:446–454.
4. Gruber A, Griffin JH. Direct detection of activated protein C in blood from human subjects. Blood. 1992;79:2340–2348.
5. Espan ˜a F, Zuazu I, Vicente V, Estelle ´s A, Marco P, Aznar J. Quantifi- cation of circulating activated protein C in human plasma by immuno- assays: enzyme levels are proportional to total protein C levels. Thromb Haemost. 1996;75:56–61.
6. Cattaneo M. Hyperhomocysteinemia: a risk factor for arterial and venous thrombotic disease. Int J Clin Lab Res. 1997;27:139–144.
7. Harpel PC, Zhang X, Borth W. Homocysteine and hemostasis: patho- genetic mechanisms predisposing to thrombosis. J Nutr. 1996;126: 1285S–1289S.
8. Rodgers GM, Conn MT. Homocysteine, an atherogenic stimulus, reduces protein C activation by arterial and venous endothelial cells. Blood. 1990;75:895–901.
9. Lentz SR, Sadler JE. Inhibition of thrombomodulin surface expression and protein C activation by the thrombogenic agent homocysteine. J Clin Invest. 1991;88:1906–1914.
10. Hayashi T, Honda G, Suzuki K. An atherogenic stimulus homocysteine inhibits cofactor activity of thrombomodulin and enhances thrombo- modulin expression in human umbilical vein endothelial cells. Blood. 1992;79:2930–2936.
saee original manuscript for further referencesz and for figures (not shown)
2012pharmaceutical
Amandeep Kaur
Aashir Awan, Phd
Abhisar Anand
Adina Hazan
Alan F. Kaul, PharmD., MS, MBA, FCCP
alexcrystal6
anamikasarkar
apreconasia
aptubman
Arav Gandhi
aviralvatsa
David Orchard-Webb, PhD
danutdaagmailcom
Demet Sag, Ph.D., CRA, GCP
Dror Nir
dsmolyar
Ethan Coomber
evelinacohn
FrasonKalapurackal
Gail S Thornton
Irina Robu
jayzmit48
jdpmd
jshertok
kellyperlman
Ed Kislauskis
larryhbern
Madison Davis
marzankhan
megbaker58
ofermar2020
Dr. Pati
pkandala
ritusaxena
Rick Mandahl
sjwilliamspa
Srinivas Sriram
stuartlpbi
Dr. Sudipta Saha
tildabarliya
zraviv06
zs22
Leave a Reply