Mitochondrial damage and DNA depletion in cord blood and umbilical cord from infants exposed in utero to Combivir
AIDS: Volume 18(7) 30 April 2004 pp 1013-1021
Divi, Rao La; Walker, Vernon Eb,c; Wade, Nancy Ad; Nagashima, Kunioe; Seilkop, Steven Kf; Adams, Mary Ellend; Nesel, Carol Jg; O'Neill, J Patrickh; Abrams, Elaine Ji; Poirier, Miriam Ca
From the aCDI Section, National Cancer Institute, NIH, Bethesda, Maryland, bLovelace Respiratory Research Institute, Albuquerque, New Mexico, cWadsworth Center, New York State Department of Health, dDepartment of Pediatrics, Children's Hospital at Albany Medical Center, Albany, New York, eSAIC, National Cancer Institute at Frederick, Frederick, Maryland, fSKS Consulting Services, Siler City, North Carolina, gWestat Inc., Rockville, Maryland, hGenetics Laboratory, University of Vermont, Burlington, Vermont, iDepartment of Pediatrics, Columbia University College of Physicians and Surgeons and Harlem Hospital Center, New York, USA.
Objective: Although most uninfected infants born to women infected with HIV-1 show no clinical evidence of mitochondrial compromise, mitochondrial dysfunction has been reported in children born to women receiving zidovudine and/or lamivudine during pregnancy. In this pilot study we examined mitochondrial integrity in HIV-1-uninfected infants born to HIV-1-infected women receiving Combivir during pregnancy.
Design: Samples of umbilical cord and cord blood were obtained from HIV-1-uninfected infants born to either HIV-1-infected women receiving Combivir therapy during pregnancy (n = 10) or HIV-1-uninfected women (n = 9).
Methods: Mitochondrial morphological integrity was examined in umbilical cords (n = 16) by electron microscopy and mtDNA quantity was determined in DNA from cord blood (n = 18) and umbilical cord (n = 18) by PCR-chemiluminescence immunoassay detection.
Results: In umbilical cords from six of nine infants born to HIV-1-infected mothers taking Combivir moderate to severe mitochondrial morphological damage was observed (P = 0.011), while none of seven unexposed infants showed similar damage. Compared to unexposed infants, statistically significant mtDNA depletion was observed in umbilical cord (P = 0.006) and cord blood (P = 0.003) from drug-exposed infants.
Conclusions: A cohort of HIV-1-uninfected Combivir-exposed infants with no clinical symptoms showed morphological and molecular evidence of mitochondrial damage.
The antiretroviral nucleoside analog zidovudine (ZDV) reduces maternal-fetal vertical viral transmission when given to pregnant women infected with HIV-1. The Centers for Disease Control and Prevention recommendations include ZDV given daily for the last 6 months of pregnancy, intravenous drug dosing during labor and delivery, and 6 weeks of post-birth ZDV administration to the infant. Currently in the USA, many HIV-1-infected pregnant women are given Combivir (GlaxoSmithKline, Research Triangle Park, North Carolina, USA), which contains two nucleoside reverse transcriptase inhibitors (NRTI), ZDV and lamivudine (3TC).Few abnormal clinical findings have been reported at birth or during the early years of life in NRTI-exposed infants. However, persistent mitochondrial dysfunction was reported in several HIV-1-uninfected children born to HIV-1-infected mothers who had received either ZDV or both ZDV and 3TC during pregnancy, and two infants died with severe persistent mitochondrial toxicity at 11 and 13 months of age. Mitochondrial DNA (mtDNA) depletion has been reported in cord blood from HIV-1-uninfected newborn infants born to HIV-1-infected mothers taking either ZDV alone or Combivir during pregnancy. In addition in utero exposure to ZDV or dual NRTI therapy has resulted in at least one abnormally high plasma lactate reading in 35 of 38 infants up to 6 months of age. In adult HIV- 1-infected patients, long-term NRTI use has been associated with skeletal muscle and cardiac wasting, elevated serum lactic acid, abnormal OXPHOS enzyme activities, mtDNA depletion and mitochondrial morphological compromise.
Because of the above mentioned human mitochondrial toxicities, and because exposure of pregnant Erythrocebus patas monkeys to human-equivalent ZDV/3TC protocols has induced mitochondrial dysfunction in the fetus, we hypothesized that mitochondrial morphological damage and mtDNA depletion might be evident in HIV-1-uninfected infants born to HIV-1-infected mothers taking Combivir during pregnancy. In this study we examined mitochondrial morphology in umbilical cord artery endothelial cells by electron microscopy (EM), and mtDNA quantity in cord blood leukocytes and umbilical cords from newborn infants. Samples were obtained from HIV-1-uninfected children, either unexposed, or exposed in utero to Combivir.
Materials and methods
Patient population and sample collection
Samples were obtained within the context of a larger observational cohort study that included infants of uninfected women (n = 75) and infants of HIV-infected women receiving triple antiretroviral therapy during pregnancy (n = 75). The larger study included infants born at several different hospitals and was designed to evaluate mutagenic and genotoxic events occurring in NRTI-exposed infants. However, in response to reports of mitochondrial toxicity in NRTI-exposed infants, the results of mitochondrial morphological damage and DNA depletion assessments performed on samples from a subgroup of unexposed infants and infants exposed to Combivir are reported here. This study used fresh cord blood and umbilical cord collected from infants born at Albany Medical Center during the day shift. After delivery of the neonate, cord blood was collected from the placenta into heparinized tubes and centrifuged to separate plasma from cells. The mononuclear cells were collected over Ficoll and cryopreserved at 70°C for future DNA extraction. Small sections of umbilical cord were placed into freshly prepared fixative (see below), and a long section of the umbilical cord was snap frozen and stored at -70oC for future DNA extraction. Samples from each newborn were coded with a patient identification number.
The protocol for the cohort study was reviewed and approved by the Institutional Review Board at Albany Medical Center. Written informed consent was obtained from each participating woman and from each child's parent/guardian. Questionnaire, medical history and chart data provided information on maternal HIV status, smoking history, alcohol consumption, medications and illicit drugs (cannabinoids, cocaine, opiates, and amphetamines) used during the pregnancy. Infant data collected from the records included, in part, birth weight, estimated gestational age at delivery, and HIV-1 status. Infants were defined as HIV- uninfected if they had at least two negative HIV PCR tests on separate occasions after the age of 1 month.
After clamping the cord at birth, a portion of umbilical cord was immediately cut into thin (~5-mm) cross-sections and fixed in 4% paraformaldehyde: 2% glutaraldehyde in 100 mM phosphate buffer, pH 7.4, at 4°C overnight. After 24 h the fixative was replaced with sodium cacodylate buffer (0.1 M). The tissues were post-fixed in 1% osmium overnight, rinsed in aqueous solution containing 4.5% sucrose and then dehydrated in a series of graded alcohol washes (35-100%) followed by 100% propylene oxide. The tissues were infiltrated in equal volumes of propylene oxide and epoxy resin (LX-112, Ladd Research, Williston, Vermont, USA), embedded in pure epoxy resin, and cured in an oven at 60°C for 48 h. Thin sections were made longitudinally and horizontally with an ultramicrotome (Ultracut E, Leica, Northvale, New Jersey, USA) and mounted on a 200-mesh grid. The thin sections were stained with uranyl acetate and lead citrate in a stainer (LKB Ultrostainer, Leica), stabilized by carbon evaporation (Denton Vacuum, Morrestown, New Jersey, USA), observed, and photographed at 30 000 ¥ magnification with an electron microscope (H7000, Hitachi, Tokyo, Japan) operated at 75 kV. Endothelial cells of the umbilical cord artery, which are rich in mitochondria, were selected for evaluation.
The photomicrographs were coded and subsequently examined by all the authors, and research collaborators at the National Cancer Institute and the Wadsworth Center. Finally, a workshop was held for a review of the photomicrographs in toto by four board-certified pathologists (see Acknowledgements). There was good agreement among the various review groups, and a simple grading scheme was devised based upon the ultrastructural pathology observed.
DNA preparation and quantification
DNA was isolated from infant umbilical cord homogenates and cord blood mononuclear cells using the QIAamp isolation kit (Qiagen, Valencia, California, USA). Briefly, the cells were washed and resuspended in phosphate buffered saline (200 ml), then lysed at 50°C in the presence of proteinase K and the Qiagen kit buffer. Samples were incubated with RNase A (Qiagen) at 37°C for 20 min before column purification of DNA. DNA concentrations (mg/ml) were determined by spectrophotometry (A260) using a Hewlett Packard 8453 Spectrophotometer.
Determination of mtDNA quantity by PCR-chemiluminescence immunoassay detection
The quantity of mtDNA was assessed relative to nuclear DNA using the PCR-chemiluminescence immunoassay detection (CID). Briefly, DNA isolated from human tissues (unknowns), and nuclear and mtDNA isolated separately from cultured human MCL5 lymphoblastoid cells (standard curves) were subjected to simultaneous PCR of the mitochondrial D loop (500 base pairs) and the nuclear 18S RNA (521 base pairs) genes. Aliquots of PCR products (amplicons) were diluted and quantified by CID.
For the CID, neutravidin-coated 96-well microtiter plates (Pierce, Rockford, Illinois, USA) were washed, blocked with 0.25% casein, and coated with either reverse primer for D loop or reverse primer for 18 S RNA, both of which were biotinylated at the 5¢ end. The PCR-amplified samples were placed in two D loop primer-coated wells and two 18S RNA primer-coated wells. Plates were hybridized, and incubated with anti-digoxigenin antibody conjugated to alkaline phosphatase (Roche, Indianapolis, Indiana, USA). CDP-Star containing Emerald II enhancer (Tropix, Bedford, Massachusetts, USA) was added and luminescence was read using a Tropix TR717 Microplate Luminometer (PE Applied Biosystems, Foster City, California, USA) at 542 nm. The luminescence is directly proportional to the gene copy numbers, which were calculated by comparison with the standard curves.Statistical analysis
Differences between unexposed infants and Combivir-exposed infants with respect to mitochondrial damage the quantity of mtDNA, and the treatment group characteristics was evaluated statistically using the Mann-Whitney U Test. Because of the small sample sizes, exact P-values were calculated. Testing was two-sided (i.e., null hypothesis of no difference between controls and treated groups), with statistical significance determined at P < 0.05. Associations between pairs of variables, such as the duration of in utero Combivir exposure and the degree of mitochondrial pathology by EM, were characterized with two measures of correlation: (i) the Pearson product-moment correlation coefficient (r) and (ii) Kendall's Tau, which provides a distribution-free test of variable independence. P < 0.05 was considered significant with respect to the null hypothesis of no correlation. Statistical analysis was performed using SAS software system (Cary, North Carolina, USA).
Maternal NRTI therapy and HIV-1 status of infants
This study included nine infants born to uninfected mothers and 10 infants born to HIV-1-infected mothers taking Combivir at the time of delivery. For eight of the infants born to HIV-1-infected mothers, Combivir was the only NRTI therapy received during the pregnancy. However, two of the HIV-1-infected mothers received combinations containing stavudine and didanosine for different periods of time before the Combivir was given for the last 0.8 months (patient 11) or 5.4 months (patient 13) of pregnancy. There was no difference between the age at delivery for uninfected women and HIV-1-infected women receiving Combivir. For HIV-1-infected women, values for plasma HIV RNA, CD4 cell count and CD4 cell percent are shown in. There were no differences in mean birth weights or estimated gestational age of unexposed versus Combivir-exposed infants. All infants born to HIV-infected mothers had two or more negative HIV PCR tests between 1 and 6 months of age and were confirmed HIV-uninfected.
Ultrastructural morphology of umbilical cord mitochondria
Samples from 16 infants were examined by transmission EM for ultrastructural changes in cellular morphology in mitochondria-rich endothelial cells of the umbilical cord artery. The photomicrographs were coded and systematically reviewed as described in Methods. Decoding of the specimens revealed that there were seven infants born to unexposed, uninfected mothers and nine infants born to mothers receiving Combivir at the time of delivery.
Examination of umbilical cord EM images revealed a range in mitochondrial morphological changes. The ultrastructural pathology in endothelial cells, with consideration of adjacent cells/structures, was graded based upon the degree of mitochondrial damage from mild (+) to severe (+++).
Six of the seven unexposed infants showed either no damage or minimal swelling of an occasional mitochondrion and were designated WNL, while one unexposed infant had mild (+) mitochondrial damage regarded as early autolytic changes (see Discussion). In contrast, increased mitochondrial damage (++ and +++) was strikingly evident in six of nine infants exposed in utero to Combivir. One infant (11), exposed in utero to other drug regimens before a final month of Combivir exposure, had mild (+) mitochondrial damage. Overall, there was substantial evidence that the higher degree of mitochondrial damage observed by EM in the Combivir-exposed group of infants relative to that observed in the unexposed group was not due to chance (P = 0.011). A strong correlation between the duration of Combivir exposure and the degree of mitochondrial damage (Pearson product-moment r = 0.85, P < 0.001; Kendall's Tau = 0.68, P = 0.002) provided additional evidence of a causal relationship.
mtDNA quantity in umbilical cord and cord blood mononuclear cells
The PCR-CID used to quantify mtDNA compares the PCR products obtained when amplifying with primers for a housekeeping gene, the 18S RNA, and a mitochondrial gene, the D loop. In both cord blood leukocytes and umbilical cord, mtDNA depletion in Combivir-exposed infants was statistically significant, compared to unexposed infants (P = 0.003 and 0.006, respectively).
Relationships between markers of maternal disease status and mitochondrial compromise in umbilical cords of Combivir-exposed infants
Maternal viral and clinical status was monitored during pregnancy using plasma HIV-1 RNA copy number, CD4 cell numbers, and percent CD4 cells positive. These parameters were assayed twice, once near initiation of NRTI therapy and a second time near delivery. Values obtained at both times were compared to the degree of mitochondrial ultrastructural damage and mitochondrial DNA depletion in umbilical cord tissue from infants born to mothers receiving Combivir (Pearson product-moment correlation). The expectation was that there might be a positive correlation between poor clinical status in HIV-1-infected mothers and greater mitochondrial damage in fetal tissues; however, the analyses showed no significant correlations between these three markers of maternal viral/clinical status and infant mitochondrial toxicity (P-values ranging from 0.324 to 0.933).
Several additional factors related to mother and child health were compared to extent of mitochondrial damage to explore differences in the Combivir-exposed and unexposed groups of infants. Self-reports of maternal smoking, alcohol use, and illicit drug use during pregnancy were limited and were not correlated with mitochondrial damage in infants. Medical chart reviews did not reveal use of legal drugs/medications, other than NRTI, that are suspected or known to affect mitochondrial structure or function. There was no significant correlation between maternal age and the degree of mitochondrial damage in Combivir-exposed infants, nor was there any correlation between prematurity and mitochondrial damage in Combivir-exposed infants. The limited size of this pilot study has precluded assessment of whether or not factors other than maternal HIV-1-status and NRTI use contributed to changes in the endpoints assessed in infants.
In previous studies mitochondrial toxicity has been: extensively documented in adult HIV-1-infected patients receiving NRTI therapy; found in a small number of infants born to HIV-1-infected women receiving ZDV or Combivir during pregnancy; and reported in infants of pregnant patas monkeys given either ZDV alone or ZDV plus 3TC during gestation. We therefore hypothesized that asymptomatic and uninfected infants born to HIV-1-infected women receiving Combivir therapy during pregnancy might sustain mitochondrial compromise that would be evident on a molecular level. The data presented here demonstrate morphological and molecular mitochondrial compromise in HIV-1-uninfected infants born to HIV-1-infected mothers who received Combivir during pregnancy. Moderate to severe mitochondrial morphological damage was found in umbilical cords from six of nine Combivir-exposed infants. In contrast, six of seven unexposed infants exhibited either no damage or minimal damage judged to be within normal limits. As the umbilical cord is fetal tissue, the ultrastructural pathology found in umbilical cord endothelial cells suggests the possibility that mitochondrial damage might also occur in tissues that are highly dependent upon mitochondrial integrity for normal function. Consistent with the mitochondrial morphological damage was the significant depletion in mtDNA quantity observed in umbilical cord and cord blood samples from Combivir-exposed infants. Taken together, the EM morphology and mtDNA depletion data demonstrate that in many HIV-1-uninfected infants exposed in utero to Combivir, mitochondrial damage may be present at the molecular level, even in the absence of adverse clinical effects.
The NRTI are 'dideoxy-type' nucleoside analogs that become phosphorylated and incorporated into viral DNA. Their antiretroviral activity is the result of termination of DNA chain elongation (replication), as well as inhibition of the viral reverse transcriptase. In the host, these drugs become incorporated into nuclear and mtDNA, inhibit DNA replication and induce genotoxic sequelae. In mitochondria, the NRTI specifically inhibit the mitochondrial DNA polymerase g, which further inhibits mtDNA replication and produces mtDNA depletion. Long-term NRTI use in adults has been associated with mitochondrial toxicity demonstrated morphologically by membrane disruption, vacuolization and cristae fragmentation. Clinically, the manifestations include skeletal muscle myopathies and cardiac insufficiency, elevated serum lactate and altered OXPHOS enzyme activities. Whereas individually the NRTI exhibit characteristic patterns of genotoxicity and mitochondrial toxicity, there is evidence that the combination of two NRTI may enhance a given toxic effect. This has been observed in cell culture models for genotoxicity and mitochondrial toxicity, and in primate studies. As most therapies for HIV-1 typically include two NRTI, it is likely that future preclinical studies involving drug combinations will assume greater importance in predicting toxic consequences of multiple-drug exposures.
Clinical evidence of mitochondrial toxicity has largely been lacking in infants born to HIV-1-infected women. A retrospective study of about 10 000 children exposed in utero to ZDV alone or to Combivir reported no evidence of mortality occurring in the first 5 years of life that could be associated with mitochondrial dysfunction. In contrast, Blanche et al. reported the deaths of two HIV-1-uninfected Combivir-exposed children at about 1 year of age with severe mitochondrial toxicity. Furthermore, several other infants under 5 years of age who were born to women taking ZDV or ZDV plus 3TC, had symptoms consistent with mitochondrial dysfunction that included seizures, abnormal magnetic resonance imaging scans, aberrant OXPHOS enzyme activities and/or persistent lactic acidosis. Depletion of mtDNA has been reported in infants born to HIV-1-infected mothers receiving ZDV plus 3TC during pregnancy. Poirier et al. found statistically significant mtDNA depletion in cord blood and peripheral blood, at 1 and 2 years of age, from HIV-1-uninfected infants born to HIV-1-infected women receiving ZDV therapy during pregnancy. In addition Shiramizu et al. reported statistically significant mtDNA depletion in cord blood and placenta from infants born to mothers taking ZDV plus 3TC during pregnancy, compared to infants born to uninfected women. Finally, elevated plasma lactate (>= 2.1 mmol/l) was documented in 35 of 38 HIV-1-uninfected infants followed to 6 months of age (median three samples per infant) after in utero and perinatal exposure to single or dual NRTI therapy. All of these studies indicate that mitochondrial compromise occurs on a molecular and biochemical level in NRTI-exposed infants in the absence of major clinical adverse effects.
Further support for NRTI-induced fetal mitochondrial compromise comes from studies in the Erythrocebus patas monkey, where drug effects can be evaluated in the absence of HIV-1 infection. Newborn fetal patas monkeys exposed during gestation to human- equivalent doses of ZDV alone or ZDV plus 3TC exhibit mitochondrial ultrastructural damage, alterations in OXPHOS enzyme specific activities and mtDNA depletion in heart, skeletal muscle and brain. In addition, the ultrastructural mitochondrial morphology observed in patas umbilical cord artery from fetuses exposed in utero to human equivalent doses of ZDV plus 3TC (Divi R.L., Kuo M., Nagashima K., St Claire M., Poirier M., unpublished data) closely resembles the human umbilical cord findings reported here. Similar to human infants, patas monkeys who were exposed to NRTI transplacentally and followed for the first year of life show no adverse clinical effects, even though mitochondrial compromise is evident on a molecular level.
The correlation between duration of Combivir therapy and degree of pathological changes in umbilical cord mitochondrial morphology in Combivir-exposed neonates suggests that EM may be a useful tool in defining the prevalence of mitochondrial damage in NRTI-exposed infants. Interestingly, an increase in the number of mitochondria per endothelial cell appeared to correlate with the duration of therapy, supporting the hypothesis that mitochondria proliferate as a compensatory mechanism in response to mitochondrial compromise.
The origin of minimal swelling of an occasional mitochondrion in umbilical cords from some unexposed infants is uncertain. Potential sources of hypoxia that could lead to this phenomenon include a delay in tissue collection at birth and/or stress associated with birth. The potential for hypoxic stress in the perinatal period is consistent with the fact that lactate and the lactate : pyruvate ratio are typically elevated at birth. While plasma lactate levels in umbilical cord blood from healthy newborns have been reported, we are unaware of other EM studies showing umbilical cord endothelial cell morphology for healthy newborn infants. The consensus of those reviewing all the coded photomicrographs was that the occurrence of minimal swelling in an occasional mitochondrion fell 'within normal limits'. The umbilical cord mitochondria from unexposed infant 4 showed multifocal areas of mildly distended rough endoplasmic reticulum (rER), disruptions and detachment of ribosomes, and minor condensation of nuclear chromatin. These findings are consistent with mild autolysis of the tissue sample (probably associated with delayed tissue collection), and, thus, are not considered to be clinically relevant. In contrast, the umbilical cord mitochondria from NRTI-exposed infant 11 had no autolysis or alterations in rER and displayed changes suggestive of an early form of the damage in samples graded ++ or +++.
In the current study, it is not possible to determine to what extent the observed mtDNA depletion may be due to maternal HIV-1 status or to antiretroviral drug administration. Two studies have reported HIV-1-related mtDNA depletion in the absence of NRTI therapy and further NRTI-associated mtDNA depletion. Poirier et al. found significant mtDNA depletion in infants of HIV-1-infected women not given antiretroviral therapy, and further significant depletion when ZDV was given during pregnancy. Cote et al. also reported significant HIV-1-related mtDNA depletion in untreated adults, with further significant mtDNA depletion in NRTI-treated patients. As HIV-1-infected patients free of antiretroviral drugs are extremely rare in most developing countries, it is frequently not possible to observe drug-induced effects in the absence of HIV-1-infection. However, evidence for a causal relationship between fetal exposure to Combivir and mitochondrial pathology is presented in this study via the significant positive correlation between exposure duration and the extent of mitochondrial damage in Combivir-exposed newborns.
Here we have shown morphological and molecular evidence of mitochondrial compromise in a small group of infants exposed to Combivir during pregnancy. The study confirms previous reports of mitochondrial toxicity in infants born to HIV-1-infected women receiving NRTI therapy, but is unique in demonstrating mitochondrial morphological damage in HIV-1-uninfected infants born to HIV-1-infected mothers. Similar to most uninfected infants of HIV-1-infected mothers, these children showed no clinical abnormalities during follow-up through 1 year of age. However, it is possible that a subtle degree of persistent mitochondrial compromise, when combined with other unknown factors, may result in mitochondrial dysfunction similar to that described by Blanche et al.